THE SCIENCE AND INFORMATION ORGANIZATION

www. t hes ai . or g | i nf o@t hes ai . or g

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(i)
www.ijacsa.thesai.org
Editorial Preface
From the Desk of Managing Editor
IJACSA seems to have a cult following and was a humungous success during 2011. We at The Science and Information
Organization are pleased to present the January 2012 Issue of IJACSA.
While it took the radio 38 years and the television a short 13 years, it took the World Wide Web only 4 years to reach 50
million users. This shows the richness of the pace at which the computer science moves. As 2012 progresses, we seem to
be set for the rapid and intricate ramifications of new technology advancements.
With this issue we wish to reach out to a much larger number with an expectation that more and more researchers get
interested in our mission of sharing wisdom. The Organization is committed to introduce to the research audience
exactly what they are looking for and that is unique and novel. Guided by this mission, we continuously look for ways to
collaborate with other educational institutions worldwide.
Well, as Steve Jobs once said, Innovation has nothing to do with how many R&D dollars you have, its about the people
you have. At IJACSA we believe in spreading the subject knowledge with effectiveness in all classes of audience.
Nevertheless, the promise of increased engagement requires that we consider how this might be accomplished,
delivering up-to-date and authoritative coverage of advanced computer science and applications.
Throughout our archives, new ideas and technologies have been welcomed, carefully critiqued, and discarded or
accepted by qualified reviewers and associate editors. Our efforts to improve the quality of the articles published and
expand their reach to the interested audience will continue, and these efforts will require critical minds and careful
consideration to assess the quality, relevance, and readability of individual articles.
To summarise, the journal has offered its readership thought provoking theoretical, philosophical, and empirical ideas
from some of the finest minds worldwide. We thank all our readers for their continued support and goodwill for IJACSA.
We will keep you posted on updates about the new programmes launched in collaboration.
We would like to remind you that the success of our journal depends directly on the number of quality articles submitted
for review. Accordingly, we would like to request your participation by submitting quality manuscripts for review and
encouraging your colleagues to submit quality manuscripts for review. One of the great benefits we can provide to our
prospective authors is the mentoring nature of our review process. IJACSA provides authors with high quality, helpful
reviews that are shaped to assist authors in improving their manuscripts.
We regularly conduct surveys and receive extensive feedback which we take very seriously. We beseech valuable
suggestions of all our readers for improving our publication.
Thank you for Sharing Wisdom and a very Happy New Year!






Managing Editor
IJACSA
Volume 3 Issue 1, January 2012
ISSN 2156-5570 (Online)
ISSN 2158-107X (Print)
2012 The Science and Information (SAI) Organization
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(ii)
www.ijacsa.thesai.org
Associate Editors
Dr. Zuqing Zhu
Service Provider Technology Group of Cisco Systems, San Jose
Domain of Research: Research and development of wideband access routers for hybrid
fibre-coaxial (HFC) cable networks and passive optical networks (PON)
Dr. Ka Lok Man
Department of Computer Science and Software Engineering at the Xi'an Jiaotong-
Liverpool University, China
Domain of Research: Design, analysis and tools for integrated circuits and systems;
formal methods; process algebras; real-time, hybrid systems and physical cyber
systems; communication and wireless sensor networks.
Dr. Sasan Adibi
Technical Staff Member of Advanced Research, Research In Motion (RIM), Canada
Domain of Research: Security of wireless systems, Quality of Service (QoS), Ad-Hoc
Networks, e-Health and m-Health (Mobile Health)
Dr. Sikha Bagui
Associate Professor in the Department of Computer Science at the University of West
Florida,
Domain of Research: Database and Data Mining.
Dr. T. V. Prasad
Dean, Lingaya's University, India
Domain of Research: Bioinformatics, Natural Language Processing, Image Processing,
Expert Systems, Robotics
Dr. Bremananth R
Research Fellow, Nanyang Technological University, Singapore
Domain of Research: Acoustic Holography, Pattern Recognition, Computer Vision, Image
Processing, Biometrics, Multimedia and Soft Computing
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(iii)
www.ijacsa.thesai.org
Reviewer Board Members

A Kathirvel
Karpaga Vinayaka College of Engineering and
Technology, India
Abbas Karimi
I.A.U_Arak Branch (Faculty Member) & Universiti
Putra Malaysia
Dr. Abdul Wahid
Gautam Buddha University, India
Abdul Khader Jilani Saudagar
Al-Imam Muhammad Ibn Saud Islamic University
Abdur Rashid Khan
Gomal Unversity
Dr. Ahmed Nabih Zaki Rashed
Menoufia University, Egypt
Ahmed Sabah AL-Jumaili
Ahlia University
Md. Akbar Hossain
Aalborg University, Denmark and AIT, Greeceas
Albert Alexander
Kongu Engineering College,India
Prof. Alcinia Zita Sampaio
Technical University of Lisbon
Amit Verma
Rayat & Bahra Engineering College, India
Ammar Mohammed Ammar
Department of Computer Science, University of
Koblenz-Landau
Arash Habibi Lashakri
University Technology Malaysia (UTM), Malaysia
Asoke Nath
St. Xaviers College, India
B R SARATH KUMAR
Lenora College of Engineering, India
Binod Kumar
Lakshmi Narayan College of Technology, India
Bremananth Ramachandran
School of EEE, Nanyang Technological University
Dr.C.Suresh Gnana Dhas
Park College of Engineering and Technology,
India
Mr. Chakresh kumar
Manav Rachna International University, India
Chandra Mouli P.V.S.S.R
VIT University, India
Chandrashekhar Meshram
Shri Shankaracharya Engineering College, India
Constantin POPESCU
Department of Mathematics and Computer
Science, University of Oradea
Prof. D. S. R. Murthy
SNIST, India.
Deepak Garg
Thapar University.
Prof. Dhananjay R.Kalbande
Sardar Patel Institute of Technology, India
Dhirendra Mishra
SVKM's NMIMS University, India
Divya Prakash Shrivastava
EL JABAL AL GARBI UNIVERSITY, ZAWIA
Dragana Becejski-Vujaklija
University of Belgrade, Faculty of organizational
sciences
Fokrul Alom Mazarbhuiya
King Khalid University
G. Sreedhar
Rashtriya Sanskrit University
Ghalem Belalem
University of Oran (Es Senia)
Hanumanthappa.J
University of Mangalore, India
Dr. Himanshu Aggarwal
Punjabi University, India
Huda K. AL-Jobori
Ahlia University
Dr. Jamaiah Haji Yahaya
Northern University of Malaysia (UUM), Malaysia
Jasvir Singh
Communication Signal Processing Research Lab
Jatinderkumar R. Saini
S.P.College of Engineering, Gujarat
Prof. Joe-Sam Chou
Nanhua University, Taiwan
Dr. Juan Jos Martnez Castillo
Yacambu University, Venezuela
Dr. Jui-Pin Yang
Shih Chien University, Taiwan
Dr. K.PRASADH
Mets School of Engineering, India
Ka Lok Man
Xian Jiaotong-Liverpool University (XJTLU)
Dr. Kamal Shah
St. Francis Institute of Technology, India
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(iv)
www.ijacsa.thesai.org
Kodge B. G.
S. V. College, India
Kohei Arai
Saga University
Kunal Patel
Ingenuity Systems, USA
Lai Khin Wee
Technischen Universitt Ilmenau, Germany
Latha Parthiban
SSN College of Engineering, Kalavakkam
Mr. Lijian Sun
Chinese Academy of Surveying and Mapping,
China
Long Chen
Qualcomm Incorporated
M.V.Raghavendra
Swathi Institute of Technology & Sciences, India.
Madjid Khalilian
Islamic Azad University
Mahesh Chandra
B.I.T, India
Mahmoud M. A. Abd Ellatif
Mansoura University
Manpreet Singh Manna
SLIET University, Govt. of India
Marcellin Julius NKENLIFACK
University of Dschang
Md. Masud Rana
Khunla University of Engineering & Technology,
Bangladesh
Md. Zia Ur Rahman
Narasaraopeta Engg. College, Narasaraopeta
Messaouda AZZOUZI
Ziane AChour University of Djelfa
Dr. Michael Watts
University of Adelaide, Australia
Miroslav Baca
University of Zagreb, Faculty of organization and
informatics / Center for biomet
Mohamed Ali Mahjoub
Preparatory Institute of Engineer of Monastir
Mohammad Talib
University of Botswana, Gaborone
Mohammed Ali Hussain
Sri Sai Madhavi Institute of Science &
Technology
Mohd Helmy Abd Wahab
Universiti Tun Hussein Onn Malaysia
Mohd Nazri Ismail
University of Kuala Lumpur (UniKL)
Mueen Uddin
Universiti Teknologi Malaysia UTM
Dr. Murugesan N
Government Arts College (Autonomous), India
Nitin S. Choubey
Mukesh Patel School of Technology
Management & Eng
Dr. Nitin Surajkishor
NMIMS, India
Paresh V Virparia
Sardar Patel University
Dr. Poonam Garg
Institute of Management Technology,
Ghaziabad
Raj Gaurang Tiwari
AZAD Institute of Engineering and Technology
Rajesh Kumar
National University of Singapore
Rajesh K Shukla
Sagar Institute of Research & Technology-
Excellence, India
Dr. Rajiv Dharaskar
GH Raisoni College of Engineering, India
Prof. Rakesh. L
Vijetha Institute of Technology, India
Prof. Rashid Sheikh
Acropolis Institute of Technology and Research,
India
Ravi Prakash
University of Mumbai
Rongrong Ji
Columbia University
Dr. Ruchika Malhotra
Delhi Technological University, India
Dr.Sagarmay Deb
University Lecturer, Central Queensland
University, Australia
Saleh Ali K. AlOmari
Universiti Sains Malaysia
Dr. Sana'a Wafa Al-Sayegh
University College of Applied Sciences UCAS-
Palestine
Santosh Kumar
Graphic Era University, India
Sasan Adibi
Research In Motion (RIM)
Saurabh Pal
VBS Purvanchal University, Jaunpur
Seyed Hamidreza Mohades Kasaei
University of Isfahan
Shahanawaj Ahamad
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(v)
www.ijacsa.thesai.org
The University of Al-Kharj
Shaidah Jusoh
University of West Florida
Sikha Bagui
Zarqa University
Dr. Smita Rajpal
ITM University
Suhas J Manangi
Microsoft
SUKUMAR SENTHILKUMAR
Universiti Sains Malaysia
Sunil Taneja
Smt. Aruna Asaf Ali Government Post Graduate
College, India
Dr. Suresh Sankaranarayanan
University of West Indies, Kingston, Jamaica
T C.Manjunath
Visvesvaraya Tech. University
T V Narayana Rao
Hyderabad Institute of Technology and
Management, India
T. V. Prasad
Lingaya's University
Taiwo Ayodele
Lingaya's University
Totok R. Biyanto
Infonetmedia/University of Portsmouth
Varun Kumar
Institute of Technology and Management, India
Vellanki Uma Kanta Sastry
Sreeneedhi
Dr. V. U. K. Sastry
SreeNidhi Institute of Science and Technology
(SNIST), Hyderabad, India.
Vinayak Bairagi
Sinhgad Academy of engineering, India
Vitus S.W. Lam
The University of Hong Kong
Vuda Sreenivasarao
St.Marys college of Engineering & Technology,
Hyderabad, India
Y Srinivas
GITAM University
Mr.Zhao Zhang
City University of Hong Kong, Kowloon, Hong
Kong
Zhixin Chen
ILX Lightwave Corporation
Zuqing Zhu
University of Science and Technology of China
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(vi)
www.ijacsa.thesai.org
CONTENTS
Paper 1: Analysis and Selection of Features for Gesture Recognition Based on a Micro Wearable Device
Authors: Yinghui Zhou, Lei Jing, Junbo Wang, Zixue Cheng
PAGE 1 7

Paper 2: A Framework for Improving the Performance of Ontology Matching Techniques in Semantic Web
Authors: Kamel Hussein Shafaamri, Jalal Omer Atoum
PAGE 8 14

Paper 3: Fingerprint Image Enhancement: Segmentation to Thinning
Authors: Iwasokun Gabriel Babatunde, Akinyokun Oluwole Charles, Alese Boniface Kayode,
Olabode Olatubosun
PAGE 15 24

Paper 4: Data Warehouse Requirements Analysis Framework: Business-Object Based Approach
Authors: Anirban Sarkar
PAGE 25 34

Paper 5: A new graph based text segmentation using Wikipedia for automatic text summarization
Authors: Mohsen Pourvali, Ph.D. Mohammad Saniee Abadeh
PAGE 35 39

Paper 6: Automated Periodontal Diseases Classification System
Authors: Aliaa A. A. Youssif, Abeer Saad Gawish, Mohammed Elsaid Moussa
PAGE 40 48

Paper 7: Communication and migration of an embeddable mobile agent platform supporting runtime code
mobility
Authors: Mohamed BAHAJ, Khaoula ADDAKIRI, Noreddine GHERABI
PAGE 49 54

Paper 8: An Adaptive parameter free data mining approach for healthcare application
Authors: Prof. Dipti Patil, Bhagyashree Agrawal, Snehal Andhalkar, Richa Biyani, Mayuri Gund, Dr.
V.M.Wadhai
PAGE 55 59

Paper 9: Question Answering System for an Effective Collaborative Learning
Authors: Prof. Kohei Arai, Anik Nur Handayani
PAGE 60 64

Paper 10: An Efficient Method For Multichannel Wireless Mesh Networks With Pulse Coupled Neural Network
Authors: S.Sobana, S.Krishna Prabha
PAGE 65 68



(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(vii)
www.ijacsa.thesai.org
Paper 11: A Congestion Avoidance Approach in Jumbo Frame-enabled IP Network
Authors: Aos Anas Mulahuwaish, Kamalrulnizam Abu Bakar, Kayhan Zrar Ghafoor
PAGE 69 75

Paper 12: Cross Layer QoS Support Architecture with Integrated CAC and Scheduling Algorithms for WiMAX
BWA Networks
Authors: Prasun Chowdhury, Iti Saha Misra, Salil K Sanyal
PAGE 76 92

Paper 13: A Conceptual Design Model for High Performance Hotspot Network Infrastructure (GRID WLAN)
Authors: Udeze Chidiebele. C, Okafor Kennedy .C, Prof. H. C Inyiama, Dr C. C. Okezie
PAGE 93 99

Paper 14: An enhanced Scheme for Reducing Vertical handover latency
Authors: Mohammad Faisal, Muhammad Nawaz Khan
PAGE 100 105

Paper 15: A Feasible Rural Education System
Authors: Lincy Meera Mathews, Dr Bandaru Rama Krishna Rao
PAGE 106 111

Paper 16: Efficient Threshold Signature Scheme
Authors: Sattar J Aboud, Mohammad AL-Fayoumi
PAGE 112 116

Paper 17: Fault Tolerant Platform for Application Mobility across devices
Authors: T. N. Anitha, Jayanth. A
PAGE 117 120

Paper 18: Viable Modifications to Improve Handover Latency in MIPv6
Authors: Mr.Purnendu Shekhar Pandey, Dr.Neelendra Badal
PAGE 121 125

Paper 19: Different Protocols for High Speed Networks
Authors: Dr.Srinivasa Rao Angajala
PAGE 126 128

Paper 20: Wideband Wireless Access Systems Interference Robustness: Its Effect on Quality of Video
Streaming
Authors: Aderemi A. Atayero, Oleg I. Sheluhin, Yuri A. Ivanov, Julet O. Iruemi
PAGE 129 136

Paper 21: Survey on Impact of Software Metrics on Software Quality
Authors: Mrinal Singh Rawat, Arpita Mittal, Sanjay Kumar Dubey
PAGE 137 141

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
(viii)
www.ijacsa.thesai.org
Paper 22: A Cost-Effective Approach to the Design and Implementation of Microcontroller-based Universal
Process Control Trainer
Authors: Udeze Chidiebele. C, Uzedeh Godwin, Prof. H. C Inyiama, Dr C. C. Okezie
PAGE 142 147

Paper 23: Self-regulating Message Throughput in Enterprise Messaging Servers A Feedback Control
Solution
Authors: Ravi Kumar G, C.Muthusamy, A.Vinaya Babu
PAGE 148 155

Paper 24: Improved Face Recognition with Multilevel BTC using Kekres LUV Color Space
Authors: H.B. Kekre, Dr. Sudeep Thepade, Sanchit Khandelwal, Karan Dhamejani, Adnan Azmi
PAGE 156 160

Paper 25: Scenario-Based Software Reliability Testing Profile for Autonomous Control System
Authors: Jun Ai, Jingwei Shang, Peng Wang
PAGE 161 165

Paper 26: Identification of Critical Node for the Efficient Performance in Manet
Authors: Shivashankar, B.Sivakumar and G.Varaprasad
PAGE 166 171

Paper 27: Secret Key Agreement Over Multipath Channels Exploiting a Variable-Directional Antenna
Authors: Valery Korzhik, Viktor Yakovlev,Yuri Kovajkin, Guillermo Morales-Luna
PAGE 172 178

Paper 28: The Relationships of Soft Systems Methodology (SSM), Business Process Modeling and e-
Government
Authors: Dana Indra Sensuse, Arief Ramadhan
PAGE 179 183

Paper 29: Re-tooling Code Structure Based Analysis with Model-Driven Program Slicing for Software
Maintenance
Authors: Oladipo Onaolapo Francisca
PAGE 184 189

Paper 30: Transform Domain Fingerprint Identification Based on DTCWT
Authors: Jossy P. George, Abhilash S. K., Raja K. B.
PAGE 190 195

Paper 31: Effective Security Architecture for Virtualized Data Center Networks
Authors: Udeze Chidiebele. C, Prof. H. C Inyiama, Okafor Kennedy .C, Dr C. C. Okezie
PAGE 196 200

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
1 | P a g e
www.ijacsa.thesai.org

Analysis and Selection of Features for Gesture
Recognition Based on a Micro Wearable Device

Yinghui Zhou
1
, Lei Jing
2
, Junbo Wang
2
, Zixue Cheng
2
1. Graduate School of Computer Science and Engineering
2. School of Computer Science and Engineering
University of Aizu
Aizu-Wakamatsu, Japan


AbstractMore and More researchers concerned about designing
a health supporting system for elders that is light weight, no
disturbing to user, and low computing complexity. In the paper,
we introduced a micro wearable device based on a tri-axis
accelerometer, which can detect acceleration change of human
body based on the position of the device being set. Considering the
flexibility of human finger, we put it on a finger to detect the
finger gestures. 12 kinds of one-stroke finger gestures are defined
according to the sensing characteristic of the accelerometer.
Feature is a paramount factor in the recognition task. In the
paper, gestures features both in time domain and frequency
domain are described since features decide the recognition
accuracy directly. Feature generation method and selection
process is analyzed in detail to get the optimal feature subset from
the candidate feature set. Experiment results indicate the feature
subset can get satisfactory classification results of 90.08%
accuracy using 12 features considering the recognition accuracy
and dimension of feature set.
KeywordsInternet of Things; Wearable Computing; Gesture
Recognition; Feature analysis and selection; Accelerometer.
I. INTRODUCTION
Internet of Things (IoTs) has become a hot topic in the
computer science field, which indicates that all objects in the
environment like human, home appliances, building, and
service equipment can be sensed, identified, even controlled via
the internet. IoTs will promote many development of
application system, such as health supporting system for elder.
Many countries are facing a serious society issue of population
ageing. One common trend is more and more elders living alone
and less able to benefit from the care and supporting that might
be available in a large household. Investigation from World
Health Organization indicates, in Japan, the proportion of
people living in 3-generation households has fallen from 46% in
1985 to 20.5% in 2006. Health care both physical and mental
becomes an important problem in current society.
Health supporting system has been studied widely in recent
years [1] [2]. Two kinds of main supporting way focus on
speech-based communication and activity-based recognition.
The former provides a direct and effective way to know users
intension, which has been used in the hospital and household [3]
[4]. However, voice signal is sensitive to environment sound
such as a TV being on, so that sometimes hard to pick out useful
speech signal made by a user. Even under certain circumstance,
the user may too weak to make a voice to call for a help.
Activity recognition provides an active and undisturbed way for
elderly care. For example, if an elder person falls down, the
system of falling recognition can send message automatically
asking for help.
Among the human activities, finger gestures are the most
flexible ones. In our daily life, most of works are performed by
hands. Gesture recognition is significant for learning user
behavior, realizing for device control, and getting user
intention. In the paper, we designed a wearable device with an
accelerometer to detect finger gestures. Based on the
accelerometer characteristic, a variety of finger gestures are
defined in a 3D space.
Particularly, gestures features are studied both in time
domain and frequency domain considering the paramount
importance of feature generation in recognition task. Each kind
of candidate features and their combination are analyzed based
on stepwise regression algorithm to form a feature vector for
accurate gesture classification and computing complexity
control.
The paper is arranged as follows. The section introduces
the related work about activity recognition and feature analysis.
The section outlines the prototype system of gesture
detection, and gives the gestures definition and data collection.
The section describes the feature analysis method in detail
including features generation process and features selection
algorithm based on stepwise regression. The section gives
the experiment and evaluation on feature generation and
selection. Conclusion is given in the last section.
II. RELATED WORKS
Two types of system are mainly used for activity
recognition. One is fixed device-based recognition system, and
the other is wearable device-based detection system.
Fixed device-based techniques have been applied widely
into various fields by different devices such as camera,
computer vision system, and so on [5] [6]. The kind of system
provides an application way of no burden to users. However,
some people do not like the way of being supervised, and some
private spaces are inconvenient to be set a camera like in
bathroom. Moreover, some issues have to be considered
including whether environment factor is fit for monitoring or
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
2 | P a g e
www.ijacsa.thesai.org

not such as surrounding light and blind corner of a camera,
image processing speed and delay, information loss of 3-D
object projecting to a 2-D image, and confusion of multiple
users in same background and so on [5].
With the development of micro-electrical technology, micro
wearable devices are penetrating into our life. They can be
attached on human body to obtain user information directly,
typically using RFID and sensor. RFID technology is attractive
for many applications since it can detect users situations by
simple ID and location information [7]. But it is difficult to
detect motion. Wearable sensors have shown their capability for
activity recognition. Some research set sensors in different
position of body to detect human daily activities. In [8] the
system sets five accelerometers on hip, wrist, arm, ankle, and
thigh for classifying 20 daily activities such as walking, sitting
& relaxing, brushing teeth, bicycling, etc, and got the
recognition accuracy ranging from 41% to 97% for different
activities. In [9] a large realistic data are collected from many
different sensors (accelerometers, physiological sensors,
environment sensors, etc.) to recognize 7 activities like lie, row,
walk, etc. with accuracy of 80% over. However, to our
knowledge, most of the researches seldom focus on the tiny
activity recognition like finger gesture.
Finger is one of the most flexible body parts. Most of works
in our daily lift are accomplished by it. Therefore, detection of
finger gestures not only helps to know user current behavior, but
also reflect user intension and carry on some operations. Data
glove, as an interactive device worn on the hand to sense
gestures, has been applied in the environment of virtual reality
[12]. However, the gestural interface required user to wear a
cumbersome device to connect with external. It is inflexible and
inconvenient for daily operation.
Moreover, most of above researches have no explained on
the process of features generation and selection while the
features are of paramount importance in any recognition works.
Based on different sensing device, some researches get features
directly from time-varying signal [10] [11] and with frequency
analysis [8] [9]. Some prefer to wavelet analysis to obtain both
spectral and temporal information [13] [14]. However, it is not
be illustrated why the features are necessary, if they can be
substituted on others, what will happen if adding or deleting one
of them.
In this paper, we introduce a wearable device in our previous
research named Magic Ring. It can be used to detect 12 kinds of
predefined one-stroke finger gestures based on a 3-axis
accelerometer [15]. A verity of gesture features are extracted for
classification evaluation. However, the process of feature
selection is a lack. In this paper, we focus on the feature analysis
method including feature generation, feature selection, and
feature evaluation taking the wearable sensor and its receiver as
a prototype of the final target devices.
III. PROTOTYPE INTRODUCTION AND DATA COLLECTION
A. Prototype Structure
The system is a ring shape sensing device based on a 3-trial
accelerometer, MMA7361L from Freescale Semiconductor,
Inc.
In its two sensitive scales of 1.5g and 1.8g, 1.5g is
adopted to detect all predefined gestures. Excepting for the
sensing unit, data processing unit is used for A/D conversion
and simple digital signal processing; and transmitting unit is for
acceleration data transmission and communication. The system
can be worn a finger with no much disturbing to daily activity as
shown in Fig. 1.

Figure 1. Sensing system on the finger
B. Gestures Definition
In the paper, the purpose of gesture recognition is to learn
user simple intension and further to apply the gestures into daily
life like controlling home appliances or calling for help.
Therefore the gestures should be easy to be done for reducing
physical load and easy to be understood and learned for
reducing conscious load. Combining the way of controlling
appliances and the characteristic of 3-axis accelerometer, 12
kinds of dynamic one-stroke gestures are designed. One-stroke
refers to dynamic gestures which are performed no more than
one degree of freedom in one direction. For example, pushing
a button, turning a knob, and pointing to a picture can be
regarded as one-stroke gestures.
The tri-axis accelerometer can detect the acceleration
change of three directions in space as shown in Fig.2.

Figure 2. Sensing direction of a tri-axis accelerometer
Considering the tri-axis characteristic of the accelerometer
and requirement of finger gestures, 12 kinds of one-stroke
finger gestures are defined as shown in Fig. 3.
The 12 kinds of gestures can be divided into 3 pairs of
gestures in X, Y and Z axis and 3 pairs of gestures in XY, YZ
and XZ axial plane. These gestures are named as Crook and
Unbend in X axis, Finger L-Shift and Finger R-Shift in Y axis,
Finger Up and Finger Down in Z axis, Wrist L-shift and Wrist
R-shift in XY plane, L-Rotate and R-Rotate in YZ plane, Wrist
Up and Wrist Down in XZ plane. The modes of motion for the
six pairs of gestures are shown in Fig. 3.




(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
3 | P a g e
www.ijacsa.thesai.org


Figure 3. Modes of motion for the six pairs of gestures
C. Data Collection
The system is attached on the middle phalanx of forefinger
since it is the most flexible in all fingers. Gestures data is
collected as sampling 50Hz. The digital signal is stored in a PC
for data analysis, features extraction and gestures classification.
20 students in the university (16 males and 4 females,
average age 25.87.8) volunteered for the experiment of data
collection under the supervision of a researcher. They are
required to perform the predefined 12 gestures by forefinger in a
natural and relax way. The gestures started in horizontal and
static state, ended with static state. Each gesture was repeated 5
times per people and 100 times for 20 people totally.
IV. FEATURES ANALYSIS
For a finger gesture, it can be expressed quantitatively as a
digital signal based on the sensing information. The features of
the signal can indicate the type of a gesture and is useful for
recognizing the gesture. A signal can be identified with various
features. Therefore the features analysis is vital for identifying
the signal. Roughly speaking, the more features are used, the
higher accuracy may be achieved, but higher complexity the
recognition is. However, not each feature can be used for
distinguishing the gesture from others, e.g. for the two signals
in Fig. 4, it is the acceleration change in X axis from two
different kinds of finger gestures. The feature of signal energy
or mean can distinguish them, but the peak value as a feature is
failed to recognize one from the other.
Moreover, the number of features may not have direct
relation with identification effects. Although it is different that
each feature and their combinations contribute to recognition
accuracy, that do not mean the more features are, the better
recognition accuracy is.

Figure 4. Two signals with same max value
Furthermore, the high dimensionality of features increases
computing cost for some recognition algorithm. Therefore,
analysis and selection of proper features of gestures to be
recognized has to be performed to get the optimal feature
vector/set for the balance between acceptable recognition rate
and computing complexity.
The feature analysis process is, first, to extract signals of
target gestures; second, based on the signals, generate candidate
features; and finally, selection of proper features from the
candidates to form feature set for recognition task. The paper,
taking an accelerometer as an example, gives the feature
analysis and selection procedure of finger gestures with an
accelerometer.
A. Extraction of Gesture signal
Extraction of gesture signal refers to get the section related
to the target gesture from successive signals, namely to detect
start and end point of the gesture. Since the gesture signal shows
a dynamic change trend from a static state to a dynamic activity,
and then back to static, therefore the short time energy (STE) of
signal is considered to distinguish the different states.
When STE in a sliding window is higher than a level, we
think a gesture start to be performed, until the STE becomes
lower than the level. We recorded the duration of each gesture
per subject. Results show it roughly ranges from 200 to 800ms.
Due to the sampling rate is 50Hz, the window size is
compromised within 10 samples with 50% overlap between two
continuous windows.
B. Feature Generation
Basically, feature can be divided into two types: features in
time domain and frequency domain. Features in time domain
show the signal characteristic varying with time. Typical
features are shown in Fig. 5.
Frequency features are used to capture the periodic nature of
a sensing signal for distinguishing some repetitive activities like
walking and running. The typical frequency-domain feature is
shown in Fig. 6.
Excepting for the features mentioned above, others can be
extracted according to different classification objects like
Euclidian distances, similarity, and so on.
Y
Z
X
Y
Z
X
Crook
Y
Z
X
Unbend
Y
Z
R-Rotate
X
X
Y
Z
L-Rotate
Y
Z
X
Wrist Down
Y
Z
X
Wrist Up
Y
Z
X
Wrist L-Shift
Y
Z
X
Wrist R-Shift
Y
Z
X
Finger L-Shift
Y
Z
X
Finger R-Shift
Finger Up
Y
Z
X
Finger Down
Y
Z
X
Y
Z
X
Y
Z
X
Crook
Y
Z
X
Crook
Y
Z
X
Unbend
Y
Z
X
Y
Z
X
Unbend
Y
Z
R-Rotate
X
Y
Z
R-Rotate
X
X
Y
Z
L-Rotate
X
Y
Z
L-Rotate
Y
Z
X
Wrist Down
Y
Z
X
Wrist Down
Y
Z
X
Wrist Up
Y
Z
X
Wrist Up
Y
Z
X
Wrist L-Shift
Y
Z
X
Wrist L-Shift
Y
Z
X
Wrist R-Shift
Y
Z
X
Wrist R-Shift
Y
Z
X
Finger L-Shift
Y
Z
X
Finger L-Shift
Y
Z
X
Finger R-Shift
Y
Z
X
Finger R-Shift
Finger Up
Y
Z
X
Finger Down
-100
-50
0
50
100
150
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49
gesture1 gesture2
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
4 | P a g e
www.ijacsa.thesai.org


Figure 5. Typical time-domain features


Figure 6. Typical frequency-domain features
All above mentioned features can be generated based on the
corresponding mathematical or statistical method. However, not
all features are necessary for a classification system. A direct
reason is whether a feature may bring good classification
significance. Even though both two features have good
classification capability, there is maybe little gain when they are
collected into a feature vector due to a high mutual correlation
[16]. Another reason is the computational complexity. The
number of features directly decides the dimensionality of
classifier parameter. Thus a feature vector as small as possible is
desired both in training process and in classifying process.
C. Feature Selection
The feature selection is very crucial, which helps us use as
less as possible of features to find out as much as possible of
classification information then to get the optimal recognition
performance. Here we try to find an optimal feature vector to
reach the balance between the acceptable recognition rate and
computational complexity. In practical application, a
satisfactory feature vector instead of an optimal vector.
In the paper, we adopt the algorithm of feature selection,
stepwise regression. It is a greedy algorithm that includes a
regression model in which candidate features are evaluated
automatically. Forward selection and backward elimination are
two main approaches to achieve the algorithm. The former
represents the procedure starting with no features in the model,
then trying to add one into the feature vector one by one until
them reaching a satisfactory significance. The latter is
contrary to the former, which including all candidate features in
the model, and deleting one that is no significant. Here, we used
the forward selection way.
Specifically, let

n
f f f F , , ,
2 1

be the candidate
feature set of using for classification design. The elements in the
set are descending order by the significance level. Let
i
f
S

denotes the significance level of feature
i
f
, for the feature set
F
,
n
f f f
S S S
2 1
. We select a classifier as the model,
then adding the feature one by one from
1
f
to test the
classification accuracy until reaching a satisfactory result.
V. EXPERIMENT AND EVALUATION
A. Features Generation
20 subjects completed the total 12 kinds of finger gestures
under the supervision of researcher. Before each one-stroke
finger gesture, it requires finger in a horizontal and static state.
When finishing a gesture, finger should maintain ending state
and stillness. In other words, the finger is dynamic just during a
gesture being performed. Therefore, it is possible to identify if a
gesture happening using a threshold-based approach.
Fig. 7 and Fig. 8 shows two finger gesture signals Finger
Left Shift and Finger Up, which is composed of three
channels and indicates the acceleration change in three axis.
From the extracted the gestures signals, various features can be
generated both in time domain and frequency domain as
described in last section. For example the mean and standard
deviation (sd) of each axis in Fig. 7 and Fig. 8 can be computed
as

1
0
1
0
) (
1
1
,
1
n
i
i
n
i
i
mean x
n
sd x
n
mean
where
x
denotes the sampling, and
n
denotes the number
of sampling in the window. Other features can also be achieved
by mathematical or statistical way.
Although any signal features can be regarded as a candidate,
in order to reduce the computing load, some obvious
insignificant features will be neglected. In our case, the finger
gestures are one-stroke type, which means each gesture is
aperiodic and instantaneous. Therefore, the features in
frequency domain are neglected.
The process of feature generation can be expressed as:
(1) Observe the signals of one gesture from different
subjects and try to describe it using some typical features. For
example, in Fig. 7, we may consider mean, energy, etc.
(2) Observe the signals of variety of kinds of gestures to find
out features with the capability of distinguishing with others
like sd of Z axis in above two figures.
(3) Abandon some insignificant features like amplitude of
each axis in our case, because even if a gesture is performed by
same person, amplitude of each time will be great different due
to the difference of performance speed.
(4) Collect the features to form a candidate feature set for
feature selection.
Time
Domain
Max
Min
Median
Difference
Mathematical
Features
Mean
Variance
Standard Variance
Energy
Entropy
Statistical
Features
Correlation
Zero-crossing Rate
Others
Time
Domain
Max
Min
Median
Difference
Mathematical
Features
Mean
Variance
Standard Variance
Energy
Entropy
Statistical
Features
Correlation
Zero-crossing Rate
Others
Frequency
Domain
DC Component
Spectral Peak
Spectral Centroid
Spectral bands
Spectral Energy
Spectral Entropy
Fourier
Transformation
Wavelet Coefficients
Wavelet
Transformation
Frequency
Domain
DC Component
Spectral Peak
Spectral Centroid
Spectral bands
Spectral Energy
Spectral Entropy
Fourier
Transformation
Wavelet Coefficients
Wavelet
Transformation
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
5 | P a g e
www.ijacsa.thesai.org


Figure 7. Acceleration curve of a finger gesture Finger Left Shift

Figure 8. Acceleration curve of a finger gesture Finger Up
In our case, time-domain features of each axis are calculated
including mean, standard deviation, energy, entropy, correlation
of any two axes, difference of peak and valley, and position of
peak and valley in the time axis, totally 8 kinds of features to
compose a candidate feature set. Each feature in the set consists
of three elements in X, Y, and Z axis, such as

1
f
{meanX,
meanY, meanZ}.
B. Features Selection
Features selection is to find a satisfactory feature subset
from the candidate feature set, so that to reach an optimal
classification accuracy and computing complexity control. It is
crucial since it decides the classification result directly. Forward
selection algorithm of stepwise regression is adopted to test
each feature and their combinations one by one. In the
algorithm, a model is required to evaluate the features. For
obtaining an objective evaluation results, here we select three
basic classifiers of machine learning, C4.5 decision tree (C4.5),
Nearest Neighbor (NN), and Nave Bayes (NB), as three test
models to calculate the classification accuracy of 12 kinds of
one-stroke finger gestures mentioned above, and the testing
average of three classifiers (Avg) is employed as the final
evaluation result.
First, each one in the candidate feature set is tested by three
models, the evaluation results are ranged in a descending order
as shown in Table 1.
TABLE 1. THE RESULTS OF SINGLE FEATURE EVALUATION
Features C4.5 NN NB Avg
1
f

meanX,
meanY,
meanZ
53.67% 56.42% 51.08% 53.72%
2
f

energyX,
energyY,
energyZ
49.83% 51.92% 49.17% 50.31%
3
f

sdX,
sdY,
sdZ
48.17% 53.25% 46.50% 49.31%
4
f

diffX,
diffY,
diffZ
38.92% 41.58% 34.92% 38.47%
5
f

col(X,Y),
col(X,Z),
col(Z,Y)
34.42% 36.58% 38.08% 36.36%
6
f

posValX,
posValY,
posValZ
26.58% 28.28% 28.58% 27.81%
7
f

entropyX,
entropyY,
entropyZ
26.83% 23.25% 29.33% 26.47%
8
f

posPeakX,
posPeakY,
posPeakZ
26.33% 27.50% 22.75% 25.53%
Second, selecting the optimal feature from the Table 1,
1
f
,
and combining with other features, the combinations will be
recomputed based on the models. The results are shown in
Table 2.
TABLE 2. THE EVALUATION RESULTS OF COMBINING TWO FEATURES
Features C4.5 NN NB Avg
2 1
f f

63.83% 71.42% 55.92% 63.72%
3 1
f f

68.25% 76.83% 66% 70.36%
4 1
f f

61.17% 72.50% 58.92% 64.20%
5 1
f f

59.92% 69.58% 59% 62.83%
6 1
f f

63.50% 68.58% 63.75% 65.28%
7 1
f f

59.17% 62.75% 58.53% 60.15%
8 1
f f

63.08% 68.67% 65.25% 65.67%
Third, repeating the above process of selection to get the
best one,
3 1
f f
, as a basic feature combination, and then
combining with rest features to reevaluate the significance of
combined features. From Table 3 to Table 7 show the evaluation
results with different combination. Each optimal combination in
each Table will be the basic of next combination.
TABLE 3. THE EVALUATION RESULTS OF COMBINING THREE FEATURES
Features C4.5 NN NB Avg
2 3 1
f f f

66.92% 77.83% 63% 69.25%
4 3 1
f f f

68.42% 79.25% 63.75% 70.47%
5 3 1
f f f

73.50% 81.92% 72.33% 75.92%
6 3 1
f f f

78.50% 84.50% 78.67% 80.56%
7 3 1
f f f

67.75% 70.25% 66.67% 68.22%
8 3 1
f f f

76.58% 86.08% 74.83% 79.16%
Finger Left Shift
-150
-100
-50
0
50
100
150
200
250
300
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49
X
Y
Z
Finger Up
-100
-50
0
50
100
150
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49
X
Y
Z
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
6 | P a g e
www.ijacsa.thesai.org

TABLE 4. THE EVALUATION RESULTS OF COMBINING FOUR FEATURES
Features C4.5 NN NB Avg
2 6 3 1
f f f f

76.92% 84.33% 76.42% 79.22%
4 6 3 1
f f f f

79.33% 85.42% 76.25% 80.33%
5 6 3 1
f f f f

80.50% 88.33% 81.75% 83.53%
7 6 3 1
f f f f

79.42% 85.58% 78.83% 81.28%
8 6 3 1
f f f f

81.08% 87.67% 82.25% 83.67%
TABLE 5. THE EVALUATION RESULTS OF COMBINING FIVE FEATURES
Features C4.5 NN NB Avg
2 8 6 3 1
f f f f f

80.50% 87.00% 79.45% 82.32%
4 8 6 3 1
f f f f f

81.17% 87.92% 79.08% 82.72%
5 8 6 3 1
f f f f f

81.42% 88.92% 84.50% 84.95%
7 8 6 3 1
f f f f f

82.92% 87.75% 82.92% 84.53%
TABLE 6. THE EVALUATION RESULTS OF COMBINING SIX FEATURES
Features C4.5 NN NB Avg
2 5 8 6 3 1
f f f f f f

82.83
%
85.50
%
82.08
%
83.47
%
4 5 8 6 3 1
f f f f f f

80.92
%
81.50
%
80.00
%
80.81
%
7 5 8 6 3 1
f f f f f f

82.00
%
81.83
%
83.33
%
82.39
%
TABLE 7. THE EVALUATION RESULTS OF COMBINING SEVEN FEATURES
Features C4.5 NN NB Avg
4 2 5 8 6 3 1
f f f f f f f

82.75
%
90.00
%
81.42
%
84.7
2%
7 2 5 8 6 3 1
f f f f f f f

83.08
%
90.28
%
84.25
%
85.8
7%
Finally, the average classification accuracy of combining all
features,
4 7 2 5 8 6 3 1
f f f f f f f f
, is 85.22%. Fig.9
shows the evaluation results under each kinds of combination.
It can be seen from the Fig. 9, with the increasing of the
number of features, the classification accuracy will increase.
However, when feature combination reaches to some extent,
such as after
6 3 1
f f f
, the accuracy has no obvious change.
That indicates not the more features are, the better classification
accuracy is. Even under certain circumstance, large number of
features will reduce the accuracy, which is the peaking
phenomenon occurring for larger features.
We can find from the Fig. 9, the best classification result is
about 85%. To improve the classification, other features are
considered to add into the candidate set. By observing the
acceleration signals of each kinds of finger gesture, we find, to a
gesture signal, the acceleration change in each axis shows great
difference. For example the gesture Finger Left Shift in Fig.
7, acceleration in Y axis shows intense change than X and Z
axis. While for Finger Up in Fig. 8, X and Z axes show
obvious fluctuation. If taking sd as the fluctuation level, the
comparison of sd between two axis is generated adding into the
candidate feature, which is sdX>sdY, sdY>sdZ, and sdX>sdZ.
we name the features as relative features. In addition, for
identifying the gestures with opposite direction, coming
sequence of peak and valley in single signal is adopted, which is
represented as posPeakX>posValX, posPeakY>posValY, and
posPeakZ>posValZ.

Figure 9. Recognition accuracy under different feature combinations
Using the relative features, candidate feature set is
regression evaluated again. Result show the recognition
accuracy reach to 90.08% with the feature subset {meanX,
meanY, meanZ, sdX, sdY, sdZ, sdX>sdY, sdY>sdZ, sdX>sdZ,
posPeakX>posValX, posPeakY>posValY,
posPeakZ>posValZ}. These relative features not only improve
the recognition accuracy, but also reduce the computing
complexity because of their alternative Boolean value. Besides,
the system robust can be increased since a relative relationship
can prevent classifier from acting of the initial state. The final
recognition matrix is shown in Table 8 based on Nearest
Neighbor classifier.
VI. CONCLUSION
In the recognition task, features are of paramount
importance. In the paper, we focus on the research of feature
analysis and selection, which includes how to generate the
candidate feature set based on the sensing information, how to
evaluate each feature and their combinations, and how to select
the optimal feature subset.
Feature generation process need to observe the activity
signal and initially select some features and neglect
insignificant one. Based on the candidate feature set, forward
selection algorithm of stepwise regression is adopted to evaluate
each feature and their combinations. The final combination with
good classification significance is selected as feature subset for
gesture recognition. Experiment result indicates the process of
feature analysis and selection is feasible to most of activity
recognition research. In the future, we plan to use our method to
evaluate other kinds of sensing device and activity data.
ACKNOWLEDGMENT
We would like to thank the subjects who help us with the
experiment in the research.
0.00%
20.00%
40.00%
60.00%
80.00%
100.00%
f
1
f
1
+
f
3
f
1
+
f
3
+
f
6
f
1
+
f
3
+
f
6
+
f
8
f
1
+
f
3
+
f
6
+
f
8
+
f
5
f
1
+
f
3
+
f
6
+
f
8
+
f
5
+
f
2
f
1
+
f
3
+
f
6
+
f
8
+
f
5
+
f
2
+
f
7
f
1
+
f
3
+
f
6
+
f
8
+
f
5
+
f
2
+
f
7
+
f
4
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
7 | P a g e
www.ijacsa.thesai.org

TABLE 8. RECOGNITION MATRIX FOR 12 KINDS OF FINGER GESTURES BASED ON THE SELECTED FEATURE SUBSET
Classified as a b c d e f g h i j k l 90.08%
a=Crook 86 0 3 8 1 1 0 0 0 0 1 0
86%
b=Unbend 0 100 0 0 0 0 0 0 0 0 0 0
100%
c=Finger Down 0 0 91 4 0 1 2 2 0 0 0 0
91%
d=Wrist Down 12 0 8 79 0 0 0 0 0 0 1 0
79%
e=L-Rotate 0 0 0 0 93 1 3 3 0 0 0 0
93%
f=Finger L-Shift 0 0 2 0 0 93 4 0 1 0 0 0
93%
g=Wrist L-Shift 0 0 1 0 2 11 86 0 0 0 0 0
86%
h=R-Rotate 1 0 0 0 3 0 1 89 0 6 0 0
89%
i=Finger R-Shift 0 0 0 0 1 1 0 0 96 2 0 0
96%
j=Wrist R-Shift 0 0 0 0 0 1 1 0 11 87 0 0
87%
k=Finger Up 0 1 0 0 0 0 0 0 3 0 94 2
94%
l=Wrist Up 0 1 0 0 0 0 0 0 0 0 12 87
87%

REFERENCES
[1] S. Consolvo, P. Roessler, et al. Technology for Care Networks of
Elders, IEEE Perv. Comp., vol. 43, no. 2, pp. 22-29, 2004.
[2] Stanford, V. Using Pervasive Computing To Deliver Elder Care, IEEE
Perv. Comp., vol. 1, no. 1, pp. 10-13, 2002.
[3] P. W. Jusczyk, and P. A. Luce, Speech Perception and Spoken Word
Recognition: Past and Present, Ear Hear, vol. 23, no. 1, pp. 2-40, Feb.
2002.
[4] A. Zafar, J. Overhage and C. McDonald, Continuous Speech
Recognition for Clinicians, J. Am. Med. Inform. Assoc., vol. 6, no. 3,
pp. 195-204, 1999.
[5] S. Mitra, and A. Tinku, Gesture Recognition: A Survey, IEEE Trans.
on System, Man, and Cybernetics Part C: Applications and Reviews,
vol. 37, no. 3, pp. 311-323, 2007.
[6] S. Helal, B. Winkler, et al. Enabling Location-aware Pervasive
Computing Applications for the Elderly, In Proc. of First IEEE
International Conference on Pervasive Computing and Communications
(PerCom03), 2003, Fort Worth, Texas, March, pp. 531-536.
[7] J. R. Smith, K. P. Fishkin, et al. RFID-based Techniques For
Human-activity Detection, Communication of ACM, vol. 48, no. 9, pp.
39-44, 2005.
[8] L. Bao, S. S. Intille, Activity recognition from user-annotated
acceleration data, in Pervasive 2004, Springer: Linz, Vienna. pp. 1-17.
[9] P. Juha, M. Ermes, et al., Activity Classification Using Realistic Data
from Wearable Sensors. IEEE Trans. on Information Technology in
Biomedicine, vol. 10, no. 1, pp. 119-128, 2006.
[10] U. Maurer, A. Rowe, A. Smailagic, and D. Siewiorek,, Location and
activity recognition using eWatch: A wearable sensor platform,
Ambient Intelligence in Everyday Life, Lecture Notes in Computer
Science, vol. 3864, pp. 86-102, 2006.
[11] X. Zhang, X. Chen, Y. Li, and etc. A Framework for Hand Gesture
Recognition Based on Accelerometer and EMG Sensors, IEEE Trans.
on Systems, Man and Cybernetics, Part A: Systems and Humans, Vol.
41, No. 6, pp. 1064-1076, 2011.
[12] T. G. Zimmerman, J. Lanier, C. Blanchard, and etc. A hand Gesture
Interface Divice, in Proc. of the SIGCHI/GI conference on Human
factors in computing systems and graphics interface, New York, USA,
1987.
[13] M. Sekine, T. Tamura, T. Togawa, and etc. Classification of waist-
acceleration signals in a continuous walking record, Med Eng Phys.
2000 May;22(4):285-91.
[14] N. Wang, E. Ambikairajah, N. H. Lovell, Accelerometry Based
Classification of Walking Patterns Using Time-frequency Analysis, in
Proc. of 29th Annual Conference of IEEE Engineering in Medicine and
Biology Sciety, pp. 4899-4902, Lyon, France, 2007.
[15] L. Jing, Y. Zhou, Z. Cheng, and J. Wang, A Recognition Method for
One-Stroke Finger Gestures Using a MEMS 3D Accelerometer, IEICE
Trans. on Information and Systems, E94.D(5), pp. 1062-1072, 2011.
[16] S. Theodoridis and K. Koutroumbas, Pattern Recogniton, 4st ed. U.K.
Elsevier, 2009, ch. 5, pp. 261-267.

AUTHORS PROFILE
Yinghui Zhou received her B.E. degree and M.E. degree in Computer Science
and Engineering from Jiamusi University and Yanshan University, China in
2001 and 2004 respectively. Now she is pursuing her ph.D. Degree in
University of Aizu. Her research is concerned with ubiquitous learning,
wearable computing and pattern recognition.
Lei Jing received his B.Eng. degree from Dalian
University of Technology, China, in 2000, M.Eng.
degree from the Yanshan University, China, in 2003, and
Ph.D from University of Aizu, Japan, in 2008. Currently
he is a special researcher at the University of Aizu. His
research interests include sensor networks, wearable
computing, and ubiquitous learning.

Junbo Wang received his B.E. in Electrical Engineering
and Automation and M.E. in Electric circuits & systems
in 2004 and 2007, from the YanShan University, China,
and received a Ph.D. degree in Computer Science at the
University of AIZU, Japan in 2011. Currently, he is a
visiting researcher in the University of Aizu. His current
research interests include IoT, ubiquitous computing,
context/situation awareness, and WSN.
Zixue Cheng received his B.Eng. degree from Northeast Institute of Heavy
Machinery in 1982, his Master degree and Ph.D degree from Tohoku
University, Japan in 1990 and 1993, respectively. Currently, he is a full
professor the School of Computer Science and Engineering, the University of
Aizu, Japan. His current interests include distributed algorithms, ubiquitous
learning, context-aware platforms, and functional safety for embedded
systems
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
8 | P a g e
www.ijacsa.thesai.org
A Framework for Improving the Performance of
Ontology Matching Techniques in Semantic Web

Kamel Hussein Shafaamri
Princess Sumaya University for Technology
Amman, Jordan


Jalal Omer Atoum
Princess Sumaya University for Technology
Amman, Jordan

AbstractOntology matching is the process of finding
correspondences between semantically related entities of
different ontologies. We need to apply this process to solve the
heterogeneity problems between different ontologies. Some
ontologies may contain thousands of entities which make the
ontology matching process very complex in terms of space and
time requirements. This paper presents a framework that
reduces the search space by removing entities (classes, properties)
that have less probability of being matched. In order to achieve
this goal we have introduced a matching strategy that uses multi
matching techniques specifically; string, structure, and linguistic
matching techniques. The results obtained from this framework
have indicated a good quality matching outcomes in a low time
requirement and a low search space in comparisons with other
matching frameworks. It saves from the search space from (43%
- 53%), and saves on the time requirement from (38% - 45%).

Keywords- Ontology matching; RDF statements;Semantic web;
Similarity Aggregation.
I. INTRODUCTION
In the current World Wide Web (WWW) computers and
machines have no idea about the semantic of the information
that are transferred through the web; the transferred
information are not machine understandable. The role of
computers is only to present the transferred information using
web browsers [19].
However, the next generation of the WWW is called a
Semantic Web. The role of the computers in the Semantic
Web is not only to present the information, but for the
computers to read and process the information in the
WebPages, and extract knowledge from this information.
The computer can understand the information in the
Semantic Web by using a data structure called Ontology.
Ontology provides a knowledge representation in a particular
domain; it defines concepts (classes and properties) in a given
domain, and shows the relationships between the defined
concepts [1], [19].
Different people may develop different ontologies that
describe a particular domain; this causes heterogeneity
problems between ontologies that describe the same domain.
In general different ontologies for a specific domain may use
different data formats, modeling languages and structures to
represent certain knowledge. The heterogeneity problem leads
to inability to get accurate search results in semantic web. For
example, some ontologies define a car as a car and another
ontologies define a car as an automobile, so if we write a
keyword car in a semantic web search engine then the result
of the search engine will be a list of all WebPages that are
based on ontologies that define car as a car, and this list will
not contain the WebPages that are based on ontologies that
define a car as automobile.
In general, ontology provides knowledge in a certain
domain to help the machines to make intelligent decisions.
Ontology consists of four components: concepts, object
properties, data properties and Individuals. In ontologies, we
can define RDF statements. An RDF statement consists of
three elements [1], [19]: Resource (Subject or Domain),
Object Property (Predicate or Property), Value (Object or
Range).
To solve the heterogeneity problem between ontologies,
we must apply a process called ontology matching process.
Ontology matching is the process of finding correspondences
between semantically related entities of different ontologies.
These correspondences stand for different relations such as
equivalence, more general, or disjointness, between ontologies
entities.
II. BACKGROUND
There are several types of matching techniques that are
used to find the correspondences entities between ontologies.
These techniques are string-based techniques, language-based
techniques and structural techniques [9].
Several ontology matching systems were developed to find
matched entities between different ontologies, such as Naive
Ontology Mapping (NOM) [7], PROMPT [15], Anchor-
PROMPT [14] and GLUE [5]. These previous systems take
two ontologies as inputs and use different ontology matching
techniques to test all entities of the first ontology with all
entities of the second ontology in order to find the matched
entities between the input ontologies [6]. So the search space
and time requirements of these previous systems are very
large.
To reduce the search space and time requirement of the
ontology matching process, we present in this paper an
ontology matching framework (system). This framework uses
a multi matching techniques specifically; string, structure, and
linguistic matching techniques, and depends on some
important features of ontologies; such as RDF statements and
class hierarchies.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
9 | P a g e
www.ijacsa.thesai.org
III. PROPOSED ONTOLOGY MATCHING FRAMEWORK
A. System Overview
As shown in Fig.1, the proposed matching system (PMS)
takes two ontologies as input, and determines the matched
entities (e.g. classes, object properties, data properties)
between these two ontologies. PMS compares entities of the
same type.
Specifically, it compares classes of ontology1 with classes
of ontology2, object properties of ontology1 with object
properties of ontology2, and data properties of ontology1 with
data properties of ontology2. PMS uses three types of
matching techniques, string and linguistic techniques in a
combined framework called "structure matching".
In order to reduce the search space of the matching
process, PMS is dependent on RDF statements and class
hierarchies which are the base components in ontologies. This
PMS matches RDF statements of ontology1 with RDF
statements of ontology2, and matches class hierarchies of
ontology1 with class hierarchies of ontology2.
The output of PMS is a set of matched entities with their
similarity values (confidence measures) each between 0 and 1.
All the entities that have similarity values greater than a pre-
defined threshold are considered to be correct matched
entities. And the output, also, includes the relationship types
between the matched entities (equivalence and subsumption).
The matching relationship is dependent on similarity values.
Our PMS focuses on one-to-one (1:1) and many-to-many (m:
m) match relationships, since they are the most commonly
used.
B. Structure of Ontology Matching Framework
The proposed system PMS has three stages: pre-
processing, matching process and post-processing.
1) Pre-processing
In this stage, PMS extracts the features of the input
ontologies. As shown in Fig.2, the system will read all RDF
statements and put them in a list that is called RDF statements
list. Each element in the RDF statements list will be one RDF
statement (subject, object properties, object).

Fig. 1. System Overview.

Fig. 2. Ontology Features Extraction.

Then, the system reads all leaves classes and their super-
classes and put them in a list that is called leaves-super list.
Each element in the leaves-super list will be an object that
contains a leaf class and its super classes. Finally, the system
reads all ontology classes and their data properties and put
them in data properties list. Each element in data properties
list will be an object that contains class and its data properties.
2) Matching Process
Structure matching consists of two stages; the first one
involves matching of RDF statements and the second one
involves matching of class hierarchies. Matching class
hierarchies also consists of two sub stages, matching of leaves-
super classes and matching of class-data properties.
a) Similarity Aggregation
In order to combine the similarity values of string matcher
and linguistic matcher, we use the following similarity
aggregation function [4]:


Where e1 is an entity of ontology1 and e2 is an entity of
ontology2, Simagg( ) is similarity combination of string
similarity Sims ( ) and linguistic similarity SimL ( ), Ws is a
string weight and WL is a linguistic weight. Ws, WL [0, 1]
and Ws + WL = 1. We used Ws = 0.3 and WL = 0.7. This
means that the linguistic matcher is more important than the
string matcher.
b) Matching RDF Statements
As mentioned earlier, an RDF statement has three
components (Subject, Object property, Object). PMS will
match every RDF statement in RDF-statements-list-A of
ontology1 with every RDF statement in RDF-statements-list-B
of ontology2 as illustrated in Fig.3.
PMS computes the similarity aggregation value of the
subject of an RDFstatement-A with the subject of an
RDFStatement-B, if their similarity aggregation value is
greater than the threshold value (matched subject) then it will
compute the similarity aggregation value of the object
property of an RDFStatement-A with the object property of an
RDFStatement-B. If their similarity aggregation value is
greater than the threshold value (matched object property) then
it will compute the similarity aggregation value of the object
of an RDFStatement-A with the object of an RDFStatement-B.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
10 | P a g e
www.ijacsa.thesai.org

Fig. 3. RDF Statements Matching Process.

Also, PMS adds the tested subjects and objects to the tested
classes list, and adds the tested object properties to the tested
object properties list. Then, PMS checks if the object, subject
classes, and object properties were tested before, by searching
the tested corresponding list. These operations are done to
prevent computing similarity aggregation values for classes
(subjects and objects) and object properties more than once.
Finally, the system will add matched subjects and objects
with their similarity aggregation values to the matched classes
list, and will add matched object properties with their
similarity aggregation values to the matched object properties
list.
Matching RDF statements using this scenario will reduce
the search space of the matching process for the following
reasons:
PMS ignores the object properties and the objects of RDF
statements if the subjects of RDF statements are not
matched.
PMS ignores the object of RDF statements if the subjects
or the objects properties of RDF statements are not
matched.
The outputs of the matching RDF statement process are the
following four lists:
1) Matched classes list.
2) Matched object properties list.
3) Tested classes list.
4) Tested Object Properties List.
c) Matching Class Hierarchies
In PMS, there are two types of class hierarchies, the first
one leaves-superClasses and the second one class-data
property. Each type has its own matching process.

Matching leaves-superClasses
In this stage, PMS reads four lists; two lists are outputted
from the previous process (Matching RDF Statements), which
are the matched classes list and the tested classes list. And the
other two lists are the leaves-superList-A of ontology1 and the
leaves-superList-B of ontology2. Then, PMS will apply the
matching process as illustrated in Fig.4.
PMS computes the similarity aggregation value of the leaf
in leaves-superList-A with the leaf of leaves-superList-B, if
their similarity aggregation value is greater than the threshold
value (matched leaves) then it will compute the similarity
aggregation value for all super classes of the matched leafs.
Also, PMS adds the tested leaves classes and super-classes
to tested classes list. Then, PMS checks if the leaves and
super-classes were tested before, by searching the tested
corresponding list. These operations are done to prevent
computing similarity aggregation values for classes more than
once.
Matching leaves-super classes using this scenario will
reduce the search space of the matching process because PMS
ignores the super-classes if the leaves are not matched.
The outputs of the matching leaves-superClasses process
are the following two lists:
1) Matched classes list.
2) Tested classes list.
Matching Class-data Property
In this stage, as illustrated in Fig.5, PMS reads three lists:
matched classes list, dataProperties list-A of ontology1 and
dataProperties list-B of ontology2. Data Properties list
contains objects. Each object presents a class and its data
properties. PMS check every pair of matched classes in the
matched classes list, if they have data properties.
Also, PMS adds the tested data properties to tested data
properties list. Then, PMS checks if the data properties were
tested before, by searching the tested corresponding list. These
operations are done to prevent computing similarity
aggregation values for data properties more than once.
Again, matching class-data property using this scenario
will reduce the search space of the matching process because
the system will ignore the data properties of non-matching
classes, and the system will try to match the data properties of
matched classes only.

Fig. 4. Leaves-SuperClasses Matching Process.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
11 | P a g e
www.ijacsa.thesai.org

Fig. 5. Data Properties Matching Process.

The outputs of the matching Class-data Property process
are the following two lists:
1) Matched data properties list.
2) Tested data properties list.
d) Final Outputs of the Matching Process
The final outputs of the PMS will be three lists as follows:
Matched classes list.
Matched object properties list.
Matched data properties list.
3) Post-Processing
In this stage, the PMS assigns matching relationship R to
the matched entities (Classes, Object properties, Data
properties) according to their similarity value [9].
If Simagg (e1, e2) = 1 then R is the equivalence
relation.
If Simagg (e1, e2) threshold then R is a
subsumption relation.
Where Simagg (e1, e2) is the similarity aggregation value
between matched entities e1 and e2.
4) String Matcher Implementation
For string matchers, the PMS uses Levenshtein distance
similarity measure [17] and soundex similarity measure [17],
[22] in combined manner as follows:


5) Linguistic Matcher Implementation
For linguistic matchers, the PMS uses Wordnet similarity
measures of Wu & Palmer similarity measure [18] and path
similarity measure [20] in combined manner as follows:

IV. PROPOSED FRAMEWORK EVALUATION
There are two types of evaluations that are used to evaluate
the PMS. The first evaluation is done by counting the number
of tested entities that were tested and by computing the time
requirement that are needed by the system to find the matched
entities. The second type of evaluation is based on compliance
measures to evaluate the quality of the matched results.
A. Compliance Measures
Following the work in [8], Compliance Measures are used
to evaluate the degree of compliance of the results of matching
algorithms. Compliance measures consist of three measures
Precision, Recall and F-measure. These measures are used to
evaluate the quality of the matching process and its results.
Precision and Recall are based on the comparison of an
expected result provided by a reference alignment and the
effective result of the evaluated system. Finally, F-measure
combines the measures of Precision and Recall as single
efficiency measure.
B. Traditional Matching System
For the purpose of evaluating our PMS we have developed
a matching system called (Traditional Matching System) TMS
that is based on the work of some existing ontology matching
systems such as NOM [7], PROMPT [15], Anchor-PROMPT
[14] and GLUE [5]. The TMS matches all classes of the first
ontology with all classes of the second ontology, and matches
all object properties of the first ontology with all object
properties of the second ontology, and matches all data
properties of the first ontology with all data properties of the
second ontology. The goal of developing the TMS is to
compare it with the PMS.
C. Conference Dataset
The conference dataset, shown in Table 1, has been
proposed in OAEI 2010 and it includes seven ontologies that
are dealing with conference organization. These ontologies
have been developed within OntoFarm project [21], and are
quite suitable for ontology matching task because of their
heterogeneous character. Every ontology in this dataset has a
number of classes, a number of object properties and a number
of data properties. The matching process will be done on each
pairs of these ontologies.
TABLE 1. CONFERENCE DATASET [21].
Ontology
Name
# of
classes
# of object
properties
# of data
properties
Cmt 29 49 10
Conference 59 46 18
ConfOf 38 13 23
Edas 103 30 20
Ekaw 73 33 0
iasted 140 38 3
sigkdd 49 17 11
D.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
12 | P a g e
www.ijacsa.thesai.org
E. TMS vs. PMS
The comparison of the PMS and the TMS was done on the
same computer system ((Intel (R) Core (TM) 2 Duo CPU,
2.4GHz, 3 GB RAM) and Windows 7)). We had applied the
matching process using the PMS and the TMS between each
pair of ontologies of the conference dataset at different
threshold values (0.5, 0.7, 0.85, and 1).
1) Search Space and Time Requirement Evaluation
Figures 6 and 7 present the average number of tested
entities and the average time requirement at different threshold
values (0.5, 0.7, 0.85, and 1) that are needed by both systems
TMS and PMS to match all the pairs of Conference dataset
ontologies.
We can notice from these two Figures, that the number of
tested entities and the time requirement that were needed to
find the matched entities in the TMS are larger than the
number of tested entities and time requirement that were
needed by the PMS, this is due to the fact that the TMS tests
more entities than the PMS.
Furthermore, we can notice that the number of tested
entities and the time requirement for the TMS remain the same
regardless of the threshold values. Whereas, in the PMS they
are inversely dependent on the threshold value.
2) Compliance Measures Evaluation
Figures 8, 9 and 10, present the average compliance
measures results (Precision, Recall and F-measure) of all the
matched pairs of Conference dataset ontologies, at different
threshold values (0.5, 0.7, 0.85, and 1) for both TMS and
PMS.


Fig. 6. Average Number of Tested Entities.


Fig. 7. Average Time Requirement.


Fig. 8. Average Precision Measure.


Fig. 9. Average Recall Measure.


Fig. 10. Average F-measure.

We can notice from these Figures the following:
At all thresholds values the Precision value of the PMS
is better than the Precision value of the TMS. Hence,
the PMS returns more accurate matching results than
the TMS.
At all thresholds values the Recall value of the TMS is
better than the Recall value of the PMS. This means
that the TMS returns more matched entities that are
existed in reference alignment R than the PMS;
At Threshold values (0.5, 0.7, and 0.85) the F-measure
value of the PMS is better than the F-measure of the
TMS. But at threshold value 1 the F-measure value of
the TMS is better than the F-measure of the PMS; this
is due to the fact that F-measure is dependent on
Recall and Precision values.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
13 | P a g e
www.ijacsa.thesai.org
F. Comparison with other Existing Matching Systems
We had made a comparison between the PMS and other
matching systems that participated in OAEI 2010 in terms of
Precision, Recall and F-measure. These systems are AgrMaker
[11], AROMA [3], ASMOV [12], CODI [13], Ef2Match [2],
Falcon [10] and GeRMeSMB [16]. This comparison is done at
threshold values of 0.5 and 0.7. Figures 11 and 12 show the
results of this comparison.
We can notice from these Figures the following:
The PMS at threshold 0.5 and at threshold 0.7 has the
lowest Precision value, because the PMS returns the
largest number of matched entities but a few of them
are existed in the reference alignment.
The PMS has the highest Recall value at threshold 0.5.
This means that the PMS returns the largest number of
matched entities that are existed in reference
alignment than the other matching systems.
The PMS has a good Recall value between the Recall
values of the other systems at threshold 0.7.
The PMS has a low F-measure value at threshold 0.5
and at threshold 0.7.


Fig. 11. Comparison with Other Matching Systems at Threshold = 0.5.



Fig. 12. Comparison with Other Matching Systems at Threshold = 0.7.
V. CONCLUSION
The main goal of this paper was to reduce the complexity
(search space and time requirement) of the ontology matching
process. This paper have introduced an ontology matching
framework that reduces the search space and the time
requirement of the matching process by removing entities
(classes, properties) that have less probability of being
matched. The proposed ontology matching framework had
used a multi matching techniques in order to find the
correspondences entities between ontologies.
The proposed matching framework saves (43% - 53%)
from the number of tested entities (search space). Furthermore,
the proposed matching framework saves on time requirement
of the matching process from (38% - 45%) in comparisons
with other matching frameworks.
The drawback of the proposed matching framework is that
it cant find all possible alignments entities between
ontologies, due to the fact that the PMS doesnt test all entities
of the matching ontologies. Hence, the PMS is recommended
to be used in matching large ontologies, since it will produce a
huge number of matching entities that could be enough for
web searching using semantic web.
REFERENCES
[1] G. Antoniou and F. Harmelen , A semantic web prime, (2
nd
ed).
London: Massachusetts Institute of Technology, 2008.
[2] W. Chua and J. Kim, Eff2Match results for OAEI 2010, proceedings
of the 5th international workshop on ontology matching, Shanghai,
China, 2010.
[3] J. David F. Guillet and H. Briand , Matching directories and OWL
ontologies with AROMA, proceedings of the 15th ACM international
conference on Information and knowledge management, ACM, New
York, USA, pp. 831-831, 2006.
[4] H. Do and E. Rahm, COMA - A system for flexible combination of
schema matching approaches, proceedings of the very larged data bases
conference (VLDB), Hong Kong, China, pp610-621, 2002.
[5] A.Doan, J. Madhavan, P. Domingos and A.Halevy, Ontology matching:
a machine learning approach In: S. Staab, and R. Studer (Ed),
handbook on ontologies in information system, (pp.385404), Berlin:
Springer-Verlag, 2003.
[6] M. Ehrig and S.Staab, QOM: quick ontology mapping, Proceedings of
the international semantic web conference (ISWC), Hiroshima, Japan,
pp. 683697, 2004.
[7] M. Ehrig and Y. Sure. Ontology mapping - an integrated approach,
proceedings of the 1st european semantic web symposium (ESWS), vol.
3053, pp.7691, Heraklion: Springer-Verlag, 2004.
[8] J. Euzenat, Semantic precision and recall for ontology alignment
evaluation, proceedings of international joint conference on artificial
intelligence (IJCAI), Hyderabad, India, pp. 248-253, 2007.
[9] J.Euzenat and P. Shvaiko, Ontology matching, (1
st
ed.). Berlin:
Springer-Verlag, 2007.
[10] W. Hu, J. Chen, G. Cheng and Y. Qu , ObjectCoref & Falcon-AO:
results for OAEI 2010, proceedings of the 5th international workshop
on ontology matching, Shanghai, China, 2010.
[11] C. Isabel. S. Cosmin, C. Michele, F. Caimi, M. Palmonari, F. Antonelli
and K. Ulas, Using agreementmaker to align ontologies for OAEI
2010, proceedings of the 5th international workshop on ontology
matching, Shanghai, China, 2010.
[12] Y. Jean-Mary P. Shironoshita and M. Kabuka, ASMOV results for
OAEI 2010, proceedings of the 5th international workshop on ontology
matching, Shanghai, China, 2010.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
14 | P a g e
www.ijacsa.thesai.org
[13] J. Noessner and M. Niepert, CODI: Combinatorial Optimization for
Data Integration results for OAEI 2010, proceedings of the 5th
international workshop on ontology matching, Shanghai, China, 2010.
[14] N. Noy and M. Musen, Anchor-PROMPT: using non-local context for
semantic matching ,proceedings of international joint conference on
artificial intelligence (IJCAI), Seattle, US, pp. 6370, 2001.
[15] N. Noy and M. Musen, The PROMPT suite: interactive tools for
ontology merging and mapping. international journal of human-
computer studies, vol. 59(6), pp. 9831024, 2003.
[16] C. Quix, A. Gal, T. Sagi and D. Kensche, An integrated matching
system: GeRoMeSuite and SMB Results for OAEI 2010, proceedings
of the 5th international workshop on ontology matching, Shanghai,
China, 2010.
[17] M. Taye, Ontology alignment mechanisms for improving web-based
searching, unpublished doctoral dissertation, De Montfort University,
United Kingdom, England, 2009.
[18] Z. Wu and M. Palmer, Verb semantics and lexical selection.
proceedings of 32nd annual meeting of the association for computational
linguistics (ACL), Las Cruces, US, pp. 133138, 1994.
[19] L. Yu, Introduction to the semantic web and semantic web services,
(1
st
ed.). Florida: Taylor & Francis Group, 2007.
[20] Pedersen, Ted. Wordnet similarity, from,
http://search.cpan.org/dist/WordNet
Similarity/lib/WordNet/Similarity/path.pm .
[21] Ontology alignment evaluation initiative,
http://oaei.ontologymatching.org/ .
[22] Soundex, from http://en.wikipedia.org/wiki/Soundex.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
15 | P a g e
www.ijacsa.thesai.org
Fingerprint Image Enhancement:
Segmentation to Thinning

Iwasokun Gabriel Babatunde
Department of Computer Science
Federal University of Technology,
Akure, Nigeria
Alese Boniface Kayode
Department of Computer Science
Federal University of Technology,

Akinyokun Oluwole Charles
Department of Computer Science
Federal University of Technology,
Akure, Nigeria
Olabode Olatubosun
Department of Computer Science
Federal University of Technology,
Akure, Nigeria


Abstract Fingerprint has remained a very vital index for
human recognition. In the field of security, series of Automatic
Fingerprint Identification Systems (AFIS) have been developed.
One of the indices for evaluating the contributions of these
systems to the enforcement of security is the degree with which
they appropriately verify or identify input fingerprints. This
degree is generally determined by the quality of the fingerprint
images and the efficiency of the algorithm. In this paper, some of
the sub-models of an existing mathematical algorithm for the
fingerprint image enhancement were modified to obtain new and
improved versions. The new versions consist of different
mathematical models for fingerprint image segmentation,
normalization, ridge orientation estimation, ridge frequency
estimation, Gabor filtering, binarization and thinning. The
implementation was carried out in an environment characterized
by Window Vista Home Basic operating system as platform and
Matrix Laboratory (MatLab) as frontend engine. Synthetic
images as well as real fingerprints obtained from the FVC2004
fingerprint database DB3 set A were used to test the adequacy of
the modified sub-models and the resulting algorithm. The results
show that the modified sub-models perform well with significant
improvement over the original versions. The results also show the
necessity of each level of the enhancement.

Keyword- AFIS; Pattern recognition; pattern matching;
fingerprint; minutiae; image enhancement.
I. INTRODUCTION
In the world today, fingerprint is one of the essential
variables used for enforcing security and maintaining a
reliable identification of any individual. Fingerprints are used
as variables of security during voting, examination, operation
of bank accounts among others. They are also used for
controlling access to highly secured places like offices,
equipment rooms, control centers and so on. The result of the
survey conducted by the International Biometric Group (IBG)
in 2004 on comparative analysis of fingerprint with other
biometrics is presented in Fig. 1. The result shows that a
substantial margin exists between the uses of fingerprint for
identification over other biometrics such as face, hand, iris,
voice, signature and middleware [1]. The following reasons
had been adduced to the wide use and acceptability of
fingerprints for enforcing or controlling security [1]-[4]:















a) Fingerprints have a wide variation since no
twopeople have identical prints.
b) There is high degree of consistency in fingerprints. A
person's fingerprints may change in scale but not in relative
appearance, which is not the case in other biometrics.
c) Fingerprints are left each time the finger contacts a
surface.
d) Availability of small and inexpensive fingerprint
capture devices
e) Availability of fast computing hardware
f) Availability of high recognition rate and speed
devices that meet the needs of many applications
g) The explosive growth of network and Internet
transactions
h) The heightened awareness of the need for ease-of-use
as an essential component of reliable security.
The main ingredients of any fingerprint used for
identification and security control are the features it possesses.
The features exhibit uniqueness defined by type, position and
orientation from fingerprint to fingerprint and they are
Figure 1: Comparative survey of fingerprint with other
biometrics
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
16 | P a g e
www.ijacsa.thesai.org
classified into global and local features [5]-[7]. Global features
are those characteristics of the fingerprint that could be seen
with the naked eye. They are the features that are
characterized by the attributes that capture the global spatial
relationships of a fingerprint. Global features include ridge
pattern, type, orientation, spatial frequency, curvature, position
and count. Others are type lines, core and delta areas.
The Local Features are also known as Minutia Points.
They are the tiny, unique characteristics of fingerprint ridges
that are used for positive identification. Local features contain
the information that is in a local area only and invariant with
respect to global transformation. It is possible for two or more
impressions of the same finger to have identical global
features but still differ because they have local features
(minutia points) that are different. In Fig. 2, ridge patterns (a)
and (b) are two different impressions of the same finger
(person). A local feature is read as bifurcation in (a) while it
appears as a ridge ending in (b).
II. FINGERPRINT IMAGE ENHANCEMENT
Reliable and sound verification of fingerprints in any AFIS
is always preceded with a proper detection and extraction of
its features. A fingerprint image is firstly enhanced before the
features contained in it could be detected or extracted. A well
enhanced image will provide a clear separation between the
valid and spurious features. Spurious features are those
minutiae points that are created due to noise or artifacts and
they are not actually part of the fingerprint. This paper adopts
with slight modifications, the algorithm implemented in [8]-
[9] for fingerprint image enhancement. The overview of the
algorithm is shown in Fig. 3. Its main steps include image
segmentation, local normalization, filtering and
binarization/thinning.











A Image Segmentation
There are two regions that describe any fingerprint image;
namely the foreground region and the background region. The
foreground regions are the regions containing the ridges and
valleys. As shown in Fig. 4, the ridges are the raised and dark
regions of a fingerprint image while the valleys are the low
and white regions between the ridges. The foreground regions
often referred to as the Region of Interest (RoI) is shown for
the image presented in Fig. 5. The background regions are
mostly the outside regions where the noises introduced into
the image during enrolment are mostly found. The essence of
segmentation is to reduce the burden associated with image
enhancement by ensuring that focus is only on the foreground
regions while the background regions are ignored.











The background regions possess very low grey-level
variance values while the foreground regions possess very
high grey-level variance values. A block processing approach
used in [8]-[9] is adopted in this research for obtaining the
grey-level variance values. The approach firstly divides the
image into blocks of size W x W and then the variance V(k)
for each of the pixels in block k is obtained from:
()

(( ) ())

()

()

( ) ()

I(i,j) and J(a,b) are the grey-level value for pixel i,j and
(a,b) respectively in block k.



(a) Bifurcation (b) Ridge ending

Figure 2: Different minutiae for different impressions
of the same finger
Figure 4: Ridges and valleys on a fingerprint image
Ridges
Valleys
Image Segmentation
Image Local Normalization
Image Filtering
Orientation
Estimation
Ridge
Frequency
Estimation
Gabor
Filtering
Image Binarization/ Thinning
Figure 3: The conceptual diagram of the fingerprint enhancement algorithm
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
17 | P a g e
www.ijacsa.thesai.org
A. Image Local Normaliuzation
Normalization is performed on the segmented fingerprint
image ridge structure so as to standardize the level of
variations in the image grey-level values. By normalization,
the grey-level values are made to fall within certain range that
is good enough for improved image contrast and brightness.
The first of the tasks of image normalization implemented in
[8]-[9] and adopted for this research is the division of the
segmented image into blocks of size S x S. The grey-level
value for each pixel is then compared with the average grey-
level value for the host block. For a pixel I(i,j) belonging to a
block of average grey-level value of M, the result of
comparison produced a normalized grey-level value N(i,j)
defined by the formula:
( )
{

(( ) )

( )

(( ) )

()

where an assumed value of M
0
is set for the desired mean
and an assumed value of V
0
is set for the desired variance.
B. Image Filtering
Normalized fingerprint image is filtered for enhancement
through removal of noise and other spurious features. Filtering
is also used for preserving the true ridge and valley structures.
The fingerprint image filtering structure adopted for this
research is in the following phases:
1) Orientation Estimation: Orientation estimation is the
first of the prerequisites for fingerprint image filtering. In
every image, the ridges form patterns that flow in different
directions. The orientation of a ridge at location x,y is the
direction of its flow over a range of pixels as shown in Fig. 6.
The Least Square Mean (LSM) fingerprint ridge
orientation estimation algorithm proposed and implemented in
[8]-[9] was slightly modified and used in this research. The
modified algorithm involves the following steps:
a) Firstly, blocks of size S x S were formed on the
normalized fingerprint image.
b) For each pixel, (p,q) in each block, the gradients

(p,q) and

(p,q) were computed as the gradient

magnitudes in the x and y directions, respectively.

(p,q) was computed using the horizontal Sobel operator

while

(p,q) was computed using the vertical Sobel operator.

c) The local orientation of a pixel in a fingerprint image
was computed by using its S x S neighborhood in [8]-[9]. This
was slightly modified in this research by dividing the image
into S x S blocks and the local orientation for each block
centered at pixel I(i,j) was then computed from:

( )

( )

( ) ()

( )

( )

( ) ()
( )

( )

( )
()
where (i, j) is the least square estimate of the local
orientation at the block centered at pixel (i, j).
d) The orientation image is then converted into a
continuous vector field defined by:

( ) (( )) ()

( ) (( )) ()

where

and

are the x and y components of

the vector field, respectively.
e) Gaussian smoothing is then performed on the vector
field as follows:

( ) ( )

( ) ()
Figure 6: The orientation of a ridge pixel in a fingerprint
Background
region
Foreground
region

Figure 5: A fingerprint image and its foreground and
background regions

0
0
0


Horizontal Sobel Operator

0 0 0



Vertical Sobel Operator

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
18 | P a g e
www.ijacsa.thesai.org

( ) ( )

( ) (0)
where G is a Gaussian low-pass filter of size

.
f) The orientation field O of the block centered at pixel
(i,j) is finally smoothed using the equation:

( )

( )

( )
()
2) Ridge Frequency Estimation: The second prerequisite
for fingerprint image filtering is the ridge frequency
estimation. In any fingerprint image, there is a local frequency
of the ridges that collectively form the ridge frequency image.
The ridge frequency is obtained from the extraction of the
ridge map from the image. The extraction of the ridge map
involves the following steps:
a) Compute the consistency level of the orientation field
obtained from the first prerequisite in the local neighborhood
of a pixel (p,q) with the following formula:

( )

|()()|

()
()
|( ) ( )| {
0
0
()

(( ) ( ) 0) 0 (14)

where W represents the local neighborhood around (p,q),
which is an n x n local window, (i,j) and (p,q) are local
ridge orientations at pixels (i,j) and (p,q) respectively.
b) If the consistency level is below a certain threshold
F
c
, then the local orientations in this region are re-estimated
at a lower image resolution level until the consistency is above
F
c
. After the orientation field is obtained, the following two
adaptive filters are applied to the image:

( )
{

()

()
0

(15)

( )
{

()

()
0

(16)
() (( )) (17)

(())
(18)
[|
(()

| |
(()

|] (19)

The two filters are capable of stressing under different
condition the local maximum grey level values along the
normal direction of the local ridge orientation. The normalized
image is first convolved with these two masks, h
t
(p,q, i, j) and
h
b
(p,q, i, j). If both the grey level values at pixel (p,q) of the
convolved images are larger than a certain threshold F
ridge
,
then pixel (p,q) is labeled as a ridge.
3) Orientation Estimation: Having obtained the
prerequisites, Gabor filtering is then used to improve or
enhance the fingerprint image to a finer structure. It involves
the removal of noise and artifacts. The general form of Gabor
filter is:
( ) {

]} () (0)

where f is the frequency of the cosine wave along the
direction from the x-axis, and
x
and
y
are the space
constants along x and y axes respectively. a= xsin + ycos
and b= xcos + ysin.
The values of the space constants
x
and
y
for the Gabor
filters were empirically determined as each is set to about half
the average inter-ridge distance in their respective direction.
x

and
y
are obtained from
x
= k
x
F and
y
= k
y
F respectively. F
is the ridge frequency estimate of the original image, and k
x

and k
y
are constant variables. The value of
x
determines the
degree of contrast enhancement between ridges and valleys
while the value of
y
determines the amount of smoothing
applied to the ridges along the local orientation.
C. Image Binarization/Thinning
The image obtained from the Gabor filtering stage is
binarized and thinned to make it more suitable for feature
extraction. The method of image binarization proposed in [10]
is employed. The Method sets the threshold (T) for making
each cluster in the image as tight as possible, thereby
minimizing their overlap. To determine the actual value of T,
the following operations are performed on set of presumed
threshold values:
a) The pixels are separated into two clusters according
to the threshold.
b) The mean of each cluster are determined.
c) The difference between the means is squared.
d) The product of the number of pixels in one cluster and
the number in the other is determined.
The success of these operations depends on the difference
between the means of the clusters. The optimal threshold is the
one that maximizes the between-class variance or, conversely,
the one that minimizes the within-class variance. The within-
class variance of each of the cluster is then calculated as the
weighted sum of the variances from:

()

()

()

()

() ()

() () ()

() () ()

()
()

()
()
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
19 | P a g e
www.ijacsa.thesai.org
p(i) is the pixel value at location i, N is the intensity level
and [0,N 1] is the range of intensity levels. The between-
class variance, which is the difference between the within-
class variance and the total variance of the combined
distribution, is then obtained from:

()

() ()

()

() (25)
(

() )

(26)
(

() )

(27)

where
2
is the combined variance,
B
(T) is the combine
mean for cluster T in the background threshold,
O
(T) is the
combine mean for cluster T in the foreground threshold and
is the combined mean for the two thresholds. The between-
class variance is simply the weighted variance of the cluster
means themselves around the overall mean. Substituting =
n
B
(T)
B
(T) +n
O
(T)
O
(T) into Equation 25, the result is:

()

()

()

()

()

()

Using the following simple recurrence relations, the
between-class variance is successfully updated by
manipulating each threshold T using a constant value p as
follows:

( )

() (29)

( )

() (30)

( )

()

()

()
(31)

( )

()

()

()
(32)

III. EXPERIMENTAL RESULTS
A slightly modified version of the fingerprint enhancement
algorithm used in [8]-[9] was implemented in this research by
using MATLAB Version 7.2 on the Windows Vista Home
Basic operating system. The experiments were performed on a
Pentium 4 1.87 GHz processor with 1024MB of RAM. The
purpose of the fingerprint enhancement experiments is to
analyze the performance of the modified algorithm under
different conditions of images as well as generate the metrics
that could serve the basis for the comparison of the results
from the research with results from related works. Two sets of
experiment were conducted for the performance analysis. The
first set of experiments was on synthetic images. The
orientation estimation, ridge frequency estimation and Gabor
filtering experiments all employed the circsine function [11] to
generate the synthetic images. The major arguments passed
into the circsine function include a number for the size of the
square image to be produced, a number for the wavelength in
pixels of the sine wave and an optional number specifying the
standard pattern of behaviour to use in calculating the radius
from the centre. This defaults to 2, resulting in a circular
pattern while large values give a square pattern. Where
necessary, the MATLAB imnoise function was also used to
generate noise and artifacts on synthetic images. The
arguments passed into the imnoise function include the image
on which noise is to be generated, noise type and noise level.
The second set of experiments was on the FVC2004
fingerprint database DB3 set A.
Fig. 7(a), 7(b) and 7(c) are synthetic images of size 200 x
200 and wavelength 8. They were obtained with the imnoise
function using the salt and pepper noise level of 0, 0.2 and 0.4
respectively. The results of the ridge orientation estimation
experiments on each of the three images are shown in Fig.
7(d), 7(e) and 7(f) respectively. These results show that for
noise levels of 0 and 0.20, the modified ridge orientation
algorithm effectively generated ridge orientation estimates that
are very close to the actual orientations. However, for image
with noise level of 0.4, the result shows a substantial number
of ridge orientation estimates that significantly differ from the
actual orientations. These results show that the performance
of the algorithm depends on the image noise level. When the
noise level on the image is within reasonable range, the
algorithm does well while it fails when the noise level rises
beyond the threshold which was found to be 0.29.
The efficiency of the ridge orientation estimation
algorithm was quantitatively measured by estimating the Mean
Square Error (MSE) which represents the difference between
the estimated and actual ridge orientation values in radians.
Mean square error estimation results for the synthetic image
shown in Fig. 7(a) under different conditions of noise for both
the pixel processing approach in [8] and the block processing
approach formulated for this research are shown in Table 1.
The increasing mean square errors in both cases indicate that
the accuracy of the two approaches decreases with increase in
the noise level. It is also revealed that the orientation estimate
is closer to the actual value for the block processing approach
at any noise level. With lower standard deviation of 0.0393 for
its MSE values, it is established that the block processing
approach performs better than the pixel processing approach
with MSE values of standard deviation of 0.1686. This higher
performance is attributed to the fact that in the block
processing approach, the degree of variation in the orientation
estimates for pixels in a block is reduced to zero as each pixel
assumes the orientation estimate for the center pixel of its host
block. This significantly increases the ability of the algorithm
to estimate the ridge orientation close to the actual value.
Fig. 7(g), 7(h) and 7(i) present the results of the ridge
frequency estimation experiments on the synthetic images of
different noise levels shown in Fig. 7(a), 7(b) and 7(c)
respectively. In the ridge frequency estimates shown in Fig.
7(g) and Fig. 7(h), there is uniformity between most of the
estimated frequency values for each 32 x 32 block as against
Fig. 7(i) which shows irregular patterns. Fig. 7(g), 7(h) and
7(i) produced MSE value of 0.0006, 0.0017 and 0.0077
respectively. The ridge frequency estimates and the increasing
MSE values reveal that the performance of the ridge frequency
estimation algorithm decreases with increase in the image
noise level. Generally, below the noise level of 0.30, it was
discovered that the algorithm produced relatively uniform
ridge frequency estimates as shown in Fig. 7(g) and 7(h).
When the noise level equals or exceeds 0.3, the orientation
estimation algorithm produced non uniform results as shown
in Fig. 7(i). Since the performance of the ridge frequency
estimation algorithm depends significantly on the performance
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
20 | P a g e
www.ijacsa.thesai.org
of the ridge orientation estimation algorithm, the failure of the
orientation estimation algorithm explains the failure of the
ridge frequency estimation algorithm for noise level above
0.29.
The MSE results of the implementation of the ridge
frequency estimation algorithm on the synthetic image shown
in Fig. 7(a) under different conditions of noise in [8] and the
current study are shown in Table 2. It is revealed that there is
increase in the MSE values as the noise level increases for the
two cases. This translates to diminishing performance due to
increasing margin between the actual and the estimated ridge
frequency values. It is also shown that the MSE value is
significantly lower in all cases for the current study than in [8].
With lower standard deviation of 0.0032 for its MSE values, it
is established that the implementation of the ridge frequency
estimation algorithm in the current study yields better results
than its implementation in [8] which yielded a standard
deviation of 2.8514 for its MSE values. This improvement is
attributed to the superior performance of the block processing
approach to ridge orientation estimation over the pixel
processing approach.
The performance of the Gabor filtering algorithm on a
zero, medium and high quality synthetic image of size 410 x
410 and wavelength 15 is presented in Fig. 8(d), 8(e) and 8(f)
respectively. Parameter values of k
x
=0.45 and k
y
=0.45 were
used to obtain these results. The results presented in Fig. 8(d)
and 8(e) reveal that with zero or medium level noise density,
the filter effectively removed the noise from the image and
enhanced it to a level that is comparable with the original
image. This effectiveness is partly due to the previous accurate
estimation of the ridge orientation and the ridge frequency for
zero or medium noise level images.
However, the experimental result presented in Fig. 8(f)
reveals that when the filter is applied to images with higher
noise level, the filter is unable to remove the noise effectively
as it produced a significant amount of spurious features. This
ineffectiveness is due to the inaccurate estimation of the ridge
orientation and the ridge frequency at higher noise levels.
When experiments were performed on real fingerprint images,
like in [8], the best results were obtained for image
segmentation with variance threshold of 100. This threshold
value provided the best segmentation results in terms of
differentiating between the foreground and the background
regions as shown in Fig. 9(b). The results of segmentation
using threshold value of 95 and 105 are presented in Fig. 9 (c)
and 9(d) respectively.

TABLE 1: COMPARISON OF MEAN SQUARE ERROR FOR
RIDGE ORIENTATION ESTIMATES
Noise level Mean Square Error
Raymond [8] Current study
0.00 0.0003 0.0000
0.05 0.0009 0.0006
0.10 0.0032 0.0008
0.15 0.0102 0.0019
0.20 0.0246 0.0015
0.25 0.0691 0.0026
0.30 0.1722 0.0405
0.35 0.2330 0.0521
0.40 0.3041 0.0661
0.45 0.4124 0.0803
0.50 0.4262 0.1085
TABLE 2: COMPARISON OF MEAN SQUARE ERROR FOR
RIDGE FREQUENCY ESTIMATES
Noise level Mean Square Error
Raymond[8] Current Study
0.00 0.0100 0.0006
0.05 0.0211 0.0009
0.10 0.0465 0.0011
0.15 0.0820 0.0012
0.20 0.1702 0.0017
0.25 1.1149 0.0033
0.30 2.0229 0.0041
0.35 4.1149 0.0060
0.40 5.8543 0.0077
0.45 6.1098 0.0080
0.50 7.2616 0.0090
(a) Original
image
Fig. 9: Results of segmentation with different threshold
(b) Segmentation
with threshold of
100
(c) Segmentation
with threshold of
95
(d) Segmentation
with threshold of
105
(a) Image with 0
noise level
(b) Image with 0.2
noise level
(c) Image with 0.4
noise level
(d) Orientation
estimate of (a)
(e) Orientation
estimate of (b)
(f) Orientation
estimate of (c)
(g) Ridge freq.
estimate of (a)
(h) Ridge freq.
estimate of (b)
(i) Ridge freq.
estimate of (c)
Figure 7: Orientation and ridge frequency estimates
for synthetic images of different noise levels
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
21 | P a g e
www.ijacsa.thesai.org
These results show inappropriate segmentation due to
inaccurate variance thresholds. With lower threshold of 95,
some of the background regions have been segmented to the
foreground while some foreground regions are also segmented
to the background under higher threshold of 105.
The result of the normalization experiment on the
fingerprint images shown in Fig. 9(a) is presented in Fig.
10(a). The desire mean of zero and variance of one used in [8]
were adopted and used to normalize the ridges in the images.
During normalization, all positions are evenly shifted along
the horizontal axis, which makes the structure of the ridges
and valleys to become well and suitably positioned.
The histogram plots of the original and the normalized
images are shown in Fig. 10(b) and 10(c) respectively. The
histogram plot of the original image shows that all the
intensity values of the ridges show irregular frequency values
and also fall within the right hand side of the 0255 scale, with
no pixels in the left hand side. This leads to an image with a
very low contrast.
The histogram plot of the normalized image shows that the
range of intensity values for the ridges has been adjusted
between 0-1 scale such that there is a more evenly and
balanced distribution between the dark and light pixels and
that the ridge frequencies fall within close values. The
normalized image histogram plots also show that the
normalization process does distribute evenly the shape of the
original image. The positions of the values are evenly shifted
along the x-axis, which means the structure of the ridges and
valleys are now well and suitably positioned. This shifted and
improved positioning lead to images with a very high contrast
shown in Fig. 10(a).












The orientation fields for the real fingerprint images were
obtained around their singular points since they are prominent
features used by any AFIS for fingerprint classification and
matching. Good quality images are shown in Fig. 11(a), 11(b)
and 11(c). Their orientation estimates are shown in Fig. 11(e),
11(f) and 11(g) respectively.
At the singular points, the orientation field is discontinuous
and unlike the normal ridge flow pattern, the ridge orientation
varies significantly. From these results, it is observed that
there exists no deviation between the actual fingerprint ridge
orientation and the estimated orientation of the vectors. In both
cases, the algorithm produces accurate estimates of the
orientation vectors such that they flow smoothly and
consistently with the direction of the ridge structures in the
images. In the superimposed version of images in Fig. 11(e),
11(f), 11(g) and 11(h), the contrast of the original image is
lowered in each case. This was done to improve the visibility
of the orientation vectors against the background.
The ridge orientation estimate for poor quality image
shown in Fig. 11(d) is presented in Fig. 11(h). The estimate
indicates a fairly smooth orientation field in some well-defined
regions while it gives misleading results in areas of very poor
quality as evident in the top-left and bottom regions of the
estimate. The orientation estimates resulting from the pixel
processing approach in [8] and the block processing approach
of the current study for the image shown in Fig. 12(a) are
presented in Fig. 12(b) and 12(c) respectively.
Visual inspection of these results reveals that the two
methods did well in the ridge orientation estimation. However,
the orientation is observed to be closer to the actual orientation
in block processing than pixel processing in some regions
especially the core areas represented by the inserted circles.
The reason adduced to this is that assigning equal orientation
estimate for pixels in a block rather than maintaining different
values is better and able to take the estimates closer to their
actual values.
(a)Norma-
lized image
Figure. 10: Normalized image and the histogram plots
(b) Histogram of
original image
(b) Histogram of
normalised image
(a) 0 noise level
image image

(b) 0.2 noise level
image
(c) 0.4 noise level
image
(d) Filtered (a) (e) Filtered (b)
(f) Filtered (c)
Figure 8: Results of applying a Gabor filter on synthetic
images of different noise levels.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
22 | P a g e
www.ijacsa.thesai.org












Visual inspection of the results for the ridge frequency
estimates for real fingerprint images shown in Fig. 11(a),
11(b), 11(c) and 11(d) are shown in Fig. 13(a), 13(b), 13(c)
and 13(d) respectively. The mean ridge frequency (MRF),
which is the average of the image ridge frequencies, is also
presented for each image. It is noted that the MRF values
differ for all the images. This difference is attributed to the
fact that fingerprints exhibit variation in their average ridge
frequency characteristics and contrast levels. The intensities of
frequency differ for blocks or regions within same image.
Some blocks or regions exhibit high contrast while others
exhibit low contrast. Based on these, the synthetic images are
more appropriate for the evaluation of the accuracy of the
ridge frequency estimation algorithm.
Fig. 14 reveals the extent to which the filtering algorithm
was able to remove noise from the images shown in Fig. 11
for different values of k
x
and k
y
. The results shown in Fig.
14(a), 14(b), 14(c) and 14(d) were obtained using parameter
values of k
x
= 0.45 and k
y
= 0.45. With these values, it is
shown that the contrast level between ridges and valleys for
each of the images is neither too high nor too low. Appropriate
level of smoothing is also applied to the ridges along the local
orientation.
With lower parameter values of k
x
= 0.40 and k
y
= 0.40, it
is shown in Fig. 14(e), 14(f), 14(g) and 14(h) that the contrast
level between ridges and valleys is too low and this explains
why there is a number of dark regions. The degree of
smoothing is also poor as there is a good number of
overlapping ridges in the filtered images. When k
x
=0.5 and
k
y
=0.5 were used, the results of the filtering experiments
shown in Fig. 14(i), 14(j), 14(k) and 14(l) reveal that there is
no significant difference from the results obtained for k
x
=0.45
and k
y
=0.45 except that some regions described by circles
appear to be excessively smoothened.
It is therefore stated that based on the modified algorithm,
parameter values of k
x
=0.45 and k
y
=0.45 are most appropriate
for image filtering as against k
x
=0.50 and k
y
=0.50 proposed in
[8]-[9]. This reduction in parameter values is due to the better
performance of the block processing approach in the
orientation and ridge frequency estimations as attested to by
the results in Table 1 and Table 2. Results presented in Fig. 14
show that the filtering algorithm is able to smoothen to a fine
level with appropriate parameter values for good quality
fingerprint images.
When the quality degrades like the one shown in Fig.
11(d), the performance of the algorithm diminishes as it
produces images with inappropriately filtered regions as
shown in Fig. 14(d). This is also corroborated with reasons
adduced to the values presented in Table 1 and Table 2.




(a) Original
Image
(b) Orientation
estimate obtained in
[8]
(c) Orientation
estimate for current
study
Fig. 12: Orientation estimates for pixel and block
processing approaches
(a) Ridge frequency estimate
of Fig. 11(a) with mean ridge
frequency of 0.0765
(c)Ridge frequency estimate
of Fig. 11(b) with mean ridge
frequency of 0.0726
(d) Ridge frequency estimate
of Fig. 11(c) with mean ridge
frequency of 0.0821
(e) Ridge frequency estimate
of Fig. 11(d) with mean
ridge frequency of
0.0637
Fig. 13: Ridge frequency estimate of selected images
Regions of false results
(a)Good quality
whorl image
(e) Orientation
estimate of (a)
Figure 11: Fingerprint images and their orientation estimates
Singular Points
(b)Good quality
left loop image
(c)Good quality
right loop image
(d)Poor quality image
(f) Orientation
estimate of (b)
(g) Orientation
estimate of (c)
(h) Orientation
estimate of (d)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
23 | P a g e
www.ijacsa.thesai.org
The results of the binarization experiments for the images
shown in Fig. 11 are presented in Fig. 15. Visual inspection of
the results shows that the binarization algorithm perfectly
separated the ridges (black pixels) from the valleys (white
pixels).
To obtain these results, the grey-level value of each pixel
in the filtered image is examined and, if the value is greater
than the threshold value 1, then the pixel value is set to a
binary value one; otherwise, it is set to zero. The threshold
value successfully made each cluster as tight as possible and
also eliminated all overlaps. The threshold value of 1 was
taken after a careful selection from a series of within and
between class variance values ranging from 0 to 1 that
optimally supported the maximum separation of the ridges
from the valleys. The clear separation of the ridges from the
valleys verifies the correctness of the algorithm as proposed in
[9] and implemented in this research.
The results of the thinning experiment on each of the
images shown in Fig. 11 are presented in Fig. 16(a-d). The
MATLABs bwmorph operation using the thin option was
used to generate the thinned images. These results show that
the ridge thickness in each of the images has been reduced to
its smallest form or skeleton (one pixel wide). It is also shown
that the connectivity of the ridge structures is well preserved.
Fig. 16(e-h) shows the results of performing binarization
experiments on the raw images without the enhancement
stages. In contrast to Fig. 15(a-d), the binary images in Fig.
16(e-h) are not well connected and contain significant amount
of noise and corrupted elements. Consequently, when thinning
is applied to these binary images, the results in Fig. 16(i-l)
show that the accurate extraction of minutiae would not be
possible due to the large number of spurious features
produced. Thus, it is shown that employing the image
enhancement stages prior to thinning is effective for accurate
and speedy extraction of minutiae. Consequently, when
thinning is applied to these binary images, the results in Fig.
16(i-l) show that the accurate extraction of minutiae would not
be possible due to the large number of spurious features
produced.
Thus, it is shown that employing the image enhancement
stages prior to thinning is effective for accurate and speedy
extraction of minutiae.





































IV. CONCLUSION
This paper discusses the results of the modification and
verification of the fingerprint enhancement algorithm
developed and implemented in [8]-[9]. Some stages of the
algorithm were slightly modified for improved performance.
For instance, block processing approach was introduced into
the orientation estimation algorithm in place of the pixel
processing approach.
(a)
(b)
(c)
(d)
(e)
(f) (g)
(h)
(i)
(j) (k)
(l)
Thinned images without the enhancement stages
Binarized images without the enhancement stages
Thinned images with the enhancement stages
Fig. 16: Thinned and Binarized images
Filtered images for k
x
= 0.45 and k
y
= 0.45

(a) (b)
(c)
(d)
Filtered images for k
x
= 0.40 and k
y
= 0.40

(e)
(f) (g) (h)
Filtered images for k
x
= 0.50 and k
y
= 0.50
(i) (j) (k) (l)

Fig. 14: Filtered images with different values for kx and ky
Fig. 15: Results of Binarization for images shown in Fig. 11(a),
11(b), 11(c) and 11(d)

(a) (b) (c) (d)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
24 | P a g e
www.ijacsa.thesai.org
While the pixel processing approach subjects each pixel in
the image to orientation estimation, the block processing
approach firstly divides the image into S x S blocks and
obtains the orientation estimate for the center pixel. This
resulted in higher performance as attested to by the Tables of
MSE for ridge orientation and frequency estimates. Parameter
values of k
x
=0.45 and k
y
=0.45 were found to perform well in
the image filtering experiments as against k
x
=0.5 and k
y
=0.5.
The results of the experiments conducted for image
segmentation, normalization, ridge orientation estimation,
ridge frequency estimation, Gabor filtering, binarization and
thinning on synthetic and real fingerprint images reveal that
with free or minimal noise level, the algorithms perform well.
Improved performance is specifically noticed for the modified
ridge orientation estimation algorithm. It is also established
that each stage of the enhancement process is important for
obtaining a perfectly enhanced image that is acceptable and
presentable to the features extraction stage. The results
obtained from the final stage of thinning show that the
connectivity of the image ridge structure has been preserved
and improved at each stage.
REFERENCES
[1] C. Roberts: Biometrics:
(http://www.ccip.govt.nz/newsroom/informoationnotes/2005/biometrics.
pdf, Accessed: July, 2009
[2] C. Michael and E. Imwinkelried: Defence practice tips, a cautionary
note about fingerprint analysis and reliance on digital technology,
Public Defense Backup Centre Report., 2006
[3] M. J. Palmiotto: Criminal Investigation. Chicago: Nelson Hall, 1994
[4] Salter D.: Fingerprint An Emerging Technology, Engineering
Technology, New Mexico State University, 2006
[5] J. Tsai-Yang, and V. Govindaraju: A minutia-based partial fingerprint
recognition system, Center for Unified Biometrics and Sensors,
University at Buffalo, State University of New York, Amherst, NY USA
14228, 2006
[6] O. C Akinyokun, C. O. Angaye and G. B. Iwasokun: A Framework for
Fingerprint Forensic; Proceeding of the First International Conference
on Software Engineering and Intelligent System, organized and
sponsored by School of Science and Technology, Covenant University,
Ota, Nigeria, 2010, pages 183-200.
[7] O. C. Akinyokun and E. O. Adegbeyeni: Scientific Evaluation of the
Process of Scanning and Forensic Analysis of Fingerprints on Ballot
Papers, Proceedings of Allied Academies International Conference on
Legal, Ethics and Regulatory Measures in New Orleans, April 8 - 10,
2009; Vol. 13; No. 1; pages 1 - 13.
[8] L. Hong, Y. Wan and A. Jain: Fingerprint image enhancement:
Algorithm and performance evaluation; Pattern Recognition and Image
Processing Laboratory, Department of Computer Science, Michigan
State University, 2006, pp1-30
[9] R. Thai: Fingerprint Image Enhancement and Minutiae Extraction, PhD
Thesis Submitted to School of Computer Science and Software
Engineering, 2003, University of Western Australia.
[10] X. Xu : Image Binarization using Otsu Method. Proceedings of NLPR-
PAL Group CASIA Conference, 2009, pp345-349
[11] P. Kovesi: MATLAB functions for computer vision and image
analysis, School of Computer Science and Software Engineering,
University of Western Australia,
http:/www.cs.uwa.edu.au/~pk/Research/MatlabFns/Index.html,
Accessed: 20

February 2010.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
25 | P a g e
www.ijacsa.thesai.org
Data Warehouse Requirements Analysis Framework:
Business-Object Based Approach

Anirban Sarkar
Department of Computer Applications
National Institute of Technology, Durgapur
West Bengal, India


AbstractDetailed requirements analysis plays a key role
towards the design of successful Data Warehouse (DW) system.
The requirements analysis specifications are used as the prime
input for the construction of conceptual level multidimensional
data model. This paper has proposed a Business Object based
requirements analysis framework for DW system which is
supported with abstraction mechanism and reuse capability. It
also facilitate the stepwise mapping of requirements descriptions
into high level design components of graph semantic based
conceptual level object oriented multidimensional data model.
The proposed framework starts with the identification of the
analytical requirements using business process driven approach
and finally refine the requirements in further detail to map into
the conceptual level DW design model using either Demand-
driven of Mixed-driven approach for DW requirements analysis.
Keywords- Requirements analysis; Business objects; Conceptual
Model; Graph Data Model; Data Warehouses.
I. INTRODUCTION
Complex, online and multidimensional analysis of data is
done by fetching just-in-time information from subjective,
integrated, consolidated, nonvolatile, historical collection of
data. Data Warehouse (DW) and On Line Analytical
Processing (OLAP) in conjunction with multidimensional
database are typically used for such analysis. DW facilitates
data navigation, analysis, and business oriented visualization of
data using multidimensional cube and OLAP query processing.
DW systems are used mainly by decision makers to analyze the
status and the development of an organization [1], based on
large amounts of data integrated from heterogeneous sources
into a multidimensional data model. As in other information
systems, requirement analysis plays a key role within DW
system development to reduce the risk of failure.
DW projects are long-term projects, so it is very difficult to
anticipate future requirements for the decision-making process
in the scenario of evolving business processes over time [2].
Further, information requirements for DW systems are difficult
to specify at the early stage because decision processes are
flexibly structured and shared across the different sectors of the
large organization. In this scenario, requirement analysis for
DW project must support reuse of domain level abstractions
and step wise refinement mechanisms strictly for mapping the
requirements in high level design. Hence, for the DW project,
requirements analysis must start focusing with early
requirements analysis and further it should move on detailed
requirements analysis. Analyzing early requirements will
significantly decrease the possibility of misunderstanding the
users requests and consequently reduce the risk of failure for
the DW project.
Several studies [2, 3, 4] indicate that improper analysis of
user requirements or avoidance of requirements analysis phase
leads to unsuccessful design of DW System. Further they
recommended that, there must be a dedicated phase of
requirement analysis for the purpose of DW system design.
The primary deliverables of the DW requirements analysis
phase is conceptual level multidimensional data model [5]. In
this course, the schemata of the available operational data
sources are also compared with the user driven information
requirements. Indeed, the approaches to DW design are usually
classified in three categories:
a) Supply-driven / Data-driven: In these approaches [6,
7, 8], the DW design starts from a detailed analysis of the
operational data sources. In order to determine the structure of
conceptual multidimensional data model the user requirements
have less impact in this approach.
b) Demand-driven / Requirements-driven: These
approaches [2, 9, 10] start from determining the information
requirements of business users or stakeholders. The problem of
mapping these requirements onto the available data sources is
faced only at the late design phase of DW system. This
approach is the only alternative whenever a deep analysis of
data sources is unfeasible, or data sources reside on legacy
systems whose inspection and normalization is not
recommendable. However, in this case, conceptual level
multidimensional data model design can be directly based on
mapping of requirements.
c) Mixed-Driven (Supply/Demand): In these approaches
[9, 11, 12], requirements analysis and source inspections are
carried out in parallel. The Supply-driven and mixed
framework is recommended when source schemata are well
known, and their size and complexity are substantial.
Majority of these approaches mainly have focused on
requirement analysis and does not provide any definite
guidelines to move from requirements model to high level
conceptual design. In [9] goal oriented approaches has been
described to support both demand driven and mixed framework
for DW system requirements analysis. It comprehensively
supports early requirement analysis phase for DW system and
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
26 | P a g e
www.ijacsa.thesai.org
provides a mechanism to map the requirements into the
conceptual model. But author has not clearly explained about
the detailed requirements analysis for DW system.
In this paper, a requirement analysis framework for DW
system has been proposed based on concept of Common
Business Objects [13]. A business object (BO) captures
information about a real world (business) concept, operations
on that concept, constraints on those operations, and
relationships between those elements and other business
concepts. The set of related BOs express an abstract view of the
businesss real world. The advantage of using this concept is
that, the set of BOs can be reusable in the context of business
domain and it ensures the resultant system will be scalable,
reliable, secure and interoperable [13].
In the perspective of requirements analysis for DW system,
our proposed framework consist of three phases, namely, (i)
Early Requirements Analysis Phase, (ii) Detailed Requirements
Analysis Phase and (iii) Mapping Phase. The early
requirements analysis phase allows for modeling and analyzing
the contextual setting of the business domain, in which the DW
will operate. In detailed requirements analysis phase, the early
requirements specifications are refined with the structural,
functional and nonfunctional features of the domain that is
relevant to the participants and their role related to the
analytical tasks. Moreover, with the aim of the user centric
requirements analysis this phase provides a guideline to
identification of materialized views [19] for the target DW
system. The refinement process is largely influenced by the
concepts of Feature Oriented Domain Analysis (FODA) [16].
The mapping phase used to map the DW requirement
specifications to the conceptual design model and can starts
just after the early requirements analysis phase. The framework
supports both demand driven and mixed demand / supply
driven requirements analysis approaches for DW system. Also
it is supported with abstraction mechanism; reuse capability
and mapping facility for requirements descriptions into high
level design components.
In the context, the preliminary version of this work has
been published in [17]. A BO based requirements analysis
framework for generic large scale information system also has
been published in [18]. Moreover, the proposed framework has
used the graph semantic based conceptual level object oriented
multidimensional data model proposed in [14, 15] for its
mapping phase. In this context, Graph Object Oriented
Multidimensional Data (GOOMD) model provides a novel
graph based semantic and simple but powerful algebra to
conceptualize the multidimensional data visualization and
operational model for OLAP, based on object oriented
paradigm.
II. GOOMD MODEL WITH EXAMPLE
In this section, we will summarize the basic concepts of
GOOMD model [14, 15]. The GOOMD model is the core of
the comprehensive object oriented model of a DW containing
all the details that are necessary to specify a data cube, a
description of the dimensions, the classification hierarchies, a
description fact and measures.
The GOOMD model allows the entire multidimensional
database to be viewed as a Graph (V, E) in layered
organization. At the lowest layer, each vertex represents an
occurrence of an attribute or measure, e.g. product name, day,
customer city etc. A set of vertices semantically related is
grouped together to construct an Elementary Semantic Group
(ESG). So an ESG is a set of all possible instances for a
particular attribute or measure. On next, several related ESGs
are group together to form a Contextual Semantic Group (CSG)
the constructs to represent any context of business analysis. A
set of vertices of any CSG those determine the other vertices of
the CSG, is called Determinant Vertices of said CSG. The most
inner layer of CSG is the construct of highest level of
granularity of fact in Multidimensional database formation.
This layered structure may be further organized by
combination of two or more CSGs as well as ESGs to represent
next upper level layers and to achieve further lower level
granularity of contextual data. From the topmost layer the
entire database appears to be a graph with CSGs as vertices and
edges between CSGs as the association amongst them.
Dimensional Semantic Group (DSG) is a type of CSG to
represent a dimension member, which is an encapsulation of
one or more ESGs along with extension and / or composition of
one or more constituent DSGs. Fact Semantic Group (FSG) is a
type of CSG to represent a fact, which is an inheritance of all
related DSGs and a set of ESG defined on measures. Two types
of edges has been used in GOOMD model, (i) directed edges
from DSGs to FSG or constituent DSG to determinant vertex
of parent DSG to represent the one to many associations
and (ii) undirected edges between constituent ESGs and
determinant ESGs to represent the association within the
members of any CSG.
Since, In order to materialize the cube, one must ascribe
values to various measures along all dimensions and can be
created from FSG. The cube will also obey a functional
constraint f: D1 D2 Dp MI. Where any Di is a
member of all related top level DSGs and MI is instances of set
of measures M. For schema containing multiple FSGs with
shared DSGs, the DSG set {D1, D2, Dp} are the common
set of DSGs for all FSGs of the schema.
Let consider an example, based on Sales Application with
Sales Amount as measure and with four dimensions
Customer, Model, Time and Location. Model, Time and
Location dimensions have upper level hierarchies as Product,
QTR and Region respectively. Then in the notation of
GOOMD model, there will be four DSGs DSales =
{DCustomer, DModel, DLocation, DTime} with hierarchies.
Each DSG will be comprised of either a set of ESGs EX
ESales or a combined set of ESGs and DSGs. As described
above, the lower layer DSG will be comprised of ESGs only.
The Product DSG DProduct is comprised of only ESGs like
EP_ID, EP_NAME and EP_DESC and will be represented as
the inner layer of the graph. In the example DModel DSG is an
extension of DProduct DSG as well as encapsulation of EM_ID
and EM_NAME. The DProduct and DModel DSG graphically
can be represented as Figure 1. The FSG for the database can
be described as FSales = {DET(DCustomer), DET(DModel),
DET(DLocation), DET(DTime), EAMOUNT}. Where
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
27 | P a g e
www.ijacsa.thesai.org
EAMOUNT is the ESGs defined on the measure. The schema
from the topmost layer has shown in Figure 2.
GOOMD model also provides algebra of OLAP operators
those will operate on different semantic groups. The dSelect ()
operator is an atomic operator and will extract vertices from
any CSG depending on some predicate P. The Retrieve ()
operator extracts vertices from any Cube using some constraint
over one or more dimensions or measures. The Retrieve
operator is helpful to realize slice and dice operation of OLAP.
The Aggregation ( and +) operators perform aggregation on
Cube data based on the relational aggregation function like
SUM, AVG, MAX etc. on one or more dimensions.
Aggregation operators are helpful to realize the roll-up and drill
down operations of OLAP. GOOMD model also provides the
definitions of the operators like Union (), Intersection (),
Difference (), Cartesian Product () and Join (||), which are
operated on any CSG or Cube.
III. PROPOSED REQUIREMENTS ANALYSIS FRAMEWORK
The core of the proposed Requirements Analysis
Framework for DW system is Business Object Model. A
business object (BO) is a conceptual object that has been
specified for the purpose of directly describing and
representing a business concept with a well-defined boundary
and identity. A BO encapsulates identity, domain specific
features and behavioral features.
The domain specific features are comprised of structural,
functional and nonfunctional features of the domain of interest.
The behavior is given by the activities that the business object
is capable of performing to fulfill its purpose including the
collaboration with other BOs. The primary issue is capturing
business semantics (for processes, events and stakeholders)
having a common idea or concept that is usable by different
parts of a business and by different participants of that
business. In the context of DW system, Business Object
Modeling is an abstraction technique that consists of
identifying the set of concepts and their contexts that belong to
some business domain. Such model is comprised of set of BOs
to characterize the set of processes and activities of that
specific business.
A. Components of Proposed Framework
The part of business object model taxonomies [13], relevant
to DW system requirements analysis are as follows,
a) Business Object (BO): This is an abstraction that
describes a concept of interest in the business itself and capable
of being specialized through inheritance mechanism. A BO is
the super type of all objects that represent business concepts
either entities or activities involved in such specific business
domain.
b) Entity BO: This is a specialize form of BO that
describes basic business concepts those are engaged in the
conduct of business processes. For example stakeholders,
products, locations etc. Two entity BOs based on some specific
role can collaborate with each other in the context of some
business process.
c) Process BO: This is a specialized form of BO that
describes a business process or workflow and is comprised of a
specified collection of Entity BOs, a pattern of interactions and
business events. For example, order fulfillment, procurement,
payment etc. Interactions represent and implement activities.
The entity BO instances are the actors with specific role and
subjects of action. A Process BO can be further refined as
collection of related sub process BOs.
d) Event BO: This is a specialized form of BO that
describes a business event, which may trigger and result from
interactions between entity BOs in the context of a process BO.
For example, inventory threshold, account overdrawn, end of
fiscal year etc. Event BOs are used to capture the business
constraints on the interactions.
Besides the above described taxonomies, DW context
requires some new taxonomy to be introduced and are as
follows,
Figure 1. Lowest Level DSG, (b) Higher Level DSG
Figure 2. (a) Schema for Sales Application in GOOMD Model
(b) SALES FSG construct after inheritance
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
28 | P a g e
www.ijacsa.thesai.org
a) Measure Attribute: This is an attribute that describes a
quantitative aspect of business process that is relevant for
decision making.
b) Fact BO: This is a specialized form of Process BO to
capture the concept of subject of analysis specific to some
business process. It encapsulates the Measure Attributes along
with other features.
c) Dimension BO: This is a specialized form of Entity
BO to capture the concept of parameters over which the Fact
BO will be analyzed using other Entity BO having relevant
role. Any behavioral features will not be included in the
Dimension BO. Further a specialized Dimension BO can be
formed from generalized one in the proposed framework.
d) Relations: several relationship types have been used
in the proposed framework like, (i) Encapsulation, (ii)
Inheritance, (iii) Collaboration and (iv) Interaction.
The graphical notations for above taxonomies have been
summarized in Table 1.
B. Early Requirements Analysis Phase
The focus of early requirements analysis phase is to analyze
the target business domain. It includes identification along with
high level of abstraction of the business processes, different
stakeholders, the interactions and collaborations between them,
and the events relevant for analytical task specific to the
business domain. This phase is important to understand how a
business is perceived by its stakeholders mainly decision
makers a as set of related business processes and events. This
phase is also the basis to identify the possible analytical
requirements of decision maker or other stakeholders from the
business process driven approach. The phase consists of five
steps and are as follows,
a) Identification of Process and Entity Level Business
Objects: In the first step the relevant business processes with
their context, possible stakeholders related to those business
processes and interactions between them are represented with
high level of abstraction. The business processes and
stakeholders can be represented in the proposed framework
using Process BOs and Entity BOs respectively. Interactions
can be represented between Process BOs and Entity BOs.
Interactions represent and implement the activities. In between
one Process BO and one Entity BO, more than one activity
may exist and this will result in more than one interaction
respectively. The interactions between the set of Entity BOs
and Process BOs of some specific business can be expressed
using business domain level Interaction Diagram. The
Interaction Diagram of each interested Process BO can be
achieved from business domain level Interaction Diagram.
For example in a Retail Organization, entire business is
comprised of several business processes like, (i) Procurements
for procuring the products for sale, (ii) Sales to handle the
customer orders and to sale the products as per order, and (iii)
Accounting to handle the bills, order payment, salaries etc.
Several stakeholders may be involved with these business
processes, e.g. Sales Manager and Customer may interact with
the Sales and Accounting processes with activities like Place
Order, Receive Product, Payment, Raise Bill, Bill adjustment
etc. Now, the business process will be mapped into the Process
BOs and the Stakeholders will be mapped into the Entity BOs.
Several other Entity BOs can be involved with each Process
BO without having any specific interactions, but they are use to
supply important information on the interactions between the
business processes and related stakeholders. For example Time,
Location, Product, Transaction Type etc. The BOs and
interactions relevant to the example can be captured using
Interaction Diagram as shown in Figure 3. In the diagram
Interactions are labeled with the activities and with direction of
initiator and receiver of the interaction. The important point to
note that, in this step all Process BOs and related Entity BOs
are the representation of high level abstraction for the set of
involved business process.

TABLE I. GRAPHICAL NOTATIONS FOR BO-BASED REQUIREMENTS
ANALYSIS FRAMEWORK
Taxonomy Graphical Notation
Entity BO


Process BO


Event BO


Measure Attribute

Fact BO


Dimension BO


Encapsulation

Inheritance

Collaboration

Interaction Relation

Figure 3. Business Domain Level Interaction Diagram





(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
29 | P a g e
www.ijacsa.thesai.org
Now let, the decision makers requirements include analyze
of the Sales of the product. So the Sales Process BO is one of
the interesting candidates of target DW system. Two
stakeholders may be involved with the process and represented
as Entity BOs namely, Sales Manager and Customer. The
Interaction Diagram related to Sales Process BO has been
shown in Figure 4.

b) Identification of Collaborations: In this step the
collaborations are identified between Entity BOs in the context
of some interested Process BO relevant for the purpose of
analytical processing. Collaborations are occurred based on
some specific roles played by participant Entity BOs. In
software engineering term role can be defined as separation of
concerns i.e. separation of behavioral characteristic of some
Entity BO. The purpose of collaboration is to fulfill some set of
activities concerned to the specific Process BO, with which the
participant Entity BOs can interact. The roles and
collaborations in the context of some Process BO can be
identified from Interaction Diagram. This will lead towards
developing the Collaboration Diagram in the context of some
specific Process BO. Further, for a specific collaboration one
can model the activities performed in the context of some
Process BO, which will result the Collaboration Interaction
Network.
In the above example Sales Manager can play the role as
Supplier to fulfill the activities like Handle Order and Deliver
Product and also can play the role as Payee for the purpose of
activity like Payment Received. Similarly two roles like Client
and Payer can be defined for Entity BO Customer. The
Collaboration Diagram and Collaboration Interaction Networks
are represented in Figure 5.
c) Identification of Measure Attributes and Fact
Business Objects: Measure attributes are the quantitative
aspects of some Process BO and are relevant for performing
analytical tasks on that specific BO. Measure attributes related
to each interested Process BO are identified in this step. This
step is also the basis of Fact BO construction. Each Process BO
relevant to analytical task can be specialized into a Fact BO by
encapsulating the related measure attributes.
For example in the running example of Retail Organization
for the Sales Process BO two possible measure attributes may
be Quantity and Amount of sales. A Fact BO for Sales can be
described from the Process BO Sales by encapsulating the
measure attributes Quantity and Amount.
d) Identification of Dimension Business Objects: In this
step Entity BOs other than representing stakeholders are
identified in the context of some specific Process BO. The set
of Entity BOs identified in this step, in general captured the
parameters over which the measure attributes related to the
specific Process BO are dependent and can be analyzed. Both
the set of Entity BOs that captured the stakeholders and other
Entity BOs that provides the important information to the
Interactions are used to describe the abstraction of Dimension
BO in the context of some Fact BO. The set of related Entity
BOs which are used to describe the dimension abstraction can
be identified from the business domain level interaction
diagram of the first step of this phase.
In the Retail Organization example in the context of Sales
Fact BO, the possible Dimension BOs can be Time, Location,
Figure 4. Interaction Diagram for Sales Process BO
Figure 5. Collaboration Diagram, (b) Collaboration Interaction
Network for Supplier / Client Collaboration, (c) Collaboration
Interaction Network for Payee / Payer Collaboration
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
30 | P a g e
www.ijacsa.thesai.org
Product and Payment Type including Sales Manager and
Customer.
e) Identification of Event Business Objects: In this step
the Event BOs are identified in the context of some specific
Process BO and related Entity BOs. Identification of Event
BOs is important to understand the rationale for recording the
process BO instances in the DW system for the analytical task.
Entity BOs are also important to realize how the specific
Process BO will react on some interaction. In the system level
view, an Event BO can be realized as a trigger which may fire
as a result of specific set of interactions between related Entity
BOs and is important in the creation of the DW system.
For example, in the Sales Process BO the possible Event
BOs may be Order Processed and Full Payment Received. This
step provides the basis of mapping the instances in DW system.
C. Detailed Requirements Analysis Phase
The focus of detailed requirements analysis phase is to
refine the Fact BOs and related Dimension BOs identified in
early requirements analysis phase to satisfy the analytical
requirements of decision makers or other stakeholders. The
Fact BOs and Dimension BOs are refined by adding the
domain specific features like structural, functional and
nonfunctional feature to capture the activities, interactions,
collaborations and other stakeholders requirements. Further
this phase provide the guidelines to select materialized views
for the target DW system. This phase is also capable to
implement the Supply-Driven part of the framework by further
refining the identified Fact BOs and Dimension BOs with the
comparison with existing operational schemas. The refined
Fact BOs and Dimension BOs produced in this case will be
more realistic and appropriate for mixed analysis approach of
DW system requirements analysis. But it is important to note
that the detailed analysis of source operational schemas must
be available a prior for this purpose. This phase consists of
four steps and are as follows,
a) Refinement of Fact Business Objects: In this step the
Process BOs identified in early requirements phase are refined
through two tasks. In the first task, Process BOs are refined as
possible collection of related sub Process BOs. The measure
attributes identified for top level Process BOs are placed with
appropriate sub Process BOs. This task may be iterative in
nature to achieve further details sub Process BOs. The
Interaction Diagram concerned to the specific Process BO
acquired from earlier phase is extended accordingly.
Further the refined Fact BOs can be realized from the
refined Process BOs to satisfy stakeholders expectations. For
example, the Sales Process BO may be thought as the
composition of two sub-Process BOs Order Processing and
Payment Processing. Measure attributes Quantity and Amount
can be associated with Order Processing and Payment
Processing Process BOs respectively. Henceforth two possible
Fact BOs can be realized based on the sub-Process BOs. In
fact, each of them encapsulates at least one measure attribute. It
will result Extended Interaction Diagram for the Sales Process
BO and has been represented in Figure 6.
In the second task, each refined Process BO can be further
refined by adding the domain level feature like structural,
functional and nonfunctional along with the constraint
specifications. The features are the attributes of the system
which directly affect the stakeholders. Structural features
describe the object level properties of some BO and may also
include constraint specification over some features. Functional
features describe the operational capabilities of some Process
BO. In the context of DW system requirements analysis,
functional features of any Process BO include the set of
possible analytical operations. Thus functional features are the
basis of identification of the possible set of OLAP operations
those are required to perform by the decision makers, over the
related Fact BO. The prime focus of Nonfunctional features of
some Process BO is related with the expected QoS
requirements of stakeholders concerned to the specific BO.
This may include the features for Security, Performance,
Usability etc., specific to some Process BO. Nonfunctional
features may be optional for certain Process BO. The features
of some Process BO can be represented using Feature Tree
Diagram.
Also, Process BO may have specialized Process BOs.
Henceforth a Process BO can contain both basic and derived
features. Besides representing the features of some business
process as per the stakeholders perspective, a Feature Tree is
capable to represent the logical grouping (AND, OR and
EXOR grouping) of the features to satisfy the decision makers
or other stakeholders need. Several features of the specific
Process BO are the basis of the set of attributes that can be
analyzed along with the parameter over which those can be
analyzed, in the context of the related Fact BO. Henceforth, the
feature tree exhibits the set of features using which the related
Fact BO can be associated with concerned Dimension BOs for
the formation of conceptual level multidimensional schema for
the DW system. The concerned high level Dimension BOs are
already identified in the first phase of the framework. Further
the feature tree may be accompanied with set of composition
rules to express the existing semantics between the subset of
features. An example of partial Feature tree for the Process BO,
Payment Processing has been shown in Figure 7.
b) Refinement of Dimension Business Objects: In this
step, the set of Dimension BOs identified in the early
Figure 6. (a) Refined Sales Process BO, (b) Extended Interaction
Diagram.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
31 | P a g e
www.ijacsa.thesai.org
requirements analysis phase are refined. The refinement
process involves two tasks. Firstly, the different levels of
granularity required for the multidimensional information are
identified for some specific business processes. Basis on that,
the set of related Dimension BOs need to refine starting from
the lowest level granularity to the highest one. The Dimension
BO with highest level granularity will exhibit lowest level
abstraction in Dimension BO instances. The Dimension BOs
with highest level granularity in the context of some Fact BO
already have been identified in early requirements analysis
phase. This task facilitates the vertical refinement of
Dimension BOs and focus on the creation of dimension
hierarchy in the context of some fact for the DW system. As an
example, the possible vertical refinement of Product
Dimension BO in the context of Sales Fact BO has been
represented in Figure 8.
In the second task, the refined set of Dimension BOs (each
of all hierarchies) are further refined by adding the domain
level features in the same way as it has been done in case of
Process BOs in the previous step of this phase. It will result the
Feature Tree diagram for each Dimension BO. This task
facilitates the horizontal refinement of Dimension BOs. The
only difference here is, the accompanied composition rules will
express the semantics between the subset of features from the
specific Dimension BO or in the context of the concerned Fact
BO. This will facilitate to realize the different constraint
requirements for the DW system.
c) Compare with Operational Schemas: This step is to
implement the Supply-Driven part to the framework. The
refined Fact BOs and Dimension BOs resulted from the last
two steps can further be modified by comparing with the
existing operational schemas of the organization. To perform
this step a prior knowledge of detailed analysis of operational
schemas are required. The set of Fact BOs and Dimension BOs
can be refined or modified or also can be filtered by navigating
the source operational schemas. Several literatures suggest
algorithmic approaches mostly based on the stakeholders
requirements using navigation through the path of many-to-one
associations from the attributes of facts.
In fact the steps can be performed by drawing the
dependencies between each of Fact BOs and related
Dimension BOs, and related operational source schemas. The
feature tree diagram of each BO on next can be modified as
per the related operational schemas features. The steps are as
followed,
1) Draw the dependencies between the Measure Attributes
of concerned Fact BO and the Appropriate Attribute of existing
source Operational Schemas. It is important to note that the
Measure Attribute names may be decided by the decision
makers or designers and henceforth attribute names may not
match.
2) From each of such Operational Schemas identified in
previous step, start navigation using the available foreign keys
to other associated Operational Schemas. The each of
associated set of Operational Schemas is required to map as
Dimensional BO. Dependency can be drawn between the
Dimension BOs achieved from the previous steps of the
framework to the Operational Schemas achieved on navigation.
3) If dependencies have not been drawn from some
Dimension BOs then those BOs are purely demanded by the
decision makers and can be labeled as Demanded Dimension
BO. On other hand, where dependencies have drawn from
some Dimension BOs, those are labeled as Supplied
Dimension BO.
4) For each Supplied Dimension BO, the attributes of the
appropriate operational schemas are compared with Feature
Tree of the specific Dimension BO. On comparison structural
features in the feature tree can be refined or modified as per the
attribute details of the related operational schemas.
5) For each operational schema, identified in step (ii) but
related Dimension BO does not exist, are required to map as
New Dimension BO. The feature tree of new Dimension BO
can be drawn from the attributes descriptions of the related
operational schemas.
6) The feature tree of each Fact BO can be modified further
from the available Dimension BOs and newly achieved
Dimension BOs from the step (v).
Figure 8. Vertical Refinement of Dimension BO Product
Figure 7. Partial Feature Tree for the Process BO Payment Processing

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
32 | P a g e
www.ijacsa.thesai.org
d) Identification of Materialized Views: For user
oriented DW requirements engineering, it is also important to
analyze that how user will efficiently interact with the DW
system to perform the necessary analysis activities.
Materialized views are the central issue for the usability of the
DW system. DW data are organized multi dimensionally to
support OLAP. A DW can be seen as a set of materialized
views defined over the source relations. Those views are
frequently evaluated by the user queries. The materialized
views need to be updated when the source relations change.
During the DW analysis and design, the initial materialized
view need to be selected to make the user's interactions simple
and efficient in terms of accomplishing user analysis
objectives. In the proposed requirements engineering
framework, the domain boundary has been drawn through
identifications of Fact BOs, Dimension BOs, Actor BOs, and
interactions between them in the first phase and which have
been further refined in this phase. The list of analysis activities
may be performed by Actor BOs based on their roles and also
Event BOs have been identified in the same phase. Moreover,
the feature tree concept explores the constraint requirements
for the interest of domain. Based on those identifications, the
different materialized views can be identified in this step. In
this step the materialized views are used to represent
semantically in the context of some Fact BO and in terms of
actor along with their roles, analysis activities those may be
performed, events those may be occurred, related Dimension
BOs involved and the related constraints. Related to one Fact
BO, there may exist several materialized views to minimize
the views level dependency and to meet the analytical
evaluation requirements of the stake holders. Semantically, a
materialized view will be represented using View Template.
The Interface Template will contain the information of View
name, identification, analysis objectives, target Fact BO, Actor
BO, roles, related activities, related Dimension BOs to realize
the source relations, related Event BOs and related constraints.
Any view template is reusable and modifiable through
iterative process to accommodate the updatable materialized
view.
In case of Retail Organization, to interact with the Sales
Fact BO, one example view template related to Customer has
been shown in Figure 9.
IV. REQUIREMENTS ANALYSIS TO CONCEPTUAL MODEL
This section will describe the Mapping Phase of our
framework. In this phase the DW requirements specifications
achieved from the early requirements analysis phase and
consequently from the detailed requirements analysis phase are
mapped into conceptual level multidimensional design model.
In this phase, we will map the requirements specifications in
the constructs and concepts of GOOMD model described in
Section 2. But there is no binding on mapping the proposed
requirements framework to any multidimensional conceptual
model. This phase can starts just after the early requirements
analysis phase and consists of two levels, namely, Early
Mapping and Detailed Mapping.
a) Early Mapping: Early mapping can be done just after
the early requirements analysis phase for DW system. The
early requirements analysis phase is business process driven
approach and is used to identify the set of Fact BOs and
related Dimension BOs relevant to analytical requirements of
decision makers and other stakeholders. Each identified Fact
BO and encapsulated Measure Attributes will be mapped as
Fact Semantic Group (FSG) and Elementary Semantic Group
(ESG) for measures respectively as described in GOOMD
model concept. Each Dimension BO related to the specific
Fact BO can be mapped as Dimension Semantic Group
(DSG). On next, each DSG need to be connected with the FSG
using Link. The early mapping will yield the topmost layer of
the GOOMD model Schema which exhibit high level
abstraction.
b) Detailed Mapping: In detailed mapping steps, the
multidimensional schema achieved from the early mapping
can be further refined to deliver the full-fledged conceptual
schema for DW system. The topmost layer GOOMD model
schema will be further modified in this step, from the
specifications available from detailed requirements analysis
phase. Firstly, the Dimension BOs of different granularity
related to each top level Dimension BO are mapped as
separate DSGs and are connected using Link to form the
dimension hierarchies. On next, the basic structural features
and identity from the Feature Tree of each Dimension BO are
mapped as ESG and Determinant ESG respectively. The
related ESGs and Determinant ESGs are connected using
Association and need to be encapsulated in the specific DSG.
Finally, each FSG can be modified from the feature tree of
related FO by extracting the relevant features into set of ESGs
and by encapsulating those ESGs. This step will exhibit all the
inner layers of GOOMD model schema.
The descriptions of possible OLAP operations for the
refined GOOMD model schema can be guided from the
Functional features of related set of Fact BO and Dimension
BOs. The DW System QoS requirements can be realized from
nonfunctional features of Fact BO.
V. FEATURES OF PROPOSED REQUIREMENTS
FRAMEWORK
The proposed requirements analysis framework for DW
system has been drawn from concepts of business object
model. Besides the advantages of using the business object
concept, the proposed framework facilitates several other
features and are as follows,
a) Process Driven Approach: The proposed
requirements analysis framework for DW system starts with
View Name and ID: Customer View 1
Target Fact BO: Sales
Requirements Objective: Analysis of Order Placing of Location XXX
Target Measures: Quantity
Actor BO:
Customer
Role:
Client
Activities:
Place Order;
Receive
Products;
Dimension BOs:
Time;
Product;
Procurements
Term ;
Event
BOs:
Order
Status

Constraint:
Construct()
invoke.
Figure 9. View Template related to Sales Fact BO
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
33 | P a g e
www.ijacsa.thesai.org
business process driven approach. The early requirements
analysis phase in the framework identify the concerned
business processes, events, stakeholders and interactions
between them possibly relevant to analytical task. The phase is
not biased from perspective of stakeholders understanding
about the system. Rather it ensures the modeling of the
business concept in high level of abstraction, how it is exist in
reality. So it is more realistic approach.
b) Object Orientation: The core of the proposed
framework is Business Object model, which supports the
general concepts and characteristics of object oriented
paradigm. Further, the requirements specification resulted
from the proposed framework can be mapped into any object
oriented conceptual level multidimensional data model
(GOOMD model as an example).
c) Abstraction: The proposed requirements analysis
framework is capable to represent the different business
concepts and stakeholders requirements in different level of
abstraction. Most abstract description of the analytical
requirements in the context of DW system is available from
the early requirements analysis phase of the framework.
Moreover the early mapping, which can be performed just
after the early requirements analysis phase, will produce the
conceptual multidimensional data model schemas with high
level of abstraction. In the detailed requirements framework
the early requirements descriptions are refined stepwise to
lower the abstraction level (see Table 2).

d) Reusability: One of the major advantages of the
proposed framework is that, it supports reuse of domain level
abstractions and step wise refinement mechanisms for
mapping the DW requirements in high level design. This is
important because the anticipation of future requirements for
the decision-making process is very difficult in large system
like DW. The support of abstraction mechanism and feature
oriented stepwise refinement of the BO based requirements
descriptions in detailed requirements analysis phase enable the
capability of reuse (see Table 2) of different types of BOs
specific to some business. A new BO can be formed from the
existing BO either at high level of abstraction in early
requirements analysis phase or by adding new features in
detailed analysis phase. The refinement processes of BOs are
iterative in nature.
e) Support for Multiple Analysis Approaches: The
proposed framework supports both Demand-driven and Mixed
analysis approaches for the requirements analysis of DW
system. Also it is capable to map the requirements to object
oriented multidimensional conceptual model by transforming
different detailed BOs into the relevant high level components
of the design model.
In the proposed framework, the early requirements analysis
phase and subsequent early mapping step are basically
business process driven and independent of any specific DW
requirements analysis approaches. But in detailed
requirements analysis, the step for comparison of refined Fact
BO and Dimension BO with the operational schemas
[subsection 3.C] is to support supply-driven part in the
proposed framework. By omitting this step from the detailed
requirements analysis phase, the proposed framework is fully
compatible with Demand-Driven DW requirements approach.
But with the presence of that step, the resultant GOOMD
model schema from the detailed mapping phase from refined
Process BOs and related Dimension BOs, are largely
influenced by the source operational schemas specifications.
In that case the proposed framework will support Mixed
Analysis approach towards DW requirements analysis.
VI. CONCLUSION
Requirements analysis plays a key role within DW system
development with the aim to reduce the risk of failure. A good
DW design method should be preceded by the requirements
elicitation and their analysis methodologies by considering
both user requirements and operational data sources for data
warehouse development. For the purpose, the Business Object
based requirements analysis framework for DW system has
been devised. The framework is comprised of three phases
namely, Early Requirements Analysis phase, Detailed
Requirements Analysis phase and Mapping phase. It starts
from understanding the set of business processes, events and
stakeholders in terms of set of well-defined BOs concerned to
some business domain in which DW system will operate. In the
framework, several requirements modeling elements like Fact
BO, Dimension BO, Measure attributes, Interaction Diagram,
Collaboration Diagram, Interaction Collaboration Network,
Feature Tree etc. have been described to express different
business concepts of the domain, relevant to DW system and in
the real business scenario. Finally the framework results the
mapping of DW requirements descriptions in high level design
components of conceptual level object oriented
multidimensional data model. The proposed framework
supports abstraction mechanism and reuse of different well
defined elements those have been used to realize the different
business concepts of the domain and useful for analytical task.
These features enable the framework to be used efficiently in
the evolving business processes over time. Further the
framework supports both demand-driven and mixed analysis
approach of DW requirements analysis.
Future work will include developing of a prototype tool in
the support of the proposed DW requirements analysis
framework. Moreover, quality evaluation of the proposed
TABLE II. ABSTRACTION AND REUSE POTENTIAL IN EACH PHASE OF
PROPOSED REQUIREMENTS ANALYSIS FRAMEWORK
Base Concepts and
Phases for proposed
Requirement analysis
framework
Level of
Abstraction Reuse
Potential
Productivity
Business Object
(Base Concept)
High High Low
Early Requirements
Analysis Phase

Early Mapping Phase
Detailed Requirements
Analysis Phase

Detailed Mapping
Phase
Low Low High
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
34 | P a g e
www.ijacsa.thesai.org
requirements analysis framework also is a prime objective of
the future work.
REFERENCES
[1] R. Kimball, M. Ross, The Data Warehouse Toolkit, Book Wiley &
Sons (2002).
[2] R. Winter and B. Strauch, A method for demand-driven information
requirements analysis in data warehousing projects, In Proc. HICSS, PP
13591365, 2003.
[3] F. R. S. Paim and J. B. Castro, DWARF: An approach for requirements
definition and management of data warehouse systems, Proc. of Int..
Conf. on Requirements Engineering, 2003.
[4] Robert Winter, Bernhard Strauch, Information requirements
engineering for data warehouse systems, Proceedings of the ACM
Symposium on Applied Computing, 2004.
[5] Rizzi, S., Abello, A., Lechtenborger, J., Trujillo, J., Research in data
warehouse modeling and design: dead or alive? Proc. of the 9th ACM
Int. Workshop on Data warehousing and OLAP, PP 310, 2006.
[6] M. Golfarelli, D. Maio, S. Rizzi, The dimensional fact model: A
conceptual model for data warehouses, Intl Journal of Cooperative
Information Systems, Vol. 7(2-3), PP 215247, 1998.
[7] B. Husemann, J. Lechtenborger, and G. Vossen. Conceptual data
warehouse design, Proc. 2nd Int. Workshop on Design and
Management of Data Warehouse, PP 39, 2000.
[8] D. Moody and M. Kortink, From enterprise models to dimensional
models: A methodology for data warehouse and data mart design, Proc.
2nd Int. Workshop on Design and Management of Data Warehouse,
2000.
[9] Paolo Giorgini, Stefano Rizzi, Maddalena Garzetti, GRAnD: A goal-
oriented approach to requirement analysis in data warehouses, Decision
Support Systems, Vol. 45(1), PP 4-21, 2008.
[10] N. Prakash and A. Gosain, Requirements Driven Data Warehouse
Development, Proc. of CAiSE Short Paper, 2003.
[11] Jose-Norberto Mazn, Juan Trujillo, Jens Lechtenbrger, Reconciling
Requirement-driven Data warehouses with Data Sources via
Multidimensional Normal Forms, Data & Knowledge Engineering,
Vol. 63(3), PP 725-751, 2007.
[12] Jose-Norberto Mazn and Juan Trujillo, An MDA approach for the
development of data warehouses, Decision Support Systems, Vol.
45(1), PP 41-58, 2008.
[13] OMG, Business Object DTF Common Business Objects, OMG
Document bom/97-11-11, ftp://ftp.omg.org/pub/docs/bom/97-11-11.pdf,
1997.
[14] A. Sarkar, S. Bhattacharya, The Graph Object Oriented
Multidimensional Data Model: A Conceptual Perspective, 16th Int.
Conf. on Software Engineering and Data Engineering (SEDE 2007), PP
165 170, 2007.
[15] A. Sarkar, S. Choudhury, N. Chaki, S. Bhattacharya, Conceptual Level
Design of Object Oriented Data Warehouse: Graph Semantic Based
Model, International Journal of Computer Science (INFOCOMP), Vol.
8(4), PP 60 70, December 2009.
[16] K. C. Kang, S. G. Cohen, J. A. Hess, W. E. Novak, A. Spencer Peterson,
"Feature-Oriented Domain Analysis (FODA) Feasibility Study",
Technical Report, Software Engineering Institute, Carnegie Mellon
University (USA), 1990.
[17] A. Sarkar, S. Choudhury, N. Chaki, S. Bhattacharya, Business-Object
Oriented Requirements Analysis Framework for Data Warehouses,
22nd International Conference on Software Engineering and Knowledge
Engineering (SEKE 2010), PP 34 37, July 2010.
[18] A. Sarkar, N. C. Debnath, Business Object Oriented Requirements
Analysis for Large Scale Information System, 20th International
Conference on Software Engineering and Data Engineering (SEDE
2011), PP 103 - 108, June 2011.
[19] A. C. Dhote, M. S. Ali, Materialized View Selection in Data
Warehousing, 4th International Conference on Information Technology
(ITNG '07), PP 843 847, 2007.
AUTHORS PROFILE
Anirban Sarkar is presently a faculty member in the
Department of Computer Applications, National
Institute of Technology, Durgapur, India. He received
his PhD degree from National Institute of
Technology, Durgapur, India in 2010. His areas of
research interests are Database Systems and Software
Engineering. His total numbers of publications in
various international platforms are about 30.



(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
35 | P a g e
www.ijacsa.thesai.org
A new graph based text segmentation using
Wikipedia for automatic text summarization

Mohsen Pourvali
Department of Electrical & Computer Engineering at
Qazvin Branch Islamic Azad University
Qazvin, Iran
Ph.D. Mohammad Saniee Abadeh
Department of Electrical & Computer Engineering at
Tarbiat Modares University
Tehran, Iran


AbstractThe technology of automatic document summarization
is maturing and may provide a solution to the information
overload problem. Nowadays, document summarization plays an
important role in information retrieval. With a large volume of
documents, presenting the user with a summary of each
document greatly facilitates the task of finding the desired
documents. Document summarization is a process of
automatically creating a compressed version of a given document
that provides useful information to users, and multi-document
summarization is to produce a summary delivering the majority
of information content from a set of documents about an explicit
or implicit main topic. According to the input text, in this paper
we use the knowledge base of Wikipedia and the words of the
main text to create independent graphs. We will then determine
the important of graphs. Then we are specified importance of
graph and sentences that have topics with high importance.
Finally, we extract sentences with high importance. The
experimental results on an open benchmark datasets from
DUC01 and DUC02 show that our proposed approach can
improve the performance compared to state-of-the-art
summarization approaches.
Keywords- text Summarization; Data Mining; Word Sense
Disambiguation.
I. INTRODUCTION
The technology of automatic document summarization is
maturing and may provide a solution to the information
overload problem. Nowadays, document summarization plays
an important role in information retrieval (IR). With a large
volume of documents, presenting the user with a summary of
each document greatly facilitates the task of finding the
desired documents. Text summarization is the process of
automatically creating a compressed version of a given text
that provides useful information to users, and multi-document
summarization is to produce a summary delivering the
majority of information content from a set of documents about
an explicit or implicit main topic [14]. Authors of the paper
[10] provide the following definition for a summary: A
summary can be loosely defined as a text that is produced
from one or more texts that conveys important information in
the original text(s), and that is no longer than half of the
original text(s) and usually significantly less than that. Text
here is used rather loosely and can refer to speech, multimedia
documents, hypertext, etc. The main goal of a summary is to
present the main ideas in a document in less space. If all
sentences in a text document were of equal importance,
producing a summary would not be very effective, as any
reduction in the size of a document would carry a proportional
decrease in its in formativeness. Luckily, information content
in a document appears in bursts, and one can therefore
distinguish between more and less informative segments.
Identifying the informative segments at the expense of the rest
is the main challenge in summarization. Assume a tripartite
processing model distinguishing three stages: source text
interpretation to obtain a source representation, source
representation transformation to summary representation, and
summary text generation from the summary representation. A
variety of document summarization methods have been
developed recently.
The paper [4] reviews research on automatic summarizing
over the last decade. This paper reviews salient notions and
developments, and seeks to assess the state of-the-art for this
challenging natural language processing (NLP) task. The
review shows that some useful summarizing for various
purposes can already be done but also, not surprisingly, that
there is a huge amount more to do. Sentence based extractive
summarization techniques are commonly used in automatic
summarization to produce extractive summaries. Systems for
extractive summarization are typically based on technique for
sentence extraction, and attempt to identify the set of
sentences that are most important for the overall understanding
of a given document. In paper [11] proposed paragraph
extraction from a document based on intra-document links
between paragraphs. It yields a text relationship map (TRM)
from intra-links, which indicate that the linked texts are
semantically related. It proposes four strategies from the
TRM: bushy path, depth-first path, segmented bushy path,
augmented segmented bushy path.
In our study we focus on sentence based extractive
summarization. We express that the lexical cohesion structure
of the text can be exploited to determine the importance of a
sentence. Eliminate the ambiguity of the word has a significant
impact on the inference sentence. In this article we will show
that the separation text into the inside issues by using the
correct concept Noticeable effect on the summary text is
created. We have used Word Sense Disambiguation [8] for
select correct concept. The experimental results on an open
benchmark datasets from DUC01 and DUC02 show that our
proposed approach can improve the performance compared to
state-of-the-art summarization approaches.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
36 | P a g e
www.ijacsa.thesai.org
II. RELATED WORK
Generally speaking, the methods can be either extractive
summarization or abstractive summarization. Extractive
summarization involves assigning salience scores to some
units (e.g. sentences, paragraphs) of the document and
extracting the sentences with highest scores, while abstraction
summarization
(e.g.http://www1.cs.columbia.edu/nlp/newsblaster/) usually
needs information fusion, sentence compression and
reformulation [14]. Sentence extraction summarization
systems take as input a collection of sentences (one or more
documents) and select some subset for output into a summary.
This is best treated as a sentence ranking problem, which
allows for varying thresholds to meet varying summary length
requirements. Most commonly, such ranking approaches use
some kind of similarity or centrality metric to rank sentences
for inclusion in the summary see, for example [1].The
centroid-based method [3] is one of the most popular
extractive summarization methods.
MEAD (http://www.summarization.com/mead/) is an
implementation of the centroid-based method for either single-
or-multi-document summarizing. It is based on sentence
extraction. For each sentence in a cluster of related documents,
MEAD computes three features and uses a linear combination
of the three to determine what sentences are most salient. The
three features used are centroid score, position, and overlap
with first sentence (which may happen to be the title of a
document). For single-documents or (given) clusters it
computes centroid topic characterizations using tfidf-type
data. It ranks candidate summary sentences by combining
sentence scores against centroid, text position value, and tfidf
title/lead overlap. Sentence selection is constrained by a
summary length threshold, and redundant new sentences
avoided by checking cosine similarity against prior ones. In
the past, extractive summarizers have been mostly based on
scoring sentences in the source document. In paper [12] each
document is considered as a sequence of sentences and the
objective of extractive summarization is to label the sentences
in the sequence with 1 and 0, where a label of 1 indicates that
a sentence is a summary sentence while 0 denotes a non-
summary sentence. To accomplish this task, is applied
conditional random field, which is a state-of-the-art sequence
labeling method.
In paper [15] proposed a novel extractive approach based
on manifoldranking of sentences to query-based multi-
document summarization. The proposed approach first
employs the manifoldranking process to compute the
manifoldranking score for each sentence that denotes the
biased information-richness of the sentence, and then uses
greedy algorithm to penalize the sentences with highest overall
scores, which are deemed both informative and novel, and
highly biased to the given query.
The summarization techniques can be classified into two
groups: supervised techniques that rely on pre-existing
document-summary pairs, and unsupervised techniques, based
on properties and heuristics derived from the text. Supervised
extractive summarization techniques treat the summarization
task as a two-class classification problem at the sentence level,
where the summary sentences are positive samples while the
non-summary sentences are negative samples. After
representing each sentence by a vector of features, the
classification function can be trained in two different manners
[7]. One is in a discriminative way with well-known
algorithms such as support vector machine (SVM) [16]. Many
unsupervised methods have been developed for document
summarization by exploiting different features and
relationships of the sentences see, for example [3] and the
references therein. On the other hand, summarization task can
also be categorized as either generic or query-based. A query-
based summary presents the information that is most relevant
to the given queries [2] and [14] while a generic summary
gives an overall sense of the documents content [2] , [4] ,
[12] , [14].
The QCS system (Query, Cluster, and Summarize) [2]
perform the following tasks in response to a query: retrieves
relevant documents; separates the retrieved documents into
clusters by topic, and creates a summary for each cluster. QCS
is a tool for document retrieval that presents results in a format
so that a user can quickly identify a set of documents of
interest. In paper [17] are developed a generic, a query-based,
and a hybrid summarizer, each with differing amounts of
document context. The generic summarizer used a blend of
discourse information and information obtained through
traditional surface-level analysis. The query-based summarizer
used only query-term information, and the hybrid summarizer
used some discourse information along with query-term
information. The article [18] presents a multi-document,
multi-lingual, theme-based summarization system based on
modeling text cohesion.
III. CREATE GRAPH AND TEXT SEGMENTATION
The algorithm presented in this paper, at first the input text
is pre-processing and the stop words is removed. Then stem of
words is found and its (POS) is tagged.
Only verbs and nouns are used in the text, in the way we
have presented. The algorithm starts from the beginning of the
main text, and take the word, and using Wikipedia knowledge
base provides a two-level tree from the links of the word
abstract. So that root of the word is the same word and tree
Children are related words (links) to the target word in the
abstract of its web page. Then it searches the children of the
target word in the input text and it creates a graph using target
word and the words that both are in the children of the
previous step tree and input text.
Let s = {s
1
, s
2
, ..., s
n
} is the set of sentences and
{

} is the set of all words are nouns or

verbs in the input text. So that w
ij
shows i-th word in the j-th
sentence. Since the goal is to extract sentences with high
importance. The sentences are considered as nodes. The
relationships between words within a sentence with other
sentences words are considered to be edges in the graph. The
algorithm is shown in Figure 1.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
37 | P a g e
www.ijacsa.thesai.org
For n=0 to EndOfSentenc
For i=0 to

Child[] = CreateTreeInWiki(

)
For r=0 to EndOfChild
For k=0 to EndOfWord
If Child[r] == AnyWordOf_W
Graph[] = Create_Or_Update_Graph(

)
Endif
EndFor
EndFor
EndFor
EndFor
Figure 1. Base algorithm for create the tree and graph
In the above algorithm, Child is the children of target word
tree in the Wikipedia, and Graph is the constructed graph
from the sentences that target word is in them. This algorithm
is implemented for all target words in the input text. Finally,
we have several independent graphs, that according to the
relationship between its nodes, each graph implies a topic in
the input text. Figure 2 shows related sentences in the text.

Figure 2. Related sentences and segments, there are two segments with two
colors (blow and red)
After extracting the graphs of the input text, the graphs
edges were given weight. According to the distance between
the words in two sentences, existed in the two sides of the
edge, the weighting to the edge is done. To do this we use
Average Google normalized distance [19]. NGD takes
advantage of the number of hits returned by Google to
compute the semantic distance between concepts. The
concepts are represented with their labels which are fed to the
Google search engine as search terms.
First, using the NGD we define the global and local
dissimilarity measure between terms (as shown in [19] the
NGD is nonnegative and does not satisfy the triangle
inequality, i.e. hence isnt distance and consequently in the
further it we shall name dissimilarity measure). According to
definition NGD the global dissimilarity measure between
terms t
k
and t
l
also is defined by the formula:

)
{(

) (

)}(

{(

) (

)}

Where

is the number of web pages containing the

search term t
k
, and

denotes the number of web pages

containing both terms t
k
and t
l
, N
Google
is the number of web
pages indexed by Google. The main properties of the NGD
[19] are listed as follows:
1) The range of the NGD is in 0 and ;
If t
k
= t
l
or if t
k
t
l
but frequency

> 0, Then

) = 0. That is, the

semantics of t
k
and t
l
, in the Google sense is the same.
If frequency

= 0, then for every term t

k
, we
have

, and the

,
which we take to be 1 by definition.
If frequency

0 and

= 0, we take

) = 1.
2) NGD (tk,tk) = 0 for every tk. For every pair t
k
and t
l
, we
have

) =

): It is symmetric.
Formula 3 is the dissimilarity measure between sentences
S
i
and S
j
.

That m
i
and m
j
are the number of words in i-th and j-th
sentences. Then, the weighting of the graph, we are selecting
the heavier graph (the graph that has heavy nods and light
edges). Using the following formula a weight is assigned to
each graph.

That L is number of nodes and E is number of edges in any
graph, d
i
is the degree of i-th node.
IV. SENTENCE EXTRACTION
Finally, the graph which is higher than other graphs
contains the main topic of the text. In formula 1, sentences can
be extracted according to the percent of summarization. If we
want to have the summarization of other topics in addition to
main topic in the text we extract important sentences from the
important graph according to the summarization percent.
Using the following formula, each node is evaluated according
to its number of incoming and outgoing edges.

(4)
Where O
i
is number of outputs from i-th sentence and I
i
is
number of inputs to i-th sentence. We use following formula
to calculate the weight of the word W
ti
.

(5)
That TF
ti
is the number of occurrences phrase t in the
sentence S
i
, and ISF is:
(

) (6)
N
t
is the number of sentences the word t
i
has occurred in it.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
38 | P a g e
www.ijacsa.thesai.org
V. EXPERIMENTS AND RESULTS
In this section, we conduct experiments to test our
summarization method empirically.
A. Datasets
For evaluation the performance of our methods we used
two document datasets DUC01 and DUC02 and corresponding
100-word summaries generated for each of documents. The
DUC01 and DUC02 are an open benchmark datasets which
contain 147 and 567 documents-summary pairs from
Document Understanding Conference (http://duc.nist.gov).
We use them because they are for generic single-document
extraction that we are interested in and they are well
preprocessed. These datasets DUC01 and DUC02 are
clustered into 30 and 59 topics, respectively. In those
document datasets, stop words were removed using the stop
list provided in ftp://ftp.cs.cornell.e-du/pub/smart/english.stop
and the terms were stemmed using Porters scheme [9], which
is a commonly used algorithm for word stemming in English.
B. Evaluation metrics
There are many measures that can calculate the topical
similarities between two summaries. For evaluation the results
we use two methods. The first one is by precision (P), recall
(R) and F1-measure which are widely used in Information
Retrieval. For each document, the manually extracted
sentences are considered as the reference summary (denoted
by Summ
ref
). This approach compares the candidate summary
(denoted by Summ
cand
) with the reference summary and
computes the P, R and F1-measure values as shown in formula
(9) [12].


|

|
|

|
(7)


|

|
|

|
(8)

(9)

The second measure we use the ROUGE toolkit [5], [6] for
evaluation, which was adopted by DUC for automatically
summarization evaluation. It has been shown that ROUGE is
very effective for measuring document summarization. It
measures summary quality by counting overlapping units such
as the N-gram, word sequences and word pairs between the
candidate summary and the reference summary. The ROUGE-
N measure compares N-grams of two summaries, and counts
the number of matches. The measure is defined by formula
(10) [5], [6].

()

()

(10)

Where N stands for the length of the N-gram, Count
match

(N-gram) is the maximum number of N-grams co-occurring in
candidate summary and a set of referencesummaries. Count
(N_gram) is the number of N-grams in the reference
summaries. We use two of the ROUGE metrics in the
experimental results, ROUGE-1 (unigram-based) and
ROUGE-2 (bigram-based).
C. Simulation strategy and parameters
The parameters of our method are set as follows: depth of
tree that is created for any word, n=3; extra value for Lesk
algorithm, =5; Finally, we would like to point out that
algorithm was developed from scratch in C#.net 2008 platform
on a Pentium Dual CPU, 1.6 GHz PC, with 512 KB cache, and
1 GB of main memory in Windows XP environment.
D. Performance evaluation and discussion
We compared our method with four methods CRF [12],
NetSum [13], ManifoldRanking [15] and SVM [16]. Tables 1
and 2 show the results of all the methods in terms ROUGE-1,
ROUGE-2, and F1-measure metrics on DUC01 and DUC02
datasets, respectively. As shown in Tables 1 and 2, on DUC01
dataset, the average values of ROUGE-1, ROUGE-2 and F1
metrics of all the methods are better than on DUC02 dataset.
As seen from Tables 1 and 2 ManifoldRanking is the worst
method, In the Tables 1 and 2 highlighted (bold italic) entries
represent the best performing methods in terms of average
evaluation metrics. Among the methods NetSum, CRF, SVM
and ManifoldRanking the best result shows NetSum. We use
relative improvement
( )

for
comparison. Compared with the best method NetSum, on
DUC01 (DUC02) dataset our method improves the
performance by 3.43% (4.82%), 7.15% (16.30%) and 3.12%
(4.28%) in terms ROUGE-1, ROUGE-2 and F1, respectively.
TABLE I. AVERAGE VALUES OF EVALUATION METRICS FOR
SUMMARIZATION METHODS (DUC01 DATASET).
Av.F1-
measure
Av.ROUGE-2 Av.ROUGE-1 Methods
0.48743 0.18962 0.48021 Our method
0.47267 0.17697 0.46427 NetSum
0.46435 0.17327 0.45512 CRF
0.45357 0.17018 0.44628 SVM
0.44368 0.16635 0.43359
Manifold
Ranking

TABLE II. AVERAGE VALUES OF EVALUATION METRICS FOR
SUMMARIZATION METHODS (DUC02 DATASET).
Av.F1-
measure
Av.ROUGE-2 Av.ROUGE-1 Methods
0.48259 0.12986 0.47129 Our method
0.46278 0.11167 0.44963 NetSum
0.46046 0.10924 0.44006 CRF
0.43095 0.10867 0.43235 SVM
0.41657 0.10677 0.42325
Manifold
Ranking
VI. CONCLUSION
We have presented the approach to automatic document
summarization based on creating graph and text segmentation
and extraction of sentences using Wikipedia. Our approach
consists of two steps. First creates a two-level tree from the
links of the words abstract, and then creates graph using of
previous phase, and finally selects important segments which
were created using of previous graph. When comparing our
methods with several existing summarization methods on an
open DUC01 and DUC01 datasets, we found that our methods
can improve the summarization results significantly. The
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
39 | P a g e
www.ijacsa.thesai.org
methods were evaluated using ROUGE-1, ROUGE-2 and F1
metrics. In this paper we also demonstrated that the
summarization result depends on the similarity measure.
Results of experiment have showed that proposed by us NGD-
based dissimilarity measure outperforms the Euclidean
distance.
REFERENCES
[1] Alguliev, R. M., & Alyguliev, R. M. (2007). Summarization of text-
based documents with a determination of latent topical sections and
information-rich sentences. Automatic Control and Computer Sciences ,
41, 132140.
[2] Dunlavy, D. M., OLeary, D. P., Conroy, J. M., & Schlesinger, J. D.
(2007). QCS: A system for querying, clustering and summarizing
documents. Information Processing and Management , 43, 15881605.
[3] Erkan, G., & Radev, D. R. (2004). Lexrank: Graph-based lexical
centrality as salience in text summarization. Journal of Artificial
Intelligence Research , 22, 457479.
[4] Jones, K. S. (2007). Automatic summarizing: The state of the art.
Information Processing and Management , 43, 14491481.
[5] Lin, C. -Y. (2004). ROUGE: A package for automatic evaluation
summaries. In Proceedings of the workshop on text summarization
branches out, (pp. 7481). Barcelona, Spain.
[6] Lin, C. -Y., & Hovy, E. H. (2003). Automatic evaluation of summaries
using N-gram co-occurrence statistics. In Proceedings of the 2003
conference of the north american chapter of the association for
computational linguistics on human language technology (HLT-NAACL
2003), (pp. 7178). Edmonton, Canada.
[7] Mihalcea, R., & Ceylan, H. (2007). Explorations in automatic book
summarization. In Proceedings of the 2007 joint conference on
empirical methods in natural language processing and computational
natural language learning (EMNLP-CoNLL 2007), (pp. 380389).
Prague, Czech Republic.
[8] Navigli, R., & Lapata, M. (2010). An Experimental Study of Graph
Connectivity for Unsupervised Word Sense Disambiguation. IEEE
Computer Society , 32.
[9] Porter, M. (1980). An algorithm for suffix stripping. Program , 14, 130
137.
[10] Radev, D., Hovy, E., & McKeown, K. (2002). Introduction to the special
issue on summarization. omputational Linguistics , 22, 399408.
[11] Salton, G., Singhal, A., Mitra, M., & Buckley, C. (1997). Automatic text
structuring and summarization. Information Processing and
Management , 33, 193207.
[12] Shen, D., Sun, J. -T., Li, H., Yang, Q., & Chen, Z. (2007). Document
summarization using onditional random fields. In Proceedings of the
20th international joint conference on artificial intelligence (JCAI
2007), (pp. 28622867). Hyderabad, India.
[13] Svore, K. M., Vanderwende, L., & Burges, C. J. C. Enhancing single-
document summarization by combining RankNet and third-party
sources. In Proceedings of the 2007 joint conference on empirical
methods in natural language processing and computational natural
language learning (EMNLP-CoNLL 2007), (pp. 448457). Prague,
Czech Republic.
[14] Wan, X. (2008). Using only cross-document relationships for both
generic and topic-focused multi-document summarizations. Information
Retrieval , 11, 2549.
[15] Wan, X., Yang, J., & Xiao, J. (2007). Manifold-ranking based topic-
focused multidocument summarization. In Proceedings of the 20th
international joint conference on artificial intelligence (IJCAI 2007),
(pp. 29032908). Hyderabad, India.
[16] Yeh, J-Y., Ke, H-R., Yang, W-P., & Meng, I-H. (2005). Text
summarization using a trainable summarizer and latent semantic
analysis. Information Processing and Management , 41, 7595.
[17] McDonald, D. M., & Chen, H. (2006). Summary in context: Searching
versus browsing. ACM Transactions on Information Systems, 24, 111
141.
[18] Fung, P., & Ngai, G. (2006). One story, one flow: Hidden Markov story
models for multilingual multi document summarization. ACM
Transaction on Speech and Language Processing, 3, 116.
[19] Cilibrasi, R. L., & Vitanyi, P. M. B. (2007). The Google similarity
measure. IEEE Transaction on Knowledge and Data Engineering, 19,
370383.
[20] Alguliev, Rasim; Aliguliyev, Ramiz. Evolutionary algorithm for
extractive text summarization. http://www.highbeam.com/doc/1G1-
214205320.html.
[21] Stergos Afantenos, Vangelis Karkaletsis, Panagiotis Stamatopoulos.
Summarization from medical documents: a survey.
[22] Xiaojun Wan. Towards a Unified Approach to Simultaneous Single-
Document and Multi-document Summarizations.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

40 | P a g e
www.ijacsa.thesai.org
Automated Periodontal Diseases Classification
System

Aliaa A. A. Youssif
Department of Computer Science,
Faculty of Computers and
Information
Helwan University
Cairo, Egypt
Abeer Saad Gawish
Department of Oral Medicine and
Periodontology, Faculty of Dental
Medicine (Girls Branch)
Al-Azhar University
Cairo, Egypt
Mohammed Elsaid Moussa
Department of Computer Science,
Faculty of Computers and
Information
Helwan University
Cairo, Egypt


Abstract This paper presents an efficient and innovative system
for automated classification of periodontal diseases, The strength
of our technique lies in the fact that it incorporates knowledge
from the patients' clinical data, along with the features
automatically extracted from the Haematoxylin and Eosin (H&E)
stained microscopic images. Our system uses image processing
techniques based on color deconvolution, morphological
operations, and watershed transforms for epithelium &
connective tissue segmentation, nuclear segmentation, and
extraction of the microscopic immunohistochemical features for
the nuclei, dilated blood vessels & collagen fibers. Also,
Feedforward Backpropagation Artificial Neural Networks are
used for the classification process. We report 100% classification
accuracy in correctly identifying the different periodontal
diseases observed in our 30 samples dataset.
Keywords-Biomedical image processing; epithelium segmentation;
feature extraction; nuclear segmentation; periodontal diseases
classification.
I. INTRODUCTION
Periodontitis is a chronic inflammatory disease of
vascularized supporting tissues of the teeth [1], Periodontal
disease occurs when inflammation or infection affect the
gingiva and extend to the periodontal apparatus [2], The 1999
classification system for periodontal diseases and conditions
listed seven major categories of periodontal diseases [3,4]:
Gingivitis, Chronic periodontitis, Aggressive periodontitis,
Periodontitis as a manifestation of systemic disease,
Necrotizing ulcerative gingivitis/periodontitis, Abscesses of
the periodontium, Combined periodontic-endodontic lesions;
The latter 4 are associated with systemic diseases [5,6,7,8,9];
Hence, this work preliminary focuses on identifying the
different types of periodontal diseases by using computer-
assisted microscopy system for automated classification of
periodontal diseases to increase the accuracy and reduce
the workload in classifying and diagnosis of the different
categories of periodontal diseases, which help in designing
the treatment plan used.
The paper is organized as follows: Section II explains the
materials and methods that we followed to get our dataset,
Section III describes the implementation details of our
proposed system, Section IV shows experimental results and
discussions about these results, and Section V concludes the
paper and introduces the future work that can be done.
II. MATERIALS AND METHODS
A. Study Cases
The study was conducted on 32 patients attending between
February 2009 and March 2011 to the Oral Medicine,
Periodontology, Oral Diagnosis and Radiology Department,
Faculty of Dental Medicine Girls' Al-Azhar University.
Patients were suffering from gingival inflammation, which
may extend to include the periodontium (different types of
periodontitis) or gingival overgrowth (due to different
etiological factors).
Patients were excluded from the present study if they were
smokers, pregnant or post-menopausal women. All selected
patients had not undergone any periodontal therapy for at least
six months.
B. Collected Clinical Data
The following clinical parameters were collected and
recorded on six sites at each tooth; all linear measurements
were recorded to the nearest 0.5 mm using William graduated
periodontal probe: Plaque index (PI) [10] Pocket Depth (PD)
and Clinical attachment level (CAL) was measured from the
CEJ to the apical part of the sulcus, All included patients were
indicated for surgical flaps as a line of their treatment.
C. Histopathological Sample Preparation
After the surgical procedures; the excised tissue
samples were immersed in 10% formalin and decalcified in
multiple baths of 10% trichloroacetic acid. The blocks were
immersed in paraffin, and semi-serial 4 m histologic sections
were stained with Haematoxylin and Eosin (H&E).
D. Image Capture
Representative sections were photographed using a Leitz
DMRD Microscope (Leica, Wetzlar, Germany) with 20 X
objective UPLanFl (resolution 0.67 m) at a size of 1600 X
1200 pixels (interpixel distance 0.62 m) using a JVC KY-
55B 3-CCD colour camera attached to a 24 bit RGB
frame grabber (Imaging Technologies IT4PCI, Bedford,
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

41 | P a g e
www.ijacsa.thesai.org
MA, U.S.A.) controlled by Optimas v. 6.51 (Media
Cybernetics, Silver Spring, MD, U.S.A.)
E. Periodontal Diseases Groups
The 32 patients were classified into 4 diagnostic groups
according to their periodontal status; 16 patients were
classified into Gingival Enlargement group, 7 patients into
Chronic Gingivitis group, another 7 patients into Chronic
Periodontitis group, and the last 2 patients into Aggressive
Periodontitis group.
We excluded the patients which were classified into the
Aggressive Periodontitis group from our classification
experiments; because of the very small number of study cases
that we have for this group; which was not sufficient for our
neural networks based classification system to work with.
Although we included these 2 cases in our preprocessing and
feature extraction experiments, to make sure that our proposed
system is generic enough to handle all of the different types of
diseases.
We also included another study case for a healthy person
to our preprocessing and feature extraction experiments.
III. PROPOSED SYSTEM
Our proposed system can be divided into several well-
defined stages as illustrated in Figure 1. We first get the H&E
stained slide, do some preprocessing steps to it to remove
background and undesired objects, segment it to the tissues
main components epithelium and connective tissue, then
extract features from both parts, and then use these features
along with the clinical data of the patient to start the
classification process.

Figure 1 Schematic structure of the proposed system
A. Pre-processing
Mainly images consist of 2 parts: Tissue (Epithelium +
Connective Tissue) and Background. In this preprocessing
phase, we need to remove the background areas from the
image, allowing the further processing to be done only on the
tissue part.
We implemented 2 different algorithms to achieve this
task; the first one is fully-automated, while the second one is
semi-automated because it requires some clicks from the user
to specify the seed points in the background area. The pseudo
codes for fully-automated and semi-automated algorithms are
shown below in Pseudocode I & Pseudocode II respectively.
0shows one sample result of the 2 algorithms. Also, sensitivity
and specificity results of these algorithms are displayed in the
Results and Discussion section.
Pseudocode I FULLY-AUTOMATED PREPROCESSING ALGORITHM
1. Get the "a" channel from L*a*b* color space of the
original image, Image
1
.
2. Smoothing of Image
1
with a 5 5 average lter to
preserve only large detail [11], Image
2
.
3. Convert Image
2
to black & white with 0.2 threshold
value, and then invert it, Mask
1
.
4. Morphologically open Mask
1
to remove all black
objects that have fewer than 1000 pixels, Mask
2
.
5. Morphologically open Mask
2
to remove all white
objects that have fewer than 1000 pixels, Mask
3
.
6. Perform a morphological dilatation with a 10 10
window size on the Mask
3
, Mask
4
.
7. Fill Holes of Mask
4
, Mask
5
.
8. Perform a morphological erosion with a 10 10 window
size on the Mask
5
, Mask
6
.
9. Remove areas from Mask
6
that are not on border [12],
Mask
7
.
10. Morphologically open Mask
7
to remove all objects that
have more than 10000 pixels, Mask
8
.
11. Apply final smoothing for Mask
8
, Mask
9
.

Pseudocode II SEMI-AUTOMATED PREPROCESSING ALGORITHM
1. Get the "Green" channel of the original image, Image
1
.
2. Get Seed Point from the pathologist, inside the
background region, then set the initial threshold to the
intensity value of this seed point.
3. Based on the initial threshold, get all the pixels that its
intensity in a range of 50 values, Mask
1
, i.e.
abs(newPixelValue - initalThreshold) <= 50
4. Perform image reconstruction on the Mask
1
to get only
the pixels that are connected to the seed point, Mask
2
.
5. Perform a morphological erosion with a 1 1 window
size on the Mask
2
, Mask
3
.
6. Morphologically open Mask
3
to remove all white
objects that have fewer than 1000 pixels, Mask
4
.
7. Perform a morphological dilation with a 2 2 window
size on the Mask
4
, Mask
5
.
8. Morphologically open Mask
5
to remove all black
objects that have fewer than 1000 pixels, Mask
6
.
9. Perform a morphological erosion with a 1 1 window
size on the Mask
6
, Mask
7
.
10. Repeat steps 2 to 10 for each separate background
region.
11. Edge shifting for the Mask
7
by performing
morphological erosion with a disk-shaped structuring
element [13], [14] with radius 2 the Mask
7
, Mask
8
.

Clinical
Data
Features
H&E Slide
Preprocessing
Epithelium
Segmentation
Feature
Extraction
Classification
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

42 | P a g e
www.ijacsa.thesai.org
For very small background regions, we skipped steps from 6 to 10 for these regions only. This
happened only 3 times over our 33 images.


(A)

(B)

(C)
Figure 2 Pre-processing sample (A) The original image. (B) After applying
fully-automated algorithm. (C) After applying semi-automated algorithm.
B. Epithelium Segmentation
In this phase of our system, we need to segment the tissue
into Epithelium and Connective Tissue; Epithelium
segmentation was targeted in some studies [15,16,17]. As
stated in [15] there are a lot of challenges doing this task, and
they are mainly caused by staining artifacts, lighting
acquisition conditions, and undesired touching objects. Some
studies [15] used the saturation channel in the HSV color
space, others [16] used multiple gray level automated
thresholding over the images green channel, while others [17]
used region growing algorithm [18] over the gray channel of
the image.
We found that 73% of our samples give better results when
using the HSV Saturation channel, while the remaining 27%
give better results when using the RGB green channel. We
implemented an automated way for this segmentation process;
Pseudocode III describes our epithelium segmentation
algorithm.
We also implemented a post-processing step, to move
some parts from epithelium to connective tissue or vice versa
manually, because our proposed algorithm didn't achieve
highly accurate results for all study cases, because the
brightness value of the infiltrate was similar to that of the
epithelium in images with high inflammatory infiltrate [16].
Information about sensitivity and specificity of our algorithm
is found in the Results and Discussion section. Figure 3 shows
one sample result of this algorithm.
C. Feature Extraction
We extracted a set of features from our Haematoxylin and
Eosin (H&E) stained slides, that we thought it will be helpful
in the classification process later. Below are the extracted
features along with the algorithms we used to extract it:
Pseudocode III EPITHELIUM SEGMENTATION ALGORITHM
1. Get the "Saturation" channel from the HSV color space
of the tissue image
a
, Image
1
, which retrieved from the
preprocessing phase, Channel
1
.
2. Adjust
b
intensity values in the grayscale Channel
1
, to
map it to new values in such that 1% of data is saturated
at low and high intensities of the Channel
1
. This
increases the contrast of the output channel, Channel
2
.
[15].
3. Apply 10 10 median lter
c
for removing the noise in
the Channel
2
without harming edges, Channel
3
. [19]
4. Convert Channel
3
to black & white by applying global
thresholding through Otsu's method [20], Mask
1
.
d

5. Fill Holes of Mask
1
, Mask
2
.
6. Perform a morphological erosion with a 10 10 window
size on the Mask
2
, Mask
3
.
7. Morphologically open Mask
3
to remove all white
objects that have fewer than 1000 pixels, Mask
4
.
8. Perform a morphological dilation with a 20 20 window
size on the Mask
4
, Mask
5
.
9. Morphologically open Mask
5
to remove all black
objects that have fewer than 1000 pixels, Mask
6
.
10. Perform a morphological erosion with a 10 10 window
size on the Mask
6
, Mask
7
.
11. Edge shifting for the mask by performing a
morphological erosion with a strel 2 2 window size on
the Mask
7
, Mask
8
. [13], [14].

a
For 9 samples of our 33 samples, we used RGB Green channel instead of HSV Saturation
channel, because we found it giving better results.
b
For 1 sample, we replaced image adjustment step with histogram equalization.
c
For 6 samples, we used a 20 20 window size for median filtering instead of the 10 10
window size, and for another 1 sample we used a 9 9 window size.
d
For 12 samples, we applied histogram equalization after step 4.
1) Epithelium Percentage
After segmentation of the tissue image into its components
(Epithelium and Connective Tissue) in the previous phase, we
calculated the percentage of the Epithelium part.
(Note that the background regions are already excluded)
2) Nuclei in Epithelium
We segmented the nuclei that found in the epithelium part,
and used its count and percentage from the whole epithelium
in our features list, Nuclei segmentation was targeted in some
studies such as [11], [21], [22] which were using nuclear
localization achieved using the colour deconvolution
algorithm developed by [23] to obtain the optical density of
the Haematoxylin stain alone plus spatial partition of the
epithelial compartment representing the exclusive area of
inuence of each nucleus prole, [24] which is based on the
multiscale Laplacian-of-Gaussian (LoG) lter, [25] which
depends on ellipse fitting and the watershed transform, [26]
which relies on some morphology-based techniques.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

43 | P a g e
www.ijacsa.thesai.org

(A)

(B)

(C)

(D)

(E)
Figure 3 Epithelium Segmentation (A) Original tissue image after pre-
processing phase. (B) Automatically segmented epithelium. (C) Automatically
segmented connective tissue. (D) Epithelium after post-processing.
(E) Connective-Tissue after post-processing.
In our proposed system, we used steps in Pseudocode IV to
segment the nuclei in the epithelium part of the image; Figure
4 shows two sample results of this algorithm.
3) Nuclei in Connective Tissue
We segmented the nuclei that found in the connective
tissue component, and used its count and percentage from the
whole connective tissue in our features list.
We do this by getting 2 masks; the first one represents the
light areas in the connective tissue (blood vessels) while the
second one represents the dark areas (nuclei). Then we need to
filter the nuclei mask to exclude the ones that are related to the
dilated blood vessels. Our steps for doing this segmentation
part are described in Pseudocode V. Figure 5 shows one
sample result of this algorithm.
4) Dilated Blood Vessels
We segmented the dilated blood vessels that found in the
connective tissue part, and used its count and percentage from
the whole connective tissue in our features list.
We do this with almost the same steps described in the
Pseudocode V that were used in the connective tissue nuclei
segmentation. We get the first mask with exactly the same
steps, while excluding steps e, f and j from the ones required
to get the second mask. And to get the final mask, we filter
parts in Mask
1
(dilated blood vessels candidates) to only the
ones that intersect at least 3 times with parts of Mask
2
(nuclei);
finally, we count and get the percentage of all objects in
Mask
3
. 0Figure 6 shows one sample result of this algorithm.
Pseudocode IV EPITHELIUM NUCLEI SEGMENTATION ALGORITHM
1. Get the optical density of the Haematoxylin as an 8 bit
grayscale image, by colour deconvolution [23] of the
epithelium part, retrieved from the epithelium
segmentation phase, Channel
1
.
a

2. Normalize intensity values in the grayscale Channel
1
, to
change the range of pixel intensity values to be between
0 and 255, Channel
2
.
b

3. Convert Channel
2
to black & white by computing the
extended-minima transform, which is the regional
minima of the H-minima transform [27], with a value of
30, Mask
1
.
4. Fill Holes of Mask
1
, Mask
2
.
5. Remove areas from Mask
2
that are not on border,
Mask
3
.
6. Morphologically open Mask
3
to remove all objects that
have less than 15 pixels, Mask
4
.
7. Morphologically open Mask
4
to remove all objects that
have more than 800 pixels, Mask
5
.
8. Convert each connected object to the ellipse that fit it,
Mask
6
.
9. Finally we count and get the percentage of all objects in
Mask
6
.

a
For 9 samples of our 33 samples, we used RGB Red channel instead of Haematoxylin channel,
because we found it giving better results.
b
For 3 samples, we also apply image adjustment after step 2.


(A)

(B)

(C)

(D)
Figure 4 Nuclei in Epithelium. (A), (C) Two samples of epithelium images.
(B), (D) After applying epithelium nuclei segmentation algorithm over (A)
and (C) respectively. [Nuclei are surrounded by blue ellipses].
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

44 | P a g e
www.ijacsa.thesai.org
Pseudocode V CONNECTIVE-TISSUE NUCLEI SEGMENTATION
ALGORITHM
1. Get Mask
1

a. Get the "Green" channel of the connective tissue part,
retrieved from the epithelium segmentation phase,
Channel
1
.
b. Enhance the contrast of the grayscale channel
Channel
1_1
by transforming the values using contrast-
limited adaptive histogram equalization (CLAHE)
[28], Channel
1_1
.
c. Apply 3 3 median filter for removing the noise [19]
in the Channel
1_1
without harming edges, Channel
1_2
.
d. Convert Channel
1_2
to black & white with 0.8
threshold value, Mask
1_1
.
e. Morphologically open Mask
1_1
to remove all objects
that have less than 80 pixels, Mask
1
.
2. Get Mask
2

a. Get the optical density of the Haematoxylin as an 8
bit grayscale image, by colour deconvolution [23] of
the connective tissue part, retrieved from the
epithelium segmentation phase, Channel
2
.
b. Adjust intensity values in the grayscale Channel
2
, to
map it to new values in such that 1% of data is
saturated at low and high intensities of the Channel
2
.
This increases the contrast of the output channel,
Channel
2_1
. [15].
c. Apply 2-D adaptive filtering with a 5 5 window size
over Channel
2_1
, Channel
2_2
.
d. Convert Channel
2_2
to black & white by computing
the extended-minima transform, which is the regional
minima of the H-minima transform [27], with a value
of 30, Mask
2_1
.
e. Morphologically open Mask
2_1
to remove all objects
that have less than 15 pixels, Mask
2_2
.
f. Morphologically open Mask
2_2
to remove all objects
that have more than 800 pixels, Mask
2_3
.
g. Apply watershed algorithm for Mask
2_3
, Mask
2_w
.
h. Perform a morphological dilation with a 10 10
window size on the Mask
2_3
, Mask
2_4
.
i. Apply AND operator between Mask
2_4
and Mask
2_w
,
Mask
2_5
.
j. Remove areas from Mask
2_5
that are not on border,
Mask
2
.
3. Get Final Mask
a. Filter parts in Mask
2
(Nuclei) to only ones that not
intersect with parts of Mask
1
(Light Parts) that
intersects with at least 3 times with parts of Mask
2
,
Mask
3_1
.
b. Apply image reconstruction algorithm between
Mask
3_1
and Mask
2_3
, Mask
3_2
.
c. Convert each connected object in Mask
3_2
to the
ellipse that fit it, Mask
3
.
4. Finally we count and get the percentage of all objects in
Mask
3
.



(A)

(B)

(C)
Figure 5 Nuclei in Connective Tissue. (A) Sample of connective tissue image.
(B) Nuclei in connective tissue surrounded by black ellipses. (C) Mask1is in
blue, Mask2 is in red and yellow, and Final Mask is in yellow.

(A)

(B)

(C)

(D)
Figure 6 Dilated Blood Vessels. (A) Sample of connective tissue image.
(B) Dilated Blood Vessels in connective tissue surrounded by black ellipses.
(C) Candidate blood vessels are in blue and light blue. Red areas are nuclei.
(D) Candidate blood vessels are in blue and light blue. Final Mask (passed
blood vessels) is in light blue.
5) Collagen Fiber
We segmented the collagen fiber that found in the
connective tissue part, and used its percentage from the whole
connective tissue in our features list.
We do this by getting 2 masks, the first one represents the
light areas in the connective tissue, and the second one
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

45 | P a g e
www.ijacsa.thesai.org
represents the dark areas, then we need to filter the nuclei
mask to exclude the ones that are related to the dilated blood
vessels. Our steps for doing this segmentation part are
described in Pseudocode VI. 0Figure 7 shows one sample
result of this algorithm.
Pseudocode VI COLLAGEN FIBER SEGMENTATION ALGORITHM
1. Get Mask
1

a. Get the Green channel of the connective tissue
part, retrieved from the epithelium segmentation
phase, Channel
1
.
b. Enhance the contrast of the grayscale channel
Channel
1_1
by transforming the values using contrast-
limited adaptive histogram equalization (CLAHE)
[28], Channel
1_1
.
c. Apply 3 3 median filter for removing the noise [19]
in the Channel
1_1
without harming edges, Channel
1_2
.
d. Convert Channel
1_2
to black & white with 0.75
threshold value, Mask
1
.
2. Get Mask
2

a. Get the optical density of the Haematoxylin as an 8
bit grayscale image, by colour deconvolution [23] of
the connective tissue part, retrieved from the
epithelium segmentation phase, Channel
2
.
b. Adjust intensity values in the grayscale Channel
2
, to
map it to new values in such that 1% of data is
saturated at low and high intensities of the Channel
2
.
This increases the contrast of the output channel,
Channel
2_1
. [15].
c. Apply 2-D adaptive filtering with a 5 5 window size
over Channel
2_1
, Channel
2_2
.
d. Convert Channel
2_2
to black & white by applying
global thresholding algorithm using the value got
from Otsu's method [20] multiplied by 1.1, and then
invert it, Mask
2
.
3. Get Final Mask by excluding Mask
1
and Mask
2
from the
original connective tissue area, Mask
3
.
4. Finally we get the percentage of all objects in Mask
3
.

D. Classification
After extracting all the features, we used Feedforward
Backpropagation Artificial Neural Networks in the
classification process.
We used 11 inputs; 3 clinical data (Plaque Index, Pocket
Depth, and Attachment Level) and 8 extracted features
(Epithelium Percentage, Epithelium Nuclei Count &
Percentage, Connective Tissue Nuclei Count & Percentage,
Dilated Blood Vessels Count & Percentage, and Collagen
Fiber Percentage), 10 hidden neural networks, and 3 outputs
representing our 3 diseases groups (Gingival Enlargement,
Chronic Gingivitis, and Chronic Periodontitis); the used neural
network model is shown in Figure 8.


(A)

(C)

(D)
Figure 7 Collagen Fiber (A) Sample of connective tissue image.
(B) Mask1 + Mask2. (C) Final Mask (Collagen Fiber).
We divided our samples randomly into 3 groups, 20
samples for the training set which were presented to the neural
network during training, and the network was adjusted
according to its error, 5 samples for the validation set which
were used to measure network generalization, and to halt
training when generalization stops improving, and the last 5
samples for the testing set which had no effect on training so
providing an independent measure of network performance
during and after training.
We wrote some algorithm to make sure that each data set
contains a relative number of study cases for each disease
group. We used scaled conjugate gradient backpropagation
function for network training. Also, we used mean squared
error as our performance function.

Figure 8 Neural Network Model
IV. RESULTS AND DISCUSSION
For background removal preprocessing phase, we applied
both fully-automated and semi-automated preprocessing
algorithms described in Pseudocode I and Pseudocode II
respectively. We considered the semi-automated results as the
ground truth, and compared the results of the fully-automated
algorithm to it; over our 33 samples, we achieved a sensitivity
of 68.42% and specificity of 98.56%. 6 samples of our
samples actually contain no background at all, and when
removing their results from our statistics, we achieved a
sensitivity of 83.33% and specificity of 98.66% over our
remaining 27 samples.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

46 | P a g e
www.ijacsa.thesai.org
For epithelium segmentation phase, we applied the
algorithm described in Pseudocode III over our date, and since
the results of this step has to be accurate since it is used in all
later phases, we also implemented a manual way for moving
parts from the epithelium part to the connective tissue part or
vice versa. We considered the results after manual processing
the ground truth, and found that our algorithm achieved a
sensitivity of 84.99% and specificity of 88.40% over our 33
samples dataset, below are some statistics about most notable
results:
16 Samples (48.5%) give Sensitivity >= 95%, 15 Samples
(45.5%) give Specificity >= 95%, and 11 Samples
(33.3%) give Sensitivity & Specificity >= 95%.
16 Samples (48.5%) give Sensitivity >= 90%, 20 Samples
(60.6%) give Specificity >= 90%, and 13 Samples
(39.4%) give Sensitivity & Specificity >= 90%.
5 Samples (15.2%) give Sensitivity <= 60%, 2 Samples
(6.1%) give Specificity <= 60%, and 2 Samples (6.1%)
give Sensitivity & Specificity <= 60%.
For nuclei in epithelium segmentation phase, we applied
the algorithm described in Pseudocode IV over our data; we
found that optical density of the Haematoxylin stain retrieved
through colour deconvolution [23] is the best channel for
nuclei segmentation in 24 samples (73%) while the RGB Red
channel is the best for the remaining 9 samples (27%).
Information regarding average (count, area, and percentage) of
nuclei found in the epithelium for each disease group can be
found in Table I.
TABLE I. EPITHELIUM NUCLEI SEGMENTATION STATISTICS.
Group
No. of
Samples
Avg.
Count
Avg.
Area
Avg.
Percentage
Normal 1 826 117887 10.31
Gingival Enlargement 16 916 108013 09.46
Chronic Gingivitis 7 1359 131323 12.41
Chronic Periodontitis 7 1020 125925 11.96
Aggressive Periodontitis 2 1284 125068 10.44

For nuclei in connective tissue segmentation phase, we
applied the algorithm described in Pseudocode IV over our
data; Information regarding average (count, area, and
percentage) of nuclei found in connective tissue for each
disease group can be found in Table II.
TABLE II. CONNECTIVE TISSUE NUCLEI SEGMENTATION STATISTICS.
Group
No. of
Samples
Avg.
Count
Avg.
Area
Avg.
Percentage
Normal 1 413 27722 03.65
Gingival Enlargement 16 409 22835 03.07
Chronic Gingivitis 7 438 26042 03.06
Chronic Periodontitis 7 510 30459 03.45
Aggressive Periodontitis 2 444 19670 02.92

For dilated blood vessels segmentation phase, we found
the following statistics that are described in Table III.
TABLE III. DILATED BLOOD VESSELS SEGMENTATION STATISTICS.
Group
No. of
Samples
Avg.
Count
Avg.
Area
Avg.
Percentage
Normal 1 92 64005 08.29
Gingival Enlargement 16 94 75001 10.57
Chronic Gingivitis 7 77 68666 08.90
Chronic Periodontitis 7 82 65443 08.53
Aggressive Periodontitis 2 117 84809 12.61
For collagen fiber segmentation phase, we applied the
algorithm described in Pseudocode VI over our data;
Information regarding average (area, and percentage) of
collagen fibers for each disease group can be found in Table
IV.
TABLE IV. COLLAGEN FIBER SEGMENTATION STATISTICS.
Group
No. of
Samples
Avg.
Area
Avg.
Percentage
Normal 1 347325 45.72
Gingival Enlargement 16 423448 58.54
Chronic Gingivitis 7 383753 51.01
Chronic Periodontitis 7 404528 51.03
Aggressive Periodontitis 2 373231 56.24
For our classification results; Figure 9 shows our neural
network performance, Figure 10 0shows the confusion matrix
for the training dataset, validation dataset, test dataset, & the
whole dataset, while 0Figure 11 shows Receiver Operating
Characteristic ROC for them.

Figure 9 Neural Network Training Performance
I. CONCLUSION AND FUTURE SCOPE
An automated system has been developed for classification
of periodontal diseases using H&E stained microscopic
images of the tissues around the affected teeth along with
clinical data. The epithelium percentage of the whole tissue,
count & percentage of nuclei in epithelium & connective
tissue, dilated blood vessels count & percentage, and collagen
fiber percentage are used as features during the classification
process which is done using Feedforward Backpropagation
Artificial Neural Networks. It was found that using these
mixed features together achieve a more accurate classification
results than using only clinical data or H&E stained images
extracted features.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

47 | P a g e
www.ijacsa.thesai.org

Figure 10 Neural Network Training Confusion Matrix

Figure 11 Neural Network Training Receiver Operating Characteristic ROC
We suggest the following as a future work for this study:
1) Get a larger dataset of diseases, to provide better training
and testing of the proposed system, also get test cases in
the Aggressive Periodontitis group to include them in the
training and testing phases of the program to make it as
generic as possible for all periodontal diseases.
2) Enhance the fully-automated background removal
algorithm found in Pseudocode I, to achieve similar
results to the semi-automated one found in Pseudocode II.
Also, enhance the epithelium segmentation algorithm
found in Pseudocode III, to achieve higher Specificity and
sensitivity results.
3) Try to find more generic algorithms' parameters, to
decrease the number of study cases that we had to alter its
parameters to get better results; as shown in the footnotes
of Pseudocode III and Pseudocode IV.
REFERENCES
[1] D. V, Prapulla, P. B. Sujatha, and A. R. Pradeep,
"Gingival Crevicular Fluid VEGF Levels in Periodontal
Health and Disease," Journal of Periodontology, vol. 78,
no. 9, pp. 1783-1787, September 2007.
[2] C. B. Wiebe and E. E. Putnins, "The periodontal disease
classification system of the American Academy of
Periodontology--an update," J Can Dent Assoc., vol. 66,
no. 11, pp. 594-597, December 2000.
[3] G. C. Armitage, "Development of a classification system
for periodontal diseases and conditions," Annals of
Periodontology, vol. 4, pp. 1-6, December 1999.
[4] (2012, January) Wikipedia, the free encyclopedia.
[Online]. http://en.wikipedia.org/wiki/Periodontitis
[5] L. J. Brown and H. Le, "Prevalence, extent, severity and
progression of periodontal disease," Periodontol 2000,
vol. 2, pp. 57-71, June 1993.
[6] G. C. Armitage, "Periodontal diseases: diagnosis," Annals
of Periodontology, vol. 1, no. 1, pp. 37-215, 1996.
[7] P. N. Papapanou, "Periodontal Diseases: Epidemiology,"
Annals of Periodontology, vol. 1, no. 1, pp. 1-36,
November 1996.
[8] Le H., Anerud A., Boysen H., and Morrison E., "Natural
history of periodontal disease in man. Rapid, moderate
and no loss of attachment in Sri Lankan laborers 14 to 46
years of age," J Clin Periodontol, vol. 13, no. 5, pp. 431-
445, May 1986.
[9] R. Attstrm and U. van der Velden, "Consensus report
(epidemiology)," in Proceedings of the 1st European
Workshop on Periodontics, 1993, London, 1994, pp. 120-
126.
[10] H. LOE and J. SILNESS, "Periodontal disease in
pregnancy. I. Prevalence and severity," Acta Odontol
Scadinavia, vol. 21, pp. 533-551, December 1963.
[11] G. Landini and I. E. Othman, "Estimation of tissue layer
level by sequential morphological," Journal of
Microscopy, vol. 209, pt 2, pp. 118 125, February 2003.
[12] P. Soille, Morphological Image Analysis: Principles and
Applications, 1st ed.: Springer-Verlag, 1999, pp. 164-
165.
[13] R. Adams, "Radial Decomposition of Discs and Spheres,"
Computer Vision, Graphics, and Image Processing:
Graphical Models and Image Processing, vol. 55, no. 5,
pp. 325332, September 1993.
[14] R. Jones and P. Soille, "Periodic lines: Definition,
cascades, and application to granulometrie," Pattern
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012

48 | P a g e
www.ijacsa.thesai.org
Recognition Letters, vol. 17, pp. 10571063, 1996.
[15] M.M. Sami, M. Saito, H. Kikuchi, and T. Saku, "A
computer-aided distinction of borderline grades of oral
cancer," in 16th IEEE International Conference on Image
Processing (ICIP), Cairo, 2009, pp. 4205 - 4208.
[16] R. Abu Eida and G. Landini, "Quantification of the
Global and Local Complexity of the Epithelial-
Connective Tissue Interface of Normal, Dysplastic, and
Neoplastic Oral Mucosae Using Digital Imaging,"
Pathology - Research and Practice, vol. 199, no. 7, pp.
475-482, 2003.
[17] M. Pal et al., "Quantitative dimensions of
histopathological attributes and status of GSTM1GSTT1
in oral submucous fibrosis," Tissue and Cell, vol. 40, no.
6, pp. 425-435, December 2008.
[18] J. Maeda, S. Novianto, S. Saga, Y. Suzuki, and V. V.
Anh, "Rough and accurate segmentation of natural
images using fuzzy region-growing algorithm," in IEEE
Proceedings Image Processing. ICIP 99, vol. 3, Kobe,
1999, pp. 227-231.
[19] R. C. Gonzalez and R. E. Woods, Digital Image
Processing, 3rd ed.: Prentice Hall, 2008.
[20] N. Otsu, "A Threshold Selection Method from Gray-
Level Histograms," IEEE Transactions on Systems, Man,
and Cybernetics, vol. 9, no. 1, pp. 62-66, 1979.
[21] G. Landini and I. E. Othman, "Quantification of Local
Architecture Changes Associated with Neoplastic
Progression in Oral Epithelium using Graph Theory," in
Fractals in Biology and Medicine.: Birkhuser Basel,
2005, vol. 4, pp. 193-201.
[22] G. Landini, "Quantitative analysis of the epithelial lining
architecture in radicular cysts and odontogenic
keratocysts," Head & Face Medicine, vol. 2, 2006.
[23] A. C. Ruifrok and D. A. Johnston, "Quantification of
histological staining by color deconvolution," Anal Quant
Cytol Histol, vol. 23, pp. 291-299, 2001.
[24] Y. Al-Kofahi, W. Lassoued, and W. Lee, "Improved
Automatic Detection and Segmentation of Cell Nuclei in
Histopathology Images," IEEE transactions on
biomedical engineering, vol. 57, no. 4, pp. 841-852, April
2010.
[25] M. Park et al., "Automatic cell segmentation in
microscopic color images using ellipse fitting and
watershed," in The 201 0 IEEEIICME International
Conference on Complex Medical Engineering, Gold
Coast, Australia, 2010, pp. 69-74.
[26] S. DiCataldo, E. Ficarra, A. Acquaviva, and E. Macii,
"Achieving the way for automated segmentation of nuclei
in cancer tissue images through morphology-based
approach - A quantitative evaluation," Computerized
Medical Imaging and Graphics, vol. 34, pp. 453461,
2010.
[27] P. Soille, Morphological Image Analysis: Principles and
Applications, 1st ed.: Springer-Verlag, 1999, pp. 170-
171.
[28] K. Zuiderveld, "Contrast Limited Adaptive Histogram
Equalization," in Graphic Gems IV. San Diego:
Academic Press Professional, 1994, pp. 474485.

AUTHORS PROFILE
Aliaa A. A. Youssif, professor of computer science,
Faculty of Computers and Information, Helwan
University. Cairo, Egypt. She received her B.Sc and MSc.
degree in telecommunications and electronics engineering
from Helwan University. Prof. A. Youssif received the
PhD degree in computer science from Helwan University
in 2000. She was a visiting professor at George
Washington University (Washington DC, USA) in 2005.
She was also a visiting professor at Cardiff University in
UK (2008). She is currently vice dean for the faculty of Computers and
Information, Helwan University. Her fields of interest include pattern
recognition, AI researches, and medical imaging. She published more than 40
papers in different fields. She can be reached at [email protected]
Abeer Saad Gawish, professor of Oral Medicine, Periodontology, Oral
Diagnosis and Radiology, Faculty of Dental Medicine
(Girls), Al-Azhar University. Cairo, Egypt. She
received her B.Sc degree in Oral and Dental Medicine
from Alexandria University. Prof. A. Gawish received
the MSc. And the PhD degree Oral and Dental Medicine
from Cairo University in 2000. She was a visiting
professor at Pan Arab Association of Oral
&Maxillofacial Surgeon KSA 2010. She is currently
vice dean for the faculty of Oral and Dental Medicine,
Sinai University. Her fields of interest include periodontal medicine, genetics
in periodontal medicine, implantology. She published more than 37 papers in
different fields. She can be reached at [email protected]
Mohammed Elsaid Moussa received his BSc.
degree in computer science from Faculty of Computers
and Information, Helwan University, Cairo, Egypt in
2005. He is currently working as a teacher assistant at
the same faculty, and a masters degree student under
the supervision of prof. Aliaa A. A.Youssif and prof.
Abeer Saad Gawish. His areas of interests include image
processing, medical imaging, cloud computing, and
programming languages. He can be reached at
[email protected]

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
49 | P a g e
www.ijacsa.thesai.org
Communication and migration of an embeddable
mobile agent platform supporting runtime code
mobility

Mohamed BAHAJ
Department of Mathematics and Computer Science,
Universit Hassan 1er, FSTS, LABO LITEN
Settat, Morocco
Khaoula ADDAKIRI
Department of Mathematics and Computer Science,
Universit Hassan 1er, FSTS, LABO LITEN
Settat, Morocco
Noreddine GHERABI
Department of Mathematics and Computer Science,
Universit Hassan 1er, FSTS, LABO LITEN
Settat, Morocco


AbstractIn this paper we present the design and the
implementation of Mobile-C, an IEEE Foundation for Intelligent
Physical Agents (FIPA) compliant agent platform for mobile
C/C++ agents. Such compliance ensures the interoperability
between a Mobile-C agent and other agents from heterogeneous
FIPA compliant mobile agent platforms. Also, the Mobile-C
library was designed to support synchronization in order to
protect shared resources and provide a way of deterministically
timing the execution of mobile agents and threads. The new
contribution of this work is to combine the mechanisms of agent
migration and their synchronization.
Keywords- Mobile agent; Mobile agent platform;Agent
communication.
I. INTRODUCTION
Mobile agent is a design program with a persistent identity
which migrates in the network and communicates with its
environment and other agent [1]. It has been applied to a
variety of distributed applications, such as manufacturing [2-4],
electronic commerce [5-7], network management [8, 9], health
care [10], and entertainment [11]. During the execution, mobile
agents can be dynamically created and sent to the destination
systems to perform tasks with the up-to-date code. The
mobility allows mobile agent to migrate from one host to
another in the network and provides a several applications with
flexibility and adaptability that are both able to satisfy the
requirement and condition in a distributed environment.
The importance of Mobile agent technology in the design of
distributed applications on the web has led the OMG (Object
Management Group) to define the specifications MASIF
(Mobile Agent System Interoperability Facility) for
interoperability between different systems to mobile agents.
Another effort is made by FIPA (Foundation for Intelligent
Physical Agents) to specify the architecture and also the
semantics of communication between mobile agents.
The majority of mobile agent platforms in use are Java-
oriented. Multiple mobile agent platforms supporting Java
mobile agent code include Mole [12], Aglets [13], Concordia
[14], JADE [15], and Agents [16]. Using a standard language
like the mobile agent code language that provides both high-
level and low-level functionalities is a good choice to treat with
the large number of distributed applications. The choice of
C/C++ is a proper for a mobile agent code language because
its provides powerful functions in terms of memory access.
Furthermore, C is a language which can easily interface with a
variety of low-level hardware devices. Ara [17, 18] and
TACOMA [19] are two mobile agent platforms supporting C
mobile agent code, while Ara also supports C++. Mobile agent
code is compiled as byte code [20] and machine code [21] for
execution in both Ara and TACOMA, respectively.
Mobile-C [22-25] was originally developed as a stand-
alone, IEEE FIPA compliant mobile agent platform to
accommodate applications where low-level hardware is
involved and embedded systems [26]. Most of the systems are
written in C/C++; Mobile-C chose C/C++ as the mobile agent
language because C has an advantage for easy interfacing with
control programs and underlying hardware. Additionally,
Mobile-C integrated an embeddable C/C++ interpreter, Ch [27-
29], as the Agent Execution Engine (AEE) in order to run the
mobile agent code. The migration of mobile agent in Mobile-C
is achieved through FIPA agent communication language
(ACL) messages. Using FIPA ACL messages for agent
migration in FIPA compliant agent systems simplifies agent
platform, reduces development effort and easily achieves inter-
platform migration through well-designed communication
mechanisms provided in the agent platform. Messages for
agent communication and migration are presented in FIPA
ACL and encoded in XML. Also, the Mobile-C library was
designed to support synchronization [26] in order to protect
shared resources and provide a way of deterministically timing
the execution of mobile agents and threads.
In this paper we present the Mobile-C library that can
embed Mobile-C into any C/C++ programs to facilitate the
design of mobile agent-based applications, also the possibility
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
50 | P a g e
www.ijacsa.thesai.org
to combine the migration of the mobile agent over the network
and the synchronization mechanism existing in Mobile-C.
Mobile agents are an application that can control the agent
platform, its modules and other mobile agents, as well as
smoothly interface with a variety of low-level hardware
devices. Using FIPA ACL messages for agent migration in
FIPA compliant agent systems simplifies agent platform since
both agent communication and migration can be achieved
through the same communication mechanism provided in the
agent platform. Flexible synchronization mechanisms have
been added for execution and interaction of several mobile
agents. This paper proposes a new approach by combining two
concept migration and synchronization supports in Mobile-C.
The remainder of the article is structured as follows. In
section 2, we present the concept and the properties of mobile
agent. Section 3 introduces the architecture of Mobile-C.
Section 3 presents the migration of mobile agent over the
network from multiple hosts. Section 4 describes the program
structure and implementation of the component of agency.
Section 5 gives an example of a mobile agent that migrates
from hosts via mobile agent messages and illustrates the
synchronization support in Mobile-C.
II. MOBILE AGENTS
An agent is defined as person whos acting on behalf of
other people [30]. In the context of computer science, mobile
agent is considered as an entity that moves from one machine
to another in the network to perform certain tasks on behalf of
the user [31].
Mobile agents have the following properties which
distinguish them from other programs [32]:
Adaptability - Mobility of agent required to learn about
user's behavior and adapt it to suit the user. Indeed, to evolve
adequately the differences between heterogeneous systems,
the agent must be able to adopt the changes during the
execution.
Autonomy- Mobile agent must be able to make his own
decision to be performed to achieve the user's tasks, also he
must be able to migrate from one machine to another in the
network and execute the user's tasks.
Communication - Mobile agent must have the ability to
communicate with others agents of the system in order to
exchange information and benefit from the knowledge and
expertise of other agents.
Mobility- Mobile agent has the ability to move from one host
to another, either by moving the agent's code or by serializing
both code and state to allow the agent to continue the
execution in a new context.
Persistence - A persistent agent it will be able to retain the
knowledge and state over extended period of time to be
accessed later on. Once the mobile agent is set up, it is not
dependent on system that has been initiated and it is
automatically recovered when the agent is terminated or
when it is flushed from memory to the database.

III. THE ARCHITECTURE OF MOBILE-C
The system of mobile-C is shown in figure1. Agencies are
the major building blocks of the system and abode in each node
of a network system in order to support Stationary Agents (SA)
and Mobile Agents (MA) at runtime. They serve for locating
and messaging agents, moving mobile agents, collecting
knowledge about other agents and providing several places
where the agent can be run. The core of an agency provides
local service for agents and proxies remote agencies. The
principle of an agency and their functionalities can be
summarized as follows [33]:
Agent Management system (AMS): The AMS manages the
life cycle of agents in the system. It relates the creation,
authentication, registration, deletion,execution, migration and
persistence of agents. AMS is also responsible for receiving
and dispatching mobile agents .Each agent must register
with an AMS in order to get a valid AID.
Agent Communication Channel (ACC): The ACC
forwardsmessages between local and remote entities. The
interaction and coordination of mobile agents and host
systems can be performed through agent communication
language (ACL).
Agent Security Manager (ASM): The ASM is responsible for
protection of access for platform and infrastructure.
Directory Facilitator: DF serves yellow page services. Agents
in the system can register their services with DF for providing
to the community. They can also look up required services
with DF.
Agent Execution Engine (AEE): AEE serves as the
workhorsefor mobile agents. Mobile agents must reside
inside an engine to execute. AEE has to be platform
independent in order to support a mobile agent executingin a
heterogeneous network.
SA
Agent Communication Channel(ACC)
Agent
Security
Manager
(ASM)
Agent
Management
System
(AMS)
Directory
Facilitator
(DF)
Agent Execution
Engine(AEE)
MA

Figure 1. The system architecture of agencies in Mobile-C.
IV. MOBILE AGENT MIGRATION
Mobile agent is a software agent who is able to migrate
from one host to another over the network and resume the
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
51 | P a g e
www.ijacsa.thesai.org
execution in the new host. The migration and the execution of
mobile agents are supported by a mobile agent system. In
previous studies, Chen et al. have developed a mobile agent
system called Mobile-C.The Mobile-
C supports weak migration. The task of a mobile agent can
be divided into several subtasks whichcan be executed
in different hosts and listed in a list of tasks as shown in figure
2. The task list can be modified by adding
new subtasks and new conditions. Changing dynamically the
task list improves the flexibility of a mobile agent. Thus,
once we start the execution of a subtask in a host, the mobile
agent cannot move until the end of execution.

Figure 2. Agent migration based on a task list and a task progress pointer.
Mobile agent migration is achieved through ACL mobile
agent messages encoded to XML, which convey mobile agents
as the content of a message. Mobile agent message containsthe
data state and the code of an agent. The data state of mobile
agent include general information about mobile agent as agent
name, agent owner and agent home , also the tasks that mobile
agent will performed on destination hosts. The data state and
code will be wrapping up into an ACL message and transmitted
to a remote host trough Agent Communication Channel.
Mobile agent migration based on ACL messages is simple and
effective for agent migration in FIPA compliant systems
because these systems have mandatory mechanisms for
message communication, transmission and procession.
V. THE PROGRAM STRUCTURE AND IMPLEMENTATION OF
COMPONENT OF AGENCY
An agency is a principle program running in each node of
the network [23]. When the execution of an agency is started,
the system is initialized and threads are created for all of the
components in the agent platform.
After the initialization of the system, the agency waits for
defined events. When the agency receives a request to run a
mobile agent, it creates a new thread and embeds an
Embeddable C/C++ Interpreter Ch into the thread for
executing mobile agent code. After the mobile agent migrates
to the other hosts, this thread is terminated automatically
(figure 3). If the agency receives a system termination request,
the execution of agent platform and the system itself will be
finished. In the current implementation, each mobile agent runs
in an embeddable Ch inside its own thread.

Figure 3. The program structure of an agency
According to the FIPA specifications, each agency should
provide mechanisms to receive and send messages. This
requirement is satisfied by three components: listening thread,
connecting thread and ACC processing thread as shown in
Figure 4. The listening thread serves to listen for client
connections. When a new connection client is accepted, it will
be added to the connection list. Also, the connecting thread is
responsible for making connections with other hosts. The ACC
processing thread processes the lists of client connections and
requests for connecting remote hosts. The ACC facilitates
remote agent to agent communication and remote agent
platform to agent platform communication via ACL messages.
Remote horizontal communication in Mobile-C is implemented
on top of TCP/IP and the transport protocol uses HTTP
(HyperText Transfer Protocol).

Agent Platform
AMS ASM DF
ACC
Listening
ACC
Processing
Connecting
T
h
r
e
a
d

1
T
h
r
e
a
d

2
T
h
r
e
a
d

3 Thread 4 Thread 5 Thread 6

Figure 4. The multi-thread implementation of an agent platform
Task 1
Task 2
Task 3
.
.
Task n
Task
list
Task
progress
Agency 1
MA
Task 1
Task 2
Task 3
.
.
Task n
Task
list
Task
progress
Agency 2
MA
Migrate
M
i
g
r
a
t
e
Task 1
Task 2
Task 3
.
.
Task n
Task
list
Task
progress
Agency n
MA
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
52 | P a g e
www.ijacsa.thesai.org
VI. SYNCHRONIZATION SUPPORT IN MOBILE-C
Amongproperties of mobile agents is the ability to
immigrate to perform tasks that exist in the remote host. The
purpose of this example is to use an embeddable mobile agent
system to protect shared resources by used the synchronization
and combines it with the migration of mobile agents from
hosts.
The Mobile-C library allows synchronization via mutex.
The mutex is a program that allows multiple threads to share
the same resource, but not simultaneously.
The example below demonstrates the capabilityof a Mobile-
C mutex to protect a resource that may be shared between two
or more agents in several hosts. As shown in program 1, a
mobile agent is transferred by an agency in the host fsts1 visits
remote host fsts2 then host fsts3 .The mobile agent message is
represented in Extensible Markup Language (XML), it contains
information of the mobile agent and tasks that will be
performed on destination hosts. The general information of a
mobile agent contains:agent name, agent owner, and the home
of the agent. The task information for example the statement
<TASK task=2 num=0> shows that this mobile agent has
two tasks to perform and no task has been done yet.The DATA
element overall information about the number of element, the
name of the tasks return variable, the completeness of the task
and the host to perform the task.The sub-element DATA
ELEMENT contains the return data from the execution task
and the sub-element AGENT_CODE contains a C program
which will be executed in remote host.
<NAME>mobileagent</NAME>
<OWNER>fsts</OWNER>
<HOME>fsts1.fsts.ac.ma:5125 </HOME>
<TASK task= "2" num= "0">
<DATA number_of_elements ="0" name = "results_fsts2"
complete = "0" server = "fsts2.fsts.ac.ma:5138">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
Mobile agent code on fsts2
</AGENT_CODE>
</DATA>
<DATA number_of_elements ="0" name = "results_fsts3"
complete = "0" server = "fsts3.fsts.ac.ma:5135">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
Mobile agent code on fsts3
</AGENT_CODE>
</DATA>
</TASK>
Program 1: The content of the mobile agent message from the host fsts1 to
fsts2 and to host fsts3
As shown in Program2, the mobile agent 1 MA1
initialize a mutex with an ID 55 via the function mc_SyncInit()
and defines two functions, SetN1() and GetN1() in the host
fsts2 . After visiting this host, the mobile agent 1 MA1 visits
the host fsts3 and defines also two functions, SetN2() and
GetN2().The result obtained from the host fsts2 is sent to the
host fsts3 and the return data will be included in the sub-
element DATA_ELEMENT.
<NAME>MA1</NAME>
<OWNER>fsts</OWNER>
<HOME> fsts1.fsts.ac.ma:5125 </HOME>
<TASK task= "2" num= "0">
<DATA number_of_elements ="0" name = "results_fsts2"
complete = "0" server = "fsts2.fsts.ac.ma:5138">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
int N1;
int main () {
intmutex_id=55;
mc_SyncInit(mutex_id);
return 0;
}
void SetN1(inti){
N1 +=i;
if(N1 > 1000){
N1=0;
}
}
intGetN(){
return N
}
]]>
</AGENT_CODE>
</DATA>
<DATA number_of_elements ="0" name = "results_fsts3"
complete = "0" server = "fsts3.fsts.ac.ma:5135">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
int N2;
int main () {
intmutex_id=55;
mc_SyncInit(mutex_id);
return 0;
}
void SetN2(inti){
N2*=i;
if(N2 > 1000){
N2=0;
}
}
intGetN(){
return N2
}
]]>
</AGENT_CODE>
</DATA>
</TASK>
Program2 .A mobile agent that contains a global variable and defines function
to access the global variables
As shown in Program 3, the task of the mobile agent
2MA2 is to perform the operation setting the variable. The
operation includes locking the mutexthrough the function
mc_MutexLock(), setting the global variable by calling
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
53 | P a g e
www.ijacsa.thesai.org
SetN1() from the host fsts2 and SetN2()from the host fsts3via
the function mc_CallAgentFunc(), and unlocking the mutex via
the function mc_MutexUnlock().
<NAME>MA2</NAME>
<OWNER>fsts</OWNER>
<HOME> fsts1.fsts.ac.ma:5125 </HOME>
<TASK task= "2" num= "0">
<DATA number_of_elements ="0" name = "results_fsts2"
complete = "0" server = "fsts2.fsts.ac.ma:5138">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
#include <stdio.h>
int main (){
MCAgent_t agent;
inti= 0,mutex_id =55 ,retval;
agent= mc_FindAgentByName("MA1");
wile(1){
mc_MutexLock(mutex_id);
mc_CallAgentFunc(agent,"SetN1",NULL,i);
printf("N1:%d\n",retval);
mc_MutexUnlock(mutex_id)
i++;
if(i==20) {
i=0;
}
}
return 0;
]]>
</AGENT_CODE>
</DATA>
<DATA number_of_elements ="0" name = "results_fsts3"
complete = "0" server = "fsts3.fsts.ac.ma:5135">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
#include <stdio.h>
int main (){
MCAgent_t agent;
inti= 0,mutex_id =55 ,retval;
agent= mc_FindAgentByName("MA1");
wile(1){
mc_MutexLock(mutex_id);
mc_CallAgentFunc(agent,"SetN2",NULL,i);
printf("N2:%d\n",retval);
mc_MutexUnlock(mutex_id)
i++;
if(i==20) {
i=0;
}
}
return 0;
]]>
</AGENT_CODE>
</DATA>
</TASK>
Program3. A mobile agent that sets a variable
Likewise, as shown in Program 4, the task of the mobile
agent 3MA3 is locks the mutex, get the global variable, and
unlocks the mutex from both the host fsts2 and fsts3
<NAME>MA3</NAME>
<OWNER>fsts</OWNER>
<HOME> fsts1.fsts.ac.ma:5125 </HOME>
<TASK task= "2" num= "0">
<DATA number_of_elements ="0" name = "results_fsts2"
complete = "0" server = "fsts2.fsts.ac.ma:5138">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
#include <stdio.h>
int main (){
MCAgent_t agent;
inti ,mutex_id =55 ,retval;
agent= mc_FindAgentByName("MA1");
mc_MutexLock(mutex_id);
mc_CallAgentFunc(agent,"GetN",&retval,NULL);
printf("N1:%d\n",retval);
mc_MutexUnlock(mutex_id)
return 0;
]]>
</AGENT_CODE>
</AGENT_CODE></DATA>
<DATA number_of_elements ="0" name = "results_fsts3"
complete = "0" server = "fsts3.fsts.ac.ma:5135">
<DATA_ELEMENT></ DATA_ELEMENT >
<AGENT_CODE>
<![CDATA[
#include <stdio.h>
int main (){
MCAgent_t agent;
inti ,mutex_id =55 ,retval;
agent= mc_FindAgentByName("MA1");
mc_MutexLock(mutex_id);
mc_CallAgentFunc(agent,"GetN",&retval,NULL);
printf("N2:%d\n",retval);
mc_MutexUnlock(mutex_id)
return 0;
]]>
</AGENT_CODE>
</DATA>
</TASK>
Program4. A mobile agent that gets a variable from another agent.
The results of the mobile agent 1MA1, mobile agent 2
MA2 and mobile agent3 Ma3 obtained from both the host
fsts2 and fts3 are send back to the home agency fsts.
VII. CONCLUSION
In this work we present the design and implementation of
an IEEE FIPA compliant agent platform, Mobile-C. Mobile-C
integrates an embeddable C/C++ interpreterChinto the
platform as a mobile agent execution engine in order to support
mobile agent. The migration of mobile agent is achieved
through ACL messages. Mobile agents, including both its data
state and code, are transported to a remote agent platform via
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
54 | P a g e
www.ijacsa.thesai.org
ACL messages which is encoded in XML, and the execution of
mobile agents is resumed by the task progress pointer. The
Mobile-C library supports the synchronization among mobile
agents and threads because the synchronization functions
protect shared resources and provide a way of deterministically
timing the execution of mobile agents and the migration to a
remote host.
In our future work, this framework will be tested and
extended in various types of industrial applications like e-
commerce and network management.
ACKNOWLEDGMENT
The authors thank the referees for valuable constructive
comments and suggestions which lead to a signicant
improvement of this paper.
REFERENCES

[1] D.Chess,C.Harrison, A.Kershenbaum,Mobile Agentes: Are They a
Good Idea?, IBM ResearchReport, 1998 [Online] Available :
http://www.research.ibm.com/iagentes/paps/mobile -idea.ps.
[2] W. Shen, D. Xue, D.H. Norrie, An agent-based manufacturing enterprise
infrastructure for distributed integrated intelligent manufacturing
systems, in: Proceedings of the 3rd International Conference on the
Practical Applications of Agents and Multi-Agent Systems (PAAM-98),
London, UK, 1998, pp. 533548.
[3] H. Wada, S. Okada, An autonomous agent approach for manufacturing
execution control systems, Integrated Computer-Aided Engineering 9 (3)
(2002) 251262.
[4] H.V.D. Parunak, A.D. Baker, S.J. Clark, The AARIA agent architecture:
from manufacturing requirements to agent-based system design,
Integrated Computer-Aided Engineering 8 (1) (2001) 4558.
[5] M. Yokoo, S. Fujita, Trends of internet auctions and agent-mediated web
commerce, New Generation Computing 19 (4) (2001) 369388.
[6] T. Sandholm, eMediator: a next generation electronic commerce server,
Computational Intelligence 18 (4) (2002) 656676.
[7] S.P.M. Choi, J. Liu, S. Chan, A genetic agent-based negotiation system,
Computer Networks: The International Journal of Computer and
Telecommunications Networking 37 (2) (2001) 195204.
[8] W.E. Chen, C. Hu, A mobile agent-based active network architecture for
intelligent network control, Information Sciences 141 (12) (2002) 335.
[9] L. Chou, K. Shen, K. Tang, C. Kao, Implementation of mobile-agent-
based network management systems for national broadband
experimental networks in Taiwan, Holonic and Multi-Agent Systems for
Manufacturing (Lecture Notes in Computer Science) 2744 (2003) 280
289.
[10] J. Huang, N.R. Jennings, J. Fox, Agent-based approach to health care
management, Applied Artificial Intelligence 9 (4) (1995) 401420.
[11] I. Noda, P. Stone, The RoboCup soccer server and CMUnited clients:
implemented infrastructure for MAS research, Autonomous Agents and
Multi-Agent Systems 7 (12) (2003) 101120.
[12] K.Straber, J.Baumann and F.Hohl.. Mole - A Java Based Mobile Agent
System. Institute for Parallel and Distributed Computer Systems,
University of Stuttgart,1997
[13] D. Lange, M.Oshima. Programming and Deploying Java Mobile Agents
with Aglets. Addison-Wesley: MA, 1998.
[14] D.Wong, N.Paciorek, T.Walsh, J.DiCelie, M.Young, B.Peet. Concordia:
An infrastructure for collaborating mobile agents. Proceedings of the
First International Workshop on Mobile Agents (MA97) (Lecture Notes
in Computer Science, vol. 1219). Springer: Berlin, 1997; 8697.
[15] F.Bellifemine, G.Caire, A.Poggi, G.Rimassa.JADE: A software
framework for developing multi-agent applications.Lessons learned.
Information and Software Technology 2008; 50(12):1021.
[16] R.Gray, G.Cybenko, D.Kotz,R.Peterson, D.Rus. DAgents: Applications
and performance of a mobile-agent system. SoftwarePractice and
Experience 2002; 32(6):543573.
[17] H.Peine. Run-time support for mobile code. PhD Dissertation,
Department of Computer Science, University of Kaiserslautern,
Germany, 2002.
[18] H.Peine .Application and programming experience with the Ara mobile
agent system. SoftwarePractice and Experience 2002; 32(6):515541.
[19] D.ohnansen, K.Lauvset, R.V.Renesse, F.B. Schneider, N.P. Sudmann,
K. Jacobsen. A TACOMA retrospective.SoftwarePractice and
Experience 2002; 32(6):605619.
[20] MACEMobile agent code environment. Available at:
http://wwwagss.informatik.uni-kl.de/Projekte/Ara/mace.html [last
modified 10 August 2004].
[21] N.P.Sudmann,D.Johansen. Adding mobility to non-mobile web robots.
Proceedings of the IEEE ICDCS00 Workshop on Knowledge Discovery
and Data Mining in the World-wide Web, Taipei, Taiwan, 2000; 7379.
[22] B.Chen, H.H.Cheng. A run-time support environment for mobile agents.
Proceedings of ASME/IEEE International Conference on Mechatronic
and Embedded Systems and Applications, No. DETC2005-85389, Long
Beach, CA, September 2005.
[23] B.Chen,H.H.Cheng,J.Palen. Mobile-C: A mobile agent platform for
mobile C/C++ agents. SoftwarePractice and Experience 2006;
36(15):17111733.
[24] B.Chen, D.Linz, H.H.Cheng. XML-based agent communication,
migration and computation in mobile agent systems. Journal of Systems
and Software 2008; 81(8):13641376.
[25] Mobile-C: A multi-agent platform for mobile C/C++ code. Available at:
http://www.mobilec.org.
[26] Y.C. Chou, D. Ko, H. H. Cheng, An embeddable mobile agent platform
supporting runtime code mobility,interaction and coordination of mobile
agents and host systems, Information and Software Technology 52
(2010) 185196
[27] H.H.Cheng. Scientific computing in the Ch programming language.
Scientific Programming 1993; 2(3):4975
[28] H.H.Cheng.Ch: A C/C++ interpreter for script computing. C/C++ Users
Journal 2006; 24(1):612.
[29] ChAn embeddable C/C++ interpreter. Available at:
http://www.softintegration.com .
[30] P.M .Reddy, Mobile Agents Intelligent Assistants on the Internet, July
2002
[31] A. Kaur , S.Kaur, Role of Mobile Agents In Mobile Computing,
Proceedings of National Conference on Challenges & Opportunities in
Information Technology (COIT-2007) RIMT-IET
[32] M. L. Griss, Software Agents as Next Generation Software Components,
Chapter 36 in Component-Based Software Engineering: Putting the
Pieces Together, Edited by George T. Heineman, Ph.D. & William
Councill, M.S., J.D., May 2001,Addison-Wesley
[33] B. Chen, H. H. Cheng, J.Palen, Integrating mobile agent technology with
multi-agent systems for distributed traffic detection and management
systems, Transportation Research Part C 17 (2009) 110
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
55 | P a g e
www.ijacsa.thesai.org
An Adaptive parameter free data mining approach for
healthcare application

Prof. Dipti Patil
#1
#
Asst. Professor, Computer Engg. Dept, MITCOE Pune,
India
Snehal Andhalkar
*3
#
Computer Engg. Dept, MITCOE Pune, India
Mayuri Gund
#5
#
Computer Engg. Dept, MITCOE Pune, India
Bhagyashree Agrawal
#2
#
Computer Engg. Dept, MITCOE Pune, India

Richa Biyani
#4
#
Computer Engg. Dept, MITCOE Pune, India

Dr. V.M.Wadhai
#6
#
Professor, Computer Engg. Dept, MITCOE Pune, India

AbstractIn todays world, healthcare is the most important
factor affecting human life. Due to heavy work load it is not
possible for personal healthcare. The proposed system acts as a
preventive measure for determining whether a person is fit or
unfit based on his/her historical and real time data by applying
clustering algorithms viz. K-means and D-stream. Both clustering
algorithms are applied on patients biomedical historical
database. To check the correctness of both the algorithms, we
apply them on patients current biomedical data. The Density-
based clustering algorithm i.e. the D-stream algorithm overcomes
drawbacks of K-means algorithm. By calculating their
performance measures we finally find out effectiveness and
efficiency of both the algorithms.
Keywords- Data stream mining; clustering; healthcare
applications; medical signal analysis.
I. INTRODUCTION
Today the health care industry is one of the largest
industries throughout the world. It includes thousands of
hospitals, clinics and other types of facilities which provide
primary, secondary & tertiary levels of care. The delivery of
health care services is the most visible part of any health care
system, both to users and the general public [2]. A health care
provider is an institution or person that provides preventive,
curative, promotional or rehabilitative health care services in a
systematic way to individuals, families or community. The
physiological signals such as SpO2, ABPsys, ABPdias, HR
affects persons health. In health care the data mining is more
popular and essential for all the healthcare applications. In
healthcare industry having the more amounts of data, but this
data have not been used properly for the application. In this
health care data is converted in to the useful purpose by using
the data mining techniques [1].
The data mining is the process of extracting or mining the
knowledge from the large amounts of data, database or any
other data base repositories. The main purpose of the data
mining is to find the hidden knowledge from the data base. In
health care industry, the data having some unwanted data,
missing values and noisy data. This unwanted data will be
removed by using preprocessing techniques in data mining.
Preprocessing is the process of removing noise, redundant data
and irrelevant data. After the preprocessing the data will be
used for some useful purpose. In recent years different
approaches are proposed to overcome the challenges of storing
and processing of fast and continuous streams of data.
Data stream can be conceived as a continuous and changing
sequence of data that continuously arrive at a system to store or
process. Traditional OLAP and data mining methods typically
require multiple scans of the data and are therefore infeasible
for stream data applications. Whereby data streams can be
produced in many fields, it is crucial to modify mining
techniques to fit data streams. Data stream mining has many
applications and is a hot research area[3]. Data stream mining
is the extraction of structures of knowledge that are represented
in the case of models and patterns of infinite streams of
information. These data stream mining can be used to form the
clusters of medical health data. This paper proposed two main
clustering algorithms namely, K-means algorithm and density
based clustering.
The K-means clustering algorithm is incompetent to find
clusters of arbitrary shapes and cannot handle outliers. Further,
they require the knowledge of k and user-specified time
window. To address these issues, D Stream, a framework for
clustering stream data using a density-based approach. The
algorithm uses an online component which maps each input
data record into a grid and an offline component which
computes the grid density and clusters the grids based on the
density. The algorithm adopts a density decaying technique to
capture the dynamic changes of a data stream. Exploiting the
intricate relationships between the decay factor, data density
and cluster structure, our algorithm can efficiently and
effectively generate and adjust the clusters in real time. Further,
a theoretically sound technique is developed to detect and
remove sporadic grids mapped to by outliers in order to
dramatically improve the space and time efficiency of the
system. The technique makes high-speed data stream clustering
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
56 | P a g e
www.ijacsa.thesai.org
feasible without degrading the clustering quality. The
experimental results show that our algorithm has superior
quality and efficiency, can find clusters of arbitrary shapes, and
can accurately recognize the evolving behaviors of real-time
data streams [4].
II. RELATED WORK
All Several health care projects are in full swing in different
universities and institutions, with the objective of providing
more and more assistance to the elderly. The application of
data clustering technique for fast retrieval of relevant
information from the medical databases lends itself into many
different perspectives.
Health Gear: a real-time wearable system for monitoring
and analyzing physiological signals [5] is a real-time wearable
system for monitoring, visualizing and analyzing physiological
signals. This system focused on an implementation of Health
Gear using a blood oximeter to monitor the user's blood oxygen
level and pulse while sleeping. The system also describes two
different algorithms for automatically detecting sleep apnea
events, and illustrates the performance of the overall system in
a sleep study with 20participants.
A Guided clustering Technique for Knowledge Discovery A
Case Study of Liver Disorder Dataset, [6] presents an
experiment based on clustering data mining technique to
discover hidden patterns in the dataset of liver disorder
patients. The system uses the SOM networks internal
parameters and k-means algorithm for finding out patterns in
the dataset. The research has shown that meaningful results can
be discovered from clustering techniques by letting a domain
expert specify the input constraints to the algorithm.
Intelligent Mobile Health Monitoring System (IMHMS),
[7] Author proposed the system which can provide medical
feedback to the patients through mobile devices based on the
biomedical and environmental data collected by deployed
sensors. The system uses the Wearable Wireless Body/Personal
Area Network for collecting data from patients, mining the
data, intelligently predicts patient's health status and provides
feedback to patients through their mobile devices.
The patients will participate in the health care process by
their mobile devices and thus can access their health
information from anywhere any time. But actual
implementation of data mining framework for decision support
system is not done.
Real-Time analysis of physiological data to support medical
applications [8], proposed a flexible framework to perform
real-time analysis of physiological data and to evaluate
peoples health conditions. Patient or disease-specific models
are built by means of data mining techniques. Models are
exploited to perform real time classification of physiological
signals and continuously assess a persons health conditions.
The proposed framework allows both instantaneous evaluation
and stream analysis over a sliding time window for
physiological data. But dynamic behavior of the physiological
signals is not analyzed also the framework is not suitable for
ECG type of signals.
Performance of Clustering Algorithms in Healthcare
Database [1], proposed a framework where they used the heart
attack prediction data for finding the performance of clustering
algorithm. In final result shows the performance of classifier
algorithm using prediction accuracy and the visualization of
cluster assignments shows the relation between the error and
the attributes. The comparison result shows that, the make
density based clusters having the highest prediction Accuracy.
III. METHOD
We present a framework that will perform clustering of
dataset available from medical database effective manner. The
flow of the system is depicted in Fig.1.


Figure 1.Flow of the system
The target is to cluster the patients records into different
groups with respect to the test report attributes which may help
the clinicians to diagnose the patients disease in efficient and
The evaluation steps are the following-
1) Data set collection
The data set contains 7 attributes, SpO2, ABPsys, ABPdias,
HR, heredity, obesity, cigarette smoking. These attributes are
the risk factors that can help in predicting the patients health
status. Attributes such as SpO2, ABPsys, ABPdias, HR can be
collected form MIMIC database [9] and the other attributes are
influenced by the persons behavior. These all attributes values
are discrete in nature .The dataset will be in preprocessed
format.
2) Model Building
In model building phase features of the available data will
be extracted and then clustering algorithm will be applied on
extracted features.
A. Feature Extraction
For each physiological signal x among the X monitored
vital signs, we extract the following features [8].
1) Offset
The offset feature measures the difference between the
current value x(t) and the moving average (i.e., mean value
over the time window). It aims at evaluating the difference
between the current value and the average conditions in the
recent past.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
57 | P a g e
www.ijacsa.thesai.org
2) Slope
The slope function evaluates the rate of the signal change.
Hence, it assesses short-term trends, where abrupt variations
may affect the patients health.
3) Dist
The dist feature measures the drift of the current signal
measurement from a given normality range. It is zero when the
measurement is inside the normality range.
B. Risk Components
The signal features contribute to the computation of the
following risk components.
1) Sharp changes
The z1 component aims at measuring the health risk
deriving from sharp changes in the signal (e.g., quick changes
in the blood pressure may cause fainting)
2) Long-term trends
The z2 component measures the risk deriving from the h
weighted offset over the time window. While z1 focuses on
quick changes, z2 evaluates long-term trends, as it is offset-
based.
3) Distance from normal behavior
The z3 component assesses the risk level given by the
distance of the signal from the normality range. A patient with
an instantaneous measurement outside the range may not be
critical, but her/his persistence in such conditions contributes to
the risk level
From above risk components, risk functions and global risk
components will be calculated. These values will be further
used in clustering algorithms as an input for cluster formation.
C. Cluster formation
The proposed flow of the system uses two algorithms K-
means and D-stream. The comparison between two clustering
algorithms will be performed using the above described
attributes.
K-MEANS ALGORITHM
1) The algorithm is composed of the following steps: [10]
Place K points into the space represented by the objects that are
being clustered. These points represent initial group centroids.
Assign each object to the group that has the closest centroid.
When all objects have been assigned, recalculate the positions
of the K centroids.
Repeat Steps 2 and 3 until the centroids no longer move. This
produces a separation of the objects into groups from which the
metric to be minimized can be calculated.
Figure 2. Algorithm K-means algorithm
D-STREAM ALGORITHM
The D-stream algorithm is explained as follows [4]
1.procedure D-Stream
2. Tc = 0;
3. Initialize an empty hash table grid list;
4. while data stream is active do
5. read record x = (x1, x2, , xd);
6. determine the density grid g that contains x;
7. if(g not in grid list) insert g to grid list;
8. update the characteristic vector of g;
9. if tc == gap then
10. call initial clustering(grid list);
11. end if
12. if tc mod gap == 0 then
13. detect and remove sporadic grids from grid list;
14. call adjust clustering(grid list);
15. end if
16. tc = tc + 1;
17. end while
18. end procedure
Figure 3: The overall process of D-Stream.
The overall architecture of D-Stream, which assumes a
discrete time step model, where the time stamp is labeled by
integers 0, 1, 2, , n, . D-Stream has an online
component and an offline component. The overall algorithm is
outlined in Figure 1.
For a data stream, at each time step, the online component
of D-Stream continuously reads a new data record, place the
multi-dimensional data into a corresponding discretized density
grid in the multi-dimensional space, and update the
characteristic vector of the density grid (Lines 5-8 of Figure 3).
The density grid and characteristic vector are to be described in
detail later. The offline component dynamically adjusts the
clusters every gap time steps, where gap is an integer
parameter. After the first gap, the algorithm generates the
initial cluster (Lines 9-11). Then, the algorithm periodically
removes sporadic grids and regulates the clusters (Lines 12-
15).
D-Stream partitions the multi-dimensional data space into
many density grids and forms clusters of these grids. This
concept is schematically illustrated in Figure 4.

Figure 4.Illustration of the use of density grid.
The input data has d dimensions, and each input data record
is defined within the space
... (1)
where Si is the definition space for the ith dimension.
In D-Stream, we partition the ddimensional space S into
density grids. Suppose for each dimension, its space Si,
I = 1, , d is divided into pi partitions as
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
58 | P a g e
www.ijacsa.thesai.org
-----(2)
then the data space S is partitioned into N density grids. For
a density grid g that is composed of

we denote it as
--------------(3)
A data record x = (x1, x2, , xd) can be mapped to a
density grid g(x) as follows:
where xi 2 Si,ji .
For each data record x, we assign it a density coefficient
which decreases with as x ages. In fact, if x arrives at time
tc, we define its time stamp T(x) = tc, and its density
coefficient D(x, t) at time t is
----------(4)
where (0, 1) is a constant called the decay factor.
Definition (Grid Density) For a grid g, at a given
time t, let E(g, t) be the set of data records that are map to
g at or before time t, its density D(g, t) is defined as the sum
of the density coefficients of all data records that mapped
to g. Namely, the density of g at t is:

1) procedure initial clustering (grid list)
2) update the density of all grids in grid list;
3) assign each dense grid to a distinct cluster;
4) label all other grids as NO CLASS;
5) repeat
6) for each cluster c
7) for each outside grid g of c
8) for each neighboring grid h of g
9) if (h belongs to cluster c)
10) if (|c| > |c|) label all grids in c as in c;
11) else label all grids in c as in c;
12) else if (h is transitional) label h as in c;
13) until no change in the cluster labels can be made
14) end procedure

Figure 3: The procedure for initial clustering.
1) procedure adjust clustering (grid list)
2) update the density of all grids in grid list;
3) for each grid g whose attribute
(dense/sparse/transitional) is changed since last call to adjust
clustering()
4) if (g is a sparse grid)
5) delete g from its cluster c, label g as NO CLASS;
6) if (c becomes unconnected) split c into two clusters;
7) else if (g is a dense grid)
8) among all neighboring grids of g, find out the grid h
whose cluster ch has the largest size;
9) if (h is a dense grid)
10) if (g is labeled as NO CLASS) label g as in ch;
11) else if (g is in cluster c and |c| > |ch|)
12) label all grids in ch as in c;
13) else if (g is in cluster c and |c| <= |ch|)
14) label all grids in c as in ch;
15) else if (h is a transitional grid)
16) if ((g is NO CLASS) and (h is an outside grid if g is
added to ch)) label g as in ch;
17) else if (g is in cluster c and |c| >=|ch|)
18) move h from cluster ch to c;
19) else if (g is a transitional grid)
20) among neighboring clusters of g, find the largest one c
satisfying that g is an outside grid if added to it;
21) label g as in c;
22) end for
23) end procedure

Figure 4: The procedure for dynamically adjusting
clusters.
The calculated values of z1, z2, z3 components will be
applied as an input for both the clustering algorithms to form
the clusters based on their risk level.
3) Patients Health status
Using clustering algorithm we form the clusters for
attributes stated above. And then for patients current input we
predict patients health status i.e. patient is fit or unfit.
IV. EXPERIMENTAL RESULT
The above described algorithms used for formation of
clusters on medical database. The data will be collected from
the Switzerland data set. The data set contains the 107
instances and the 14 attributes. The attributes are age, sex,
Blood Pressure, Cholesterol, Chest Pain and etc. The
performance of these algorithms will be computed by using
correctly predicted instance. [1]
Performance Accuracy= correctly predicted Instance/
Total Number of Instance
TABLE I. PERFORMANCE OF CLUSTERING ALGORITHM
Cluster
Category
Cluster
Algorithm
Measures


Clusters
Correctly
Classified
Instance
In
correctly
Classified
Instance
Prediction
Accuracy
Simple
K-means
89 18 83.18
D-stream 94 13 87.85
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
59 | P a g e
www.ijacsa.thesai.org
From above table we observed that the performance of
density based algorithm is better than simple K-means.
Accuracy of D-stream algorithm is more than K-means.
V. FUTURE SCOPE AND CONCLUSION
K-means is unable to handle arbitrary cluster formation
because prediction of the number of classes to be formed is not
fixed. The D-stream algorithm has superior quality and
efficiency, can find clusters of arbitrary shapes, and can
accurately recognize the evolving behaviors of real-time data
streams. Therefore, D-stream will perform better in biomedical
applications. This system can be further developed for real time
analysis of biomedical data to predict patients current health
status.
The proposed system can be used for monitoring elderly
people, Intensive Care Unit (ICU) Patient. Also the system
gives the health status of patient, it can be used be used by
clinicians to keep the records of patients.
The proposed system is adaptive since it can handles more
than one physiological signal. The proposed system uses
historical biomedical data which is very useful for prediction of
current health status of a patient by using clustering algorithms
like K-means, D-stream, etc. Prediction of health status is very
sensitive job, D-stream will perform better here, as it supports
arbitrary cluster formation which is not supported by K-means.
Also D-stream is particularly suitable for users with little
domain knowledge on the application data that means it wont
require the K-values. Hence D-stream is parameter free and
proves to give more accurate results than K-means when used
for cluster formation of historical biomedical data.
REFERENCES
[1] P.Santhi, V.Murali Bhaskaran Computer Science & Engineering
Department Paavai Engineering College, Performance of Clustering
Algorithms in Healthcare Database, International Journal for Advances
in Computer Science, Volume 2, Issue 1 March 2010
[2] Sellappan Pandian, Rafigh Awang,Heart Disease Prediction System
using Data Mining Techniques,IEEE Computer, Vol 7, PP.295-
304,August2008.
[3] Mahnoosh Kholghi, Mohammadreza Keyvanpour, An Analytical
Framework For Data Stream Mining Techniques Based On Challenges
And Requirements, Mahnoosh Kholghi et al. / International Journal of
Engineering Science and Technology (IJEST), Vol. 3 No. 3 Mar 2011
[4] Yixin Chen, Li Tu, Density-Based Clustering for Real-Time Stream
Datain Proceedings of the 13th ACM SIGKDD international
conference on Knowledge discovery and data mining, 2007
[5] Nuria Oliver , Fernando Flores-Mangas, HealthGear: A Real-time
Wearable System for Monitoring and Analyzing Physiological signals
in proceeding BSN06 Proceedings of the International Workshop on
Wearable and Implantable Body Sensor Networks, 2006
[6] Vikram Singh, Sapna Nagpal A Guided clustering Technique for
Knowledge Discovery A Case Study of Liver Disorder Dataset,
International Journal of Computing and Business Research, Vol.1, no.
1, Dec 2010
[7] RifatShahriyar, Md. Faizul Bari, GourabKundu, Sheikh IqbalAhamed
and Md. Mustofa Akbar 5,Intelligent Mobile Health Monitoring
System(IMHMS), International Journal of Control and Automation, vol
2,no.3, Sept 2009, pp 13-27.
[8] Daniele Apiletti, Elena Baralis, Member, IEEE, Giulia Bruno, and Tania
Cerquitelli, Real-Time Analysis of Physiological Data to Support
Medical Applications, IEEE Transactions On Information Technology
In Biomedicine, Vol. 13, No. 3, May 2009.
[9] The MIMIC database on PhysioBank (2007, Oct.) [Online]. Available:
http://www.physionet.org/physiobank/database/mimicdb
[10] Jiawei Han and Micheline Kamber, Data Mining: Concepts and
Techniques , Second Edition
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
60 | P a g e
www.ijacsa.thesai.org
Question Answering System for an Effective
Collaborative Learning

Prof. Kohei Arai
1

Information Science,
Saga University - Japan

Anik Nur Handayani
1,2

Electrical and Information Technology
State University of Malang - Indonesia


Abstract The increasing advances of Internet Technologies in
all application domains have changed life styles and interactions.
With the rapid development of E-Learning, collaborative
learning is an important for teaching, learning methods and
strategies. Interaction between the students also student with the
teacher is important for student to gain knowledge. Based on the
four basic teaching styles formal authority, demonstrator or
personal model, facilitator and delegator, today combined
between facilitator and delegator style is responsible for student
learning. It is student centered and the teacher as facilitates the
material and activities, but learning becomes part of valuable and
effective when they collaborate with each other, and as the
teacher who will delegates and facilitates the responsibility of
learning to the students. In this paper, we introduce an effective
question answering Q&A system for collaborative learning,
which can act not just like a virtual teacher, but also virtual
discussion for student. With the proposed system, brings a new
Q&A system, student can attach their question when they want
collaborate using collaborative learning capitalize on one
anothers resources and skills. Students can ask their questions to
the group when they want to collaborate with others, asking one
another for information, evaluating one anothers ideas, then
each of the answer will compare with encyclopedia data base. In
this research, the Q&A system for the Senior High School in
Indonesia, in this subject of Information Communication
Technology implemented. From the the 40 question and 120
answer, the result is 90,48% precision 50% recall.
Keywords component; E-Learning; Collaborative Learning;
Q&A; Knowledge Base.
I. INTRODUCTION
The concept of Collaborative Learning is two or more
people learn or attempt to learn something together than
independent. Different with individual learning, in
collaborative learning people can exploit and share their
resources and skills by asking for information, evaluating,
monitoring one anothers information and idea, etc [1].
Collaborative Learning is a model that knowledge can be
created by sharing experiences within a population where
members actively interact [2] [3]. In the Collaborative Robotic
Instruction (A Graph Teaching Experience Computers and
Education), the goal of collaborative learning is methodologies
and environments which learners or students where each
depends on and is responsible to each other [4]. Including both
directly with face-to-face conversations [5] or using computer
discussions (online forums, chat rooms, etc.) [6].
In [3] authors indicate that when they found some problem,
students learn better when they learn together more frequently
than working individually as members in a group. Indeed, the
effectiveness of collaborative learning on the internet has been
identified by various studies. Interaction among learners is
fostered as communication over the internet is unpretentious
and convenient when addressing to a single user or multiple
users. However, interaction students and a teacher address a
problem, the teacher cannot constantly online every time, and it
is not possible for the teacher to deal with lots of question from
students all the time and in a timely manner. In the
collaborative learning, students are encouraged to ask question
[7]. Therefore, there is a need to describe an automated Q and
A system to support learning efficiency of collaborative
learning.
In this paper, we proposed question answering system for
an effective collaborative learning. With the proposed system,
brings a new Q&A system, student can attach their question
when they want collaborative using collaborative learning
capitalize on one anothers resources and skills. Students can
ask their questions to the group when they want to collaborate
with others, asking one another for information, evaluating one
anothers ideas. The method also considers that students can
communication through Q&A interaction such a discussion
forum to support information sharing. The paper is organized
as follows. First is the motivation for the question and
answering system. Section 2 presents the related works on
collaborative learning. Section 3 explains about the question
answering mechanism proposed. In section 4, present the
architecture of system that we proposed an interactive effective
Q&A system for collaborative learning. The user interactions
between user and the system are explained. Finally, section 6,
are summary and conclusion of this paper.
II. RELATED WORK
Collaborative learning is one of the study groups. Some
studies show that students get the most current learning through
group rather than independently [8] [9]. Studies by the OTTER
Group [10] have shown that the ideal class is organized around
50/50 rule. At least 50% of the time students spend is spent
interacting with and learning about the other student in the
virtual classroom. The social aspect of the classroom is an
important factor. If social aspect missing, than student
dissatisfaction rises dramatically, as does the attrition rate. In
this learning mode, which is collaborative learning, students
who are interested in sharing their knowledge from a learning
group to communicate and discuss all kinds of questions,
asking one another for information, evaluating one anothers
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
61 | P a g e
www.ijacsa.thesai.org
idea for help and teach each other. Therefore, learning is both a
group activity and a social process and thus learning
performance is strongly affected peers [11].
In the development of networks, comprise all forms of
electronically supported learning and teaching that can
eliminate the obstacles of time and space. In the collaborative
learning, students can take part by computer at anyplace, at the
same time or different time (synchronous and asynchronous).
Researchers have used activity theory to analyze Computer
Supported Collaborative Work (CSCW) system [12].
Group communication relationship [13] refers to the intra-
group relationships determined by the interactions among
members. However, how to form a learning group after the
group is a problem in collaborative learning.
Several study about Q&A for collaborative learning had
been done. An application of Question Answering System for
Collaborative Learning has been designed. In this application
learners can attach their question to the group when they want
to collaborate with others, and the teacher providing answers to
them. In this case, the collaborative between the students and
the teacher to gain knowledge, and becomes question
answering system like a virtual teacher [7] [14]. In this paper,
we proposed question answering system for an effective
collaborative learning. The originality of the system is bringing
a new Q&A system, in which students can attach their
questions to the group when they want to gaining and sharing
knowledge with others, by collaborative learning capitalize on
one anothers resources and skills (asking and evaluating one
anothers idea), then each of the answer will compare with
encyclopedia data base.
III. PROPOSED E-LEARNING SYSTEM WITH KNOWLEDGE-
BASE SYSTEM
E-learning system, which is proposed here is based on
knowledge base system which allows acquires knowledge
about students and uses the acquired knowledge. Once students
learn with the proposed system through questions and answers,
then all students can use the previously acquired Q&A
information which relates to the current question. Figure 1
shows the complete overview proposed of Question Answering
Mechanism.
A. Proposed Question Answering System
An online Wikipedia-encyclopedia was chosen as the
corpus for the task. Wikipedia is a free, web-based,
collaborative, multilingual encyclopedia project supported by
the non-profit Wikimedia Foundation. Indonesian Wikipedia is
the edition of Wikipedia in the Indonesian language; there were
over 100.000 articles in the Indonesian Wikipedia project. In
the research proposed, we using Indonesian Q&A system
specifically in the Information Communication Technology
Subject, for the Senior High School.
An automated Q&A system in collaborative learning
operates proposed work based on the question answering
knowledge base. When a student needs some information, he or
she can ask a question through a designed interface. When a
new asked question enters the system, query is created. Then
other students will response the question with answering and
evaluating one anothers ideas by vote. This representation is
then compared with the representations of Wikipedia data base.
A similarity percentage is given between the student answer
and any existing Wikipedia data base. Based on the biggest
vote some time is not describing the best answer, after time for
answering and voting finish, similarity percentage will be show
for every answer. It aims to estimate whether which one of the
answered considered with the Wikipedia data base. Also,
student can access the topic question more clearly through the
link provided. Such as [7], if no any information from
Wikipedia comparing with neither student answer, nor the
student is satisfied with any the answers (no proper match from
the knowledge base), and then the teacher will answer or might
be sent to the student. After the teacher manually answer the
question, the new Q&A set is formed and entered into the
question answering knowledge base. When student meeting
some difficulties or having no difficulties, a student can see
what problems other students have encountered in learning
now and in the past and see the answers or solutions the teacher
offered by browsing the knowledge base. The whole process of
question answering in the effective of collaborative learning is
shown in figure 1.
Student
Information
Needs
Student Answer
Student Vote
Q/A Knowledge Base
Instructor
Manually Answer
Q/A Answer
Engine
Information
Requirde Knowledge
Accumulation
Required
Knowledge
Beyond Scope
Required Knowledge
Browse Q/A
Ask Question
Compare
with
Wikipedia
Database

Figure 1. Question Answering Mechanism
B. Architecture of Question and Answering System
The architecture of the question answering system is shown
in figure 2. There are eight main components in the system,
including the student agent Q, the student agent A, question
analysis and query generation, Q&A acquirer, question
answering knowledge base, similarity machine, Q&A browsing
component, answer generator.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
62 | P a g e
www.ijacsa.thesai.org
Student Agent Q Student Agent A Instructor Agent
Answer generator
Question Answering Knowledge Base
Question Analysis
Query Generation
Answer Generator
Q&A Browsing
Answer Question
Vote
Similarity Machine
Q&A Acquirer

Figure 2. Architecture of Question Answering System
The functions of the components are briefly described as
follows,
Student Agent Q
Student agent Q is to be the interface between student and
the system. Student can send his or her question and receive
answers (feedback from other student) from this interface.
Beside from student also provide feedback from the teacher for
manually answering and question if the system are not
satisfied.
Student Agent A
Student agent A is the component interface between student
and the system provides an answer and voting question. In this
component also provide similarity machine to match student
answer with Wikipedia database, to ensure the answer
considering biggest vote not necessary correct answer.
Question analysis, query generation
An analyzed question represents an asyntactic and semantic
analysis of a question. It serves as an interface between the
question analysis and query generation stages. A query is a
search engine query generated at the query generation stage
and executed at the search stage.
Q&A Acquirer
If there is an unanswered question, the questions will
temporarily stores at this component. When the teacher gets
online, and then will manually answer the question with an
interface (Instructor Agent). This component are very
important for a reason that students question might be not
answer by other student or students answer cannot compare
with Wikipedia database, cause the information not relevant. A
new Q&A will be formed and saved to the Q&A knowledge
base, after a question replied by the teacher.
Question Answering Knowledge Base
A knowledge base provides a means for collected,
organized, shared, searched and utilized information. Q&A
knowledge base is a knowledge base where questions
corresponding with answer. The knowledge will be
accumulated and rich, through the process of student question
answer and manually teacher answer.
Q&A Browsing Component
Beside asking a question, when student meeting some
difficulties or having no difficulties, a student can see what
problems other students have encountered in learning now and
in the past and see the answers or solutions the teacher offered
by browsing the knowledge base.
Answer Generator
Several interfaces have been designed, to let students and
teacher is able to interact with the system. The interfaces for
the student to ask question and get answers (Student Agent A)
are developed as shown in Figure 3. Under the interface in
figure 3, student can key in an Indonesian question and get an
answer immediately without waiting for the teacher to get
online. Then other student can answer the question and can
discuss whether which one of their answer as same as their
opinion. After that, Q&A system makes a point for each of
student answer that closely with Q&A system answer. If no
satisfactory answer is found, then it will present the teacher
demanding a manual answer whenever he or she is online.
Similarity Machine
This component uses Levenstein and Word similarity
method to calculate the similarity between students answer and
Wikipedia database. In the question analysis, extracting
question query is necessary. Specific interrogation all has
question query focus, usually question word (Qw) is the
question feature, such Qw as what, when, where, who, why and
how. In this system we using what, when, where, who type of
question. How and why, are notable mission because its
procedural answer. Using Levenstein for what question
(Definition expectations), to calculate percentage Levenstein
distance between two length string (students answer and the
knowledge base answer). And using Word similarity for when,
where, and who question. In the word similarity, the distance of
words is a real number in [0, ] a word and its own distance is
zero. The following equation is word similarity formula.
( )

( )
()
Where is an adjustable parameter.
C. User Interactions with Question and Answering System
Several interfaces have been designed, to let students and
teacher be able to interact with the system. The interfaces for
the student to ask questions and get answers (Student Agent A)
are developed as shown in figure 3. Under the interface in
figure 3, student can key in an Indonesian question and get
answer immediately without waiting for the teacher to get
online. Then other student can answer the question and can
discuss whether which one of their answer as same as their
opinion. After that, Q&A system makes point for each of
student answer that closely with Q&A system answer. If no
satisfactory answer is found, then student can press a button to
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
63 | P a g e
www.ijacsa.thesai.org
send the question to the Q&A acquirer, which will then present
the answered questions to the teacher demanding for a manual
answer whenever he is online.





(a)




(a)











(b)
Figure 3. Students interfaces question and get the answer
Beside asking a question, when student meeting some
difficulties or having no difficulties, through the interfaces in
figure 4, a student can see what problems other students have
encountered in learning now and in the past and see the
answers or solutions the teacher offered by browsing the
knowledge base.













(a)












(b)
Figure 4. Browsing activities
IV. EXPERIMENT
According to the method described above and the structure
of question and answering system, we build an experimental
system in Indonesian Q&A system specifically in the
Information Communication Technology Subject, for the
Senior High School. We choose 40 questions, and there are two
indexes in this experimental, precision and recall. The
calculation formula is as follow:

()

(3)
In the formula, is the number of right matching of
questions; b is the number of without matching of question; c is
the numbers of wrong matching questions. Through the
experiment, we can get the data of precision and recall, the
result is in Table 1.
TABLE 1. RESULT OF EXPERIMENT
Question 40
Answer 120
Right Matching 57
Without Matching 57
Wrong Matching 6
Precision 90,48 %
Recall 50 %
From the table, from the 40 question and 120 answer there
are 57 answer are without matching because where the machine
similarity get the keyword from the student question, there is
not information about keyword question in the Wikipedia data
based. In this case, the teacher will provide answers to students.
After a question is replied by the teacher, not only the student
will be notified, but also a new Q&A set will be formed and
saved to the question answering knowledge base. We have this
component for a reason that students questions might be
unanswered by the system due to no suitable or desirable
answers for them in the knowledge base.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
64 | P a g e
www.ijacsa.thesai.org
V. CONCLUSION
In this research, a system for online automatically
answering students questions in the collaborative learning
environment has been designed. The system operated upon the
question answering knowledge base. In the knowledge base,
pairs of question with its corresponding answer (Q&A sets)
were collected through the process of students asking questions
and other students will response the question with answering
and evaluating one anothers ideas by vote. This representation
is then compared with the representations of Wikipedia data
base. A similarity percentage is given between the student
answer and any existing Wikipedia data based. Based on the
biggest vote some time is not describe the best answer, after
time for answering and voting finish, similarity percentage will
show for every answer. It aims to the students who asked the
question and information for students who answer and vote can
estimate approximately correct answer. Also, students can
access the topic question more clearly through the link
provided. It was very important to have such a system in
collaborative learning environment. It benefited both the
teacher and the students; students can look for the answers to
their questions without the constraint of time and space.
REFERENCES
[1] Dillenbourg, P. (1999). Collaborative Learning: Cognitive and
Computational Approaches. Advances in Learning and Instruction
Series. New York, NY: Elsevier Science, Inc.
[2] Chiu, M. M. (2000). Group problem solving processes: Social
interactions and individual actions. Journal for the Theory of Social
Behavior, 30, 1, 27-50.600-631.
[3] Chiu, M. M. (2008).Flowing toward correct contributions during
groups' mathematics problem solving: A statistical discourse analysis.
Journal of the Learning Sciences, 17 (3), 415 463
[4] Mitnik, R., Recabarren, M., Nussbaum, M., & Soto, A. (2009).
Collaborative Robotic Instruction: A Graph Teaching Experience.
Computers & Education, 53(2), 330-342.
[5] Chiu, M. M. (2008). Effects of argumentation on group micro-creativity.
Contemporary Educational Psychology, 33, 383 402.
[6] Chen, G., & Chiu, M. M. (2008). Online discussion processes.
Computers and Education, 50, 678 692
[7] Wang, C.C., Hung J.C., Yang C.Y., Shih T.K. (2006). An Apllication of
Question Answering System for Collaborative Learning. IEEE
Conference on ICDCSW06
[8] Johnson, D.W. and Johnson R.T. (1999) Cooperation and competition:
Theory and research. Edina. MN: Interaction Book Company
[9] Slavin, R. (1996) Research on cooperative learning and achievement:
what we know, what we need to know. Contemporary Educational
Pshchology, 21, 1, pp. 43-69.
[10] Gilrory, K.(2001). Collaborative e-learning: the right approach
([Online].Available at : http://www.ottergroup.com/otter-with-
comments/right_approach.html).
[11] Lave, J., and Wenger, E. (1991) Situated Learning: Legitimate
Peripheral Participation. Cambridge University Press, Cambridge
[12] Kuutti, K. (1991) The concept of activity as a basic unit of analysis for
CSCW research. Proceedings of the Second European Conference on
Computer-Supported Co-operative Work: EC-CSCW91 (eds. L.J.
Bannon, M. Robinson &K. Schmidt) pp. 249-264, Kluwer, Dordrecht.
[13] Watzlawick, P. (1967) Pragmatics of Human Communications: A Study
of Interactional Patterns. Pathologies and Paradoxes. W.W. Norton, New
York.
[14] Bahreininejad A, Alinaghi Tanaz, 2011, A Multi Agent Question
Answering System for E-Learning nad Collaborative Learning,
nternational Journal of Distance Education Technologies, 9(2), 23-39,
April-June 2011.
[15] Xu Jinzhong, Jia Keliang, Fu Jibin. 2008. Research of Automatic
Answering System in Network Teaching. The 9th International
Conference for Young Computer Scientist.
AUTHORS PROFILE
Kohei Arai received BS, MS and PhD degrees in 1972,
1974 and 1982, respectively. He was with The Institute
for Industrial Science and Technology of the University
of Tokyo from April 1974 to December 1978 and also
was with National Space Development Agency of Japan
from January, 1979 to March, 1990. During from 1985 to
1987, he was with Canada Centre for Remote Sensing as
a Post Doctoral Fellow of National Science and
Engineering Research Council of Canada.
He moved to Saga University as a Professor in Department of Information
Science on April 1990. He was a councilor for the Aeronautics and Space
related to the Technology Committee of the Ministry of Science and
Technology during from 1998 to 2000. He was a councilor of Saga University
for 2002 and 2003. He also was an executive councilor for the Remote
Sensing Society of Japan for 2003 to 2005. He is an Adjunct Professor of
University of Arizona, USA since 1998. He also is Vice Chairman of the
Commission A of ICSU/COSPAR since 2008. He wrote 26 books and
published 227 journal papers.
Anik Nur Handayani received the B.E. degree in
electronics engineering from Brawijaya University, and
the M.S. degree in Electrical Engineering, from Institute
of Technology Sepuluh Nopember, Surabaya, Indonesia,
in 2004 and 2008, respectively. She is currently a PhD
Student at Information Science in Saga University, Japan.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
65 | P a g e
www.ijacsa.thesai.org
An Efficient Method For Multichannel Wireless Mesh
Networks With Pulse Coupled Neural Network

S.Sobana
Assistant professor
Department of ECE
PSNA College of Engg & Tech-Dindigul 624622
S.Krishna Prabha
Associate professor
Department of Mathematics
PSNA College of Engg & Tech-Dindigul 624622


AbstractMulti cast communication is a key technology for
wireless mesh networks. Multicast provides efficient data
distribution among a group of nodes, Generally sensor networks
and MANETs uses multicast algorithms which are designed to
be energy efficient and to achieve optimal route discovery among
mobile nodes whereas wireless mesh networks needs to maximize
throughput. Here we propose two multicast algorithms: The
Level Channel Assignment (LCA) algorithm and the Multi-
Channel Multicast (MCM) algorithm to improve the throughput
for multichannel sand multi interface mesh networks. The
algorithm builds efficient multicast trees by minimizing the
number of relay nodes and total hop count distance of the trees.
Shortest path computation is a classical combinatorial
optimization problem. Neural networks have been used for
processing path optimization problem. Pulse Coupled Neural
Networks (PCNNS) suffer from high computational cast for very
long paths we propose a new PCNN modal called dual source
PCNN (DSPCNN) which can improve the computational
efficiency two auto waves are produced by DSPCNN one comes
from source neuron and other from goal neuron when the auto
waves from these two sources meet the DSPCNN stops and then
the shortest path is found by backtracking the two auto waves.
Keywords-Wireless Mesh Networks; Multicast; Multichannel;
Multiinterface; Shortest path; DSPCNN; Auto wave; Search space.
I. INTRODUCTION
Unlike mobile adhoc networks or wireless sensor networks
route recovery are energy efficiency is not the major concern
for mesh network due to limited mobility and the rechargeable
characteristics of mesh nodes. Moreover supporting major
applications such as video on demand poses a significant
challenge for the limited bandwidth of WMNs it is necessary to
design an effective multicast algorithm for mesh networks. It
improves the system throughput by allowing simultaneous
close-by transmissions with multichannel and multi
interfaces. It assigns all the available channels to the interfaces
instead of just the non-overlapping channels.
We propose level channel assignment algorithm
multichannel multicast algorithm to improve throughput for
multichannel and multi interface mesh networks. Our design
builds a new multicast backbone - tree mesh which partitions
mesh network into different levels based on the Breadth First
Search (BFS), and then heuristically assigns channel to
different interfaces. The Pulse Coupled Neural Network is a
very active neural network .The PCNN is modified so that the
output pulses decay in times. These modified PCNN models
need fewer neurons than other approach. This paper proposes a
faster PCNN model, which can improve the computational
efficiency significantly.
II. LEVEL CHANNEL ASSIGNMENT ALGORITHM
The nodes obtain their level information. The BFS is used
to traverse the whole network. All the nodes are portioned into
different levels according to the hop count distances between
the source and the nodes.
If node a (in level i) and b (in level i+1) are within each
others communication range, then a is called the parent of
b, and b is called the child of a.
We build a multicast tree based on the node level
information. Initially, the source and all the receivers are
included in the tree. Then, for each multireceiver v, if one of its
parents is a tree node then connect it with that parent, and stop.
Otherwise randomly choose one of its parents, say fv, as relay
node on the tree, and connect v and fv. Afterwards, we try to
find out the relay node for fv recursively. The process repeats
until all the multireceivers are included in the multicast tree.
The tree nodes decide their channel assignment with the
level information.
The source node (level 0) only uses one interference,
which is assigned channel 0. This interference is
responsible for sending packets to the tree nodes in
level 1.
The internal tree node in level i (i1) uses two
interfaces: one is assigned channel i-1, which is used to
receive packets from the upper level; the other is
assigned channel 1, which is used to forward the
packets to the tree nodes at level i+1.
The leaf in the level I (i1) uses two interfaces: one
uses Channel i-1 to receive the packets from level i-
1,the other uses channel I to forward the packets to the
mesh clients within the communication range that
desire to receive the packets.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
66 | P a g e
www.ijacsa.thesai.org
Fig.1.An example for LCA and tree mesh,(a)Network topology ,(b) multicast tree, (c) channel assignment, and (d) tree mesh
For example in fig.1, the node s is the source and nodes f,
g, e are the multireceivers. In fig1.a {s, f, g, and e} are included
in the multicast tree. Since nodes of gs parents are tree nodes,
it randomly selects d as a parent node and connects node g with
d. Then choose ds parent b as a tree node and connect d with
b. Since bs parent s is a tree node connect b with s. Next, we
start from multireceiver e. Connect e with its parent node b and
stop because b is already connected with tree node s. Similarly
the third multireceiver f, connect f with c, c with a and then a
with s. Thus the tree construction is completed by connecting
all the receivers with the tree.
The LCA algorithm has two advantages: simple
implementation and throughput improvement. At the same time
the use of multiple channels reduces the close-by interference
and allows more simultaneous transmissions.
To improve the system throughput by the following ways,
first LCA cannot diminish the reference among the same level
s since it uses the same channel at the same level. Second,
when the number of available channels is more than that of the
levels, some channels will not be utilized, which is a waste of
channel diversity. Third, the channel assignment does not take
the overlap property of the two adjacent channels into account.
For all I, channel and channel i+1 are adjacent in frequency, so
they partially interfere with each other. Thus, the channel I for
level i sti ll has some interference effect with the channel i+1
for level i +1.
III. MULTICHANNEL MULTICAST
A. Algorithm
To improve the system throughput, the MCM algorithm is
proposed to minimize the number of relay nodes and the hop
count distances between the source and the destinations, and
further reduce the interference by exploiting all the partially
overlapping channels instead of just the orthogonal channels.
B. Construction of multicast protocol
When all the Nodes are multireceivers, the multicast
problem becomes the broadcast problem. We can say that the
broadcast is a special case of multicast. The broadcast structure
in the mesh network is built by the following steps.
After the BFS traversal, all the nodes are divided into
different levels. Delete the edges between any two nodes of the
same level, with which we get the elementary communication
structure tree mesh.Fig.1a and 1d given an example of the
original network topology and its corresponding tree mesh.
Identify the minimal number of relay nodes that form the
broadcast tree. Using more relay node means more
transmissions in the network. Because the number of available
channel is limited by current technical conditions, more
transmissions would result in more interference and result in
more bandwidth cost. Hence, minimizing the multicast tree size
helps to improve the throughput. The purpose of this step is to
identify the relay node for a node that has more than one parent
nodes so that the number of relay node is minimal.
C. Structure of Multicast protocol
In broadcast structure unnecessary branches are present if
the destinations do not involve all the nodes. Hence, we
propose to construct a slim structure y using the MCM Tree
Construction algorithm.
The goal of the algorithm is to discover the minimal
number of relay nodes needed to construct a multicast tree. The
search process starts from the bottom to the top. We use a
simple example to explain the process in a tree mesh in Fig.3a,
where nodes 6, 7, and 8 are the multireceivers. First select
node 4 at level 2 because it covers all the multireceivers at
level 3.Next select node 2 at level 1, which covers all the multi
receivers and the relay node at level 2.By doing these steps
finally we get the multicast tree in Fig.3b
D. Channel Assignment
Multi receivers can be connected with the gateway through
mini al hop count distance as discussed earlier. Now we
discuss about how to assign channels to the interference of tree
nodes, for that we propose two allocation algorithms:
ascending channel allocation and heuristic channel assignment.

a
a
b
d
e
g
h
i
j
g
f
c
s
s
b
d
h
i j
s
b
a

d
c
e
g
f
e
f
c
0
0
1
1
1
2
0
2
3 3
2
1
0
1
2
2
(a)
(b) (c) (d)

a
b
d e
g
h
i
c
s
f




j
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
67 | P a g e
www.ijacsa.thesai.org
a) Ascending channel Allocation
The interference that a node uses to receive packets from its
relay node at the upper layer, called as Receiver Interference
(RI), is disjoint from the interference the node uses to forward
packets to children, called Send-Interference (SI).To guarantee
that the relay node can communicate with its children, each
nodes RI is associated with the SI of its relay node, i.e., they
should be assigned the same channel.
The algorithm is explained as follows: From the top to
bottom in the tree, the channels are assigned to the interfaces in
the ascending order until the maximum channel number is
reached, then start from the channel 0 again. Although simple,
this approach avoids the situation that the same channel is
assigned to two nearby links that interfere with each other. In
Fig .4, the numbers of orthogonal channels are three, the
number above the node represents the channel number used for
its RI, and the number below the node represents the channel
number for its SI.
b) Heuristic Channel Assignment
We noticed that the interference range decreases with the
increase of the channel separation for two wireless links which
have short physical distance. To minimize the sum of the
interference area of all the transmissions this algorithm is
proposed.
We use IR (uv) to indicate the interference range of sender
u of one link with respect to sender v of another link According
to our consideration all the have the same transmission range
R, IR (u v) =R* [iu-iv].iu and iv are the channels of u and v
for their SIs, and t is the interference factor. When allocating
a channel for relay node u, the channel assignment should take
a channel that minimize the sum of the square of the IRs
between u and us neighboring relay nodes, i.e., minimize
,where N(u) is the set of neighboring relay nodes
of u. This is because the bigger the interference area means
bigger chance two transmissions may interfere. The
interference area is approximated as a circle whose area is
determined by IR2 (uv).Since
,
The heuristic Channel assignment is used to minimize



Figure 2. Relay node search example

Fig.4. Ascending channel allocation example
IV. DSPCNN MODEL
A. Preliminaries
The input to the preprocessing stage is an undirected graph
G = (V, E) with n vertices and m edges, and non-negative
lengths l(e) for each edge e. Another two inputs are a source
node s and a goal node g. The goal of this algorithm is to find a
shortest path from s to g. Let dist(s, g) denote the shortest path
length from s to g with respect to l. i.e. dist(s, g) =dist (g, s).
a
b c d
1 2 3
4 5
6 7
b
c
d d
4 5 6
7
6 7
b
1
2
3 4 5 6 7
a
c d
(a)
(b)
(c)
(d)
0 2
1
4
1
0
1
0
2
2
1
1
0
0
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
68 | P a g e
www.ijacsa.thesai.org
B. Model Of DSPCNN
To compute the point-point shortest path more efficiently a
Dual Source Pulse Coupled Neural Network (DSPCNN) model
is proposed. This model can produce two Auto waves from two
different firing sources. At t=t0, the source neuron and goal
neuron fire and emit pulses simultaneously. Then, the two Auto
waves propagate in parallel by their neighboring neurons at
next instant till they meet together. In order to differentiate two
auto waves, the auto wave propagating from source neuron is
denoted as Ps, and the auto wave propagating from goal is
denoted as Pg. If neuron fires on the simulating of Ps auto
wave it outputs Ps pulses. If a neuron fires on the simulating of
Pg auto wave, it output Pg pulses. If a neuron fires on the
simulating of both Ps and Pg pulses, it indicates that the two
auto waves meet and the model should stop.
C. Shortest Path Computation Using DSPCNN
To compute the shortest path for networks, first they are all
transformed to a graph for further processing. The next step is
to map the graph into DSPCNN model. Each node in the graph
corresponds to a DSPCNN neuron, and each edge associated
with a link between neurons. The cost of an edge can be
viewed as an external input for the two neurons connected by
the edge.
During time t=0, Source neuron and Goal neuron fire
simultaneously. Then the auto waves Ps and Pg from the firing
sources propagate to their neighbors. A variant meeting is used
to determine whether the two auto waves meet together, and
the meeting neutron is denoted by Nm. If Nj fires on the
simulation of Ni, we call Ni is the precursor of neuron Nj.
V. CONCLUSION
In our paper we investigate the multicast algorithm wireless
networks. In order to achieve efficient multicast in WMNs, two
multicast algorithms are proposed by using multichannel and
multi interfaces.
These algorithms are focused on increasing the throughput
and decreasing delay. With neural networks the proposed
DSPCNN is used to achieve higher efficiency and involve
lower search space, which can save the run time significantly.
ACKNOWLEDGMENTS
We would like to thank the anonymous reviewers for their
helpful comments. We would like to thank our mother for her
kind support. We would like to thank our institution and
management for their cooperation.
REFERENCES
[1] J.So and N.Vaidya, 2004.Multichannel Mac for Ad Hoc Networks:
Handling Multi-channel Hidden Terminals Using a single transceiver.
[2] K.Ramachandran, E.M.Belding, K.Almeroth,and M.Buddhiko,
interference- Aware channel Assignment in multiradio wireless
messnetworksProc.IEEE INFOCOM,2006.
[3] http://www.seattlewireless,net,2009.
[4] A.Mishra,V.Shrivatsava,and S.Banarjee,Partialy Overlapped channels
Not Considred Harmful,
[5] E.W.Dijkstra .A note on two prolems in connection with
aphs.Numeische Mathematick.1959,(1):269-271
[6] R.CM.Folyd.Algorthms 97;shortest path Communications of the
ACM.1962,6(5):345
[7] H.Qu.Z.Yi.A new algorithm for finding the shortest paths using PCNN
s.Chaos, solutions and fractals.2007,4(33):1220-1229
[8] T.H.Cormen,C.E.Leiserson,R.L.Rivest Introduction to Algorithms.The
MIT press,2001.
[9] B.Yu,L.Zhang.pulse coupled Neural Networks for contour and motion
matchings.IEEE Transanctions on Newral networks. 2004, 5(15):1186-
1201
[10] Guokai Zeng, Bo Wang, Yong Ding, Li Xiao, Matt Mutka, Multicast
Algorithms for Multi-Channel Wireless Mesh Networks
AUTHORS PROFILE
S.SOBANA.-received the B.E (with distinction) in Electronics
andCommunication Engineering and M.E (with distinction) in Applied
Electronics from RV.S College of Engineering and Technology, Anna
University Chennai, Tamil Nadu, India .in 2005 and 2007 respectively.
Currently working as an assistant professor in Electronics and Communication
Engineering Department, P.S.N.A College of Engineering and Technology,
Tamil Nadu, India. Her research mainly focuses on ad hoc networks,
congestion control, power management techniques in wireless networks.
S.Krishna Prabha-receiver her B.Sc Mathematics degree in 2000 from
G.T.N Arts College, MK University, Tamil Nadu, India. MSc and MPhil
degree in Mathematics from M.K. University, Tamil Nadu, India in 2002 and
2004 respectively. Currently she is pursuing the M.E degree from System
Engineering and Operation Research department, Anna University, Trichy,
Tamil Nadu, India. Currently working as an associate professor in
Department of Mathematics, P.S.N.A College of Engineering and
Technology, Tamil Nadu, India. Her research interests include graph theory,
operation research, boundary value problems, ad hoc networks.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
69 | P a g e
www.ijacsa.thesai.org
A Congestion Avoidance Approach in Jumbo Frame-
enabled IP Network
Aos Anas Mulahuwaish, Kamalrulnizam Abu Bakar, Kayhan Zrar Ghafoor
Faculty of Computer Science and Information System
Universiti Teknologi Malaysia
Johor, Malaysia

Abstract Jumbo frame is as an approach that allows for higher
utilization of larger packet sizes on a domain-wise basis,
decreasing the number of packets processed by core routers while
not having any adverse impact on the link utilization of fairness.
The major problem faced by jumbo frame networks is packet
loss during queue congestion inside routers is as the RED
mechanism that is recommended to combine with jumbo frame
treats jumbo frame encapsulation as one packet by drop the
whole jumbo frame with packets encapsulate during the
congestion time. RED dropping the whole jumbo frame
encapsulation randomly from head, middle and tail inside queue
of router during periods of router congestion, leading to affect
the scalability and performance of the networks by decreasing
throughputs and increasing queue delay. This work proposes the
use of two AQM techniques with jumbo frame, modified Random
Early Detection MRED and developed drop Front technique
DDF, which are used with the jumbo frame network to reduce
packet drop and increase throughput by decreasing overhead in
the network. For the purpose of evaluation, network simulator
NS-2.28 was set up together with jumbo frame and AQM
scenarios. Moreover, for justification objectives, the proposed
algorithm and technique for AQM with jumbo frame were
compared against the existing AQM algorithm and techniques
that are found in the literature using metrics such as packet drop
and throughput.
Keywords- Jumbo Frame; Queue Congestion; AQM; RED.
I. INTRODUCTION
Computer networks have experienced rapid growth over the
years, from transferring simple email messages to now being a
full media resource where full length movies are commonly
transmitted. Many users have begun to use the internet for
many things; as a result, the connections of internet have
started to become strained where before the common solution
of the internet service provider (ISP) was capable of providing
sufficient bandwidth to users in the network. However, recent
research has found that the users access speed has increased
and thus affects the efficiency of the network. Therefore new
techniques need to improve the efficiency of the network
traffic. Many techniques from multicasting to packet caching
have been used to improve the efficiency of the network, but
with limited success as these techniques suffer from one or
more drawbacks including global network support, application
support, asymmetric and computation overhead. The current
assumption with networking research is also that it affects an
individual network flows quality of service (QoS) including
the packet loss, end to end delay and jitter; however these
researches presented techniques that investigate the possibility
of trading a minimal amount of an individual flows QoS
typical delay so as to obtain better overall network performance
[1].
One of the issues facing networks is the number of packets
required to be processed per second, whereby the gigabit link
core router may have to route anywhere from 90,000 to
2,000,000 packets per second. As line speed increases to
greater rates, so does the number of packets that need to be
processed; one way to reduce the load on the router is to
increase the maximum transmission unit (MTU) of the
network. Unfortunately while the MTU of Ethernet is 1500
bytes, up to 50% of the packets transferred across the network
are 64 bytes or less.
Jumbo frame is a technique that aims to reduce the number
of packets processed by the core routers, by reducing the
number of packets. This is accomplished by transmitting many
packets in the domain into a single large jumbo frame for
transmission across the core network. In ordering the aggregate
packets together in a jumbo frame, incoming packets are
queued briefly by egress point. Once the jumbo frame reaches
the egress of the domain, the original packets are rebuilt and
transmitted on toward their final destination.
II. RELATED WORK
A jumbo frame has a common size of 9000 bytes, which is
exactly six times the size of a standard Ethernet frame [5]. A 9k
byte jumbo frame would be 9014-9022 bytes together with the
Ethernet headers. This makes it large enough to encapsulate a
standard network file system (NFS) data block of 8192 bytes,
yet not large enough to exceed the 12,000 byte limit of
Ethernet's error checking in cyclic redundancy check algorithm
(CRC) [5].Undoubtedly, smaller frames usually mean more
CPU interruptions and more processing overhead for a given
data transfer size [9]. When a sender sends a data, every data
unit plus its headers have to be processed and read by the
network components between the sender and the receiver. The
receiver then reads the frame and TCP/IP headers before
processing the data. This whole process, plus that of adding the
header to frames and packets from the sender to the receiver
consumes CPU cycles and bandwidth [13]. For these reasons,
increasing the frame size by sending data in jumbo frames
means fewer frames are sent across the network when
considering the fact of high processing cost of network packets
[3]. These generate improvements in CPU utilization and
bandwidth by allowing the system to concentrate on the data in
the frames, instead of the frames around the data. The
justification behind increasing the frame size is clear; larger
frames reduce the number of packets to be processed per
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
70 | P a g e
www.ijacsa.thesai.org
second. A single 9k Jumbo Frame replaces six 1.5k standard
frames, producing a net reduction of five frames as only one
TCP/IP header and Ethernet header is required instead of six,
resulting in 290 (5*(40+18)) fewer bytes transmitted over the
network [14].
In terms of improving bandwidth, it takes over 80,000
standard Ethernet frames per second to fill a gigabit Ethernet
pipe, which in turn consumes a lot of CPU cycles and
overhead. By sending the same data with 9,000 bytes jumbo
frames, only 14,000 frames need to be generated and the
reduction in header bytes frees up 4 Mbps of bandwidth. The
resources used by the server to handle network traffic are
proportional to the number of frames it receives. Therefore,
using fewer large frames dramatically improves server and
application performance, compared to a larger number of
smaller frames [14]. Jumbo frame improves core router
scalability, by encapsulating packets with the same next
autonomous systems (AS) and egress point into larger packets
for transmission across the domain. Critically, the design of
jumbo frame functions on a domain-wise scale, instead of end-
to-end; the external entities (other domains and end hosts) are
unaware that any conversion took place. The overall jumbo
frame shown in Figure 1:
Figure 1. Jumbo Frame Structure
The description of the structure as shown in Figure 1 is that
when packets arrive at an ingress node to the domain, the
ingress node and the packets are sorted into queues based upon
their egress point of the network in their path that is obtained
from the border gateway protocol (BGP) routing table [12].
A jumbo frame Encapsulation Timer (JFET) is started for the
queue. Packets that are being sent through the same egress
point are combined into the same jumbo frame, subject to
MTU; once the JFET for the queue has expired, the Jumbo
Frame is released towards the next AS. The jumbo frame is
routed through the core of the network, with the routing
provided by using the standard routing mechanism of the
network. The jumbo frame arrives at the egress node and the
original packets are separated out, after which the original
packets are forwarded onto their final destinations. There are
two main benefits of using jumbo frame [1]. The first benefit is
that jumbo frame lowers the number of packets that the core
routers are responsible for processing, thus allowing better
scaling for the network as line speeds increase. The second
beneficial aspect of jumbo frame is that data is more efficiently
transferred by reducing the number of physical layer headers
used (due to a lower number of packets).
A. Fast Packet Encapsulation
The jumbo frame is structured to allow for efficient
encapsulation, inspection, and de-capsulation [5]; packet
overhead is minimal and is offset by the reduction in physical
layer headers. The structure of the jumbo frame is shown in
Figure 2 containing the following fields:












Figure 2. Jumbo Frame Structure
The destination address of the jumbo frame is the same as
the first packet stored in the jumbo frame. For a multiprotocol
label switching (MPLS) network, the destination address is the
MPLS address of the first packet stored in the group. This
ensures proper routing for all packets as all encapsulated
packets contained in the jumbo frame would arrive at the next
correct AS in their path. The design of the jumbo frame allows
the original packets to be de-capsulated with minimal effort
while also keeping the overhead of the jumbo frame to a
minimum. As shown, the overhead of the jumbo frame is
6 + 4N bytes. However, the overhead is offset by the reduction
in physical layer headers. The net cost (or benefit) of jumbo
frame can be stated as:
Equation 1: Cost = HIP + HJG + (N _ 4) (Hp _ (N 1)) (1)
The cost of jumbo frame in the above equation comes from
the size of the IP header (HIP), the jumbo frame header (HJG),
and the number of encapsulated packets (N). The reduction in
bandwidth comes from the reduction in physical headers 5
(Hp). For example, if the network is an Ethernet network and
two packets were encapsulated into a jumbo frame, then HIP
= 20, HJG = 6, N = 2, Hp = 38, and the total cost would be 4
bytes. In other words, 4 bytes of bandwidth would be
conserved.
B. Egress Shaping
When the jumbo frame reaches its destination, the packets
need to be de-capsulated and released to the next node on their
path to the destination. If all the packets are released as soon as
they are removed from the jumbo frame, this can lead to
dropped packets at the client due to the receive buffer
overflowing [2]. Hence packets are shaped at the egress
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
71 | P a g e
www.ijacsa.thesai.org
according to the differences in their arrival time (pQTime). In
other words, if two packets arrive at the ingress node 4 ms
apart, they are released from the egress node 4 ms apart.
C. Active Queue Management (AQM) with Jumbo Frame
The structure of the jumbo frame allows active queue
management AQM scheme techniques and methods are an
important type of technology with aims to improve the
utilization of the network [4] and [8]. While jumbo frame can
be combined with AQM techniques and methods, this allows
the combination between the jumbo frame and AQM
techniques and methods to solve many problems in jumbo
frame networks, and also enhance the efficiency and scalability
of jumbo frame network by decreasing the packet loss and end
to end delay, reducing the overhead and increasing throughput
for jumbo frame networks to perform optimally, RED is one of
the AQM methods that work with jumbo frame [4] and [8], for
preventing the gateway router from becoming full and ensuring
that jumbo frame can transmit to the destinations.
In [6] and [11], two different methods that RED queues can
use to determine the queue utilization are presented. The first is
through the number of packets in the queue and the second is to
determine queue utilization by number of bytes in the queue.
RED detect the congestion in jumbo frame networks, and
decrease the congestion of overflow by randomly drop whole
jumbo frame, RED treat jumbo frame as a one big packet, so
when the drop occur RED will used the same drop operation
with standard packet.
D. Random Early Detection (RED) and Drop from Front
Random early detection (RED) Algorithm was first
proposed by [6]. This discipline maintains a moving average of
the queue length to manage congestion. If this moving average
of the queue length lies between a minimum threshold value
and a maximum threshold value, then the packet is either
marked or dropped with a probability. If the moving average of
the queue length is greater than or equal to the maximum
threshold then the packet is dropped. Even though it tries to
avoid global synchronization and has the ability to
accommodate transient bursts, in order to be efficient RED
must have sufficient buffer spaces and must be correctly
parameterized. In contrast, RED algorithm uses packet loss and
link utilization to manage congestion. RED gateways can be
useful in gateways with a range of packet-scheduling and
packet-dropping algorithms. For example, RED congestion
control mechanisms could be implemented in gateways with
drop preference, where packets are marked as either essential
or optional, and optional packets are dropped first when the
queue exceeds a certain size. Similarly, for the example of a
gateway with separate queues for real time and non-real time
traffic, RED congestion control mechanisms could be applied
to the queue for one of these traffic classes.
The RED congestion control mechanisms monitor the
average queue size for each output queue, and by using
randomization chooses connections to notify of that congestion.
Transient congestion is accommodated by a temporary increase
in the queue. Longer-lived congestion is reflected by an
increase in the computed average queue size, and results in
randomized feedback to some of the connections to decrease
their windows. The probability that a connection is notified of
congestion is proportional to that connections share of the
throughput through the gateway. In addition, gateways
detecting congestion before the queue overflows are not limited
to packet drops as the method for notifying connections of
congestion. RED gateways can mark a packet by dropping it at
the gateway or by setting a bit in the packet header, depending
on the transport protocol. When the average queue size exceeds
a maximum threshold, the RED gateway marks every packet
that arrives at the gateway. If RED gateways mark packets by
dropping them, rather than by setting a bit in the packet header,
then the RED gateway controls the average queue size even in
the absence of a cooperating transport protocol when the
average queue size exceeds the maximum threshold. One
advantage of a gateway congestion control mechanism is that it
works with current transport protocols and does not require that
all gateways in the internet use the same gateway congestion
control mechanism; instead it could be deployed gradually in
the current Internet. RED gateways are a simple mechanism for
congestion avoidance that could be implemented gradually in
current TCP/IP networks with no changes to transport
protocols.
Drop from front technique drops the head of the queue if
the incoming packet sees the queue as full. With the drop from
front policy that governs when a packet arrives to a full buffer,
the arriving packet is allowed in, with space being created by
discarding the packet at the front of the buffer. This shows that
for networks using TCP, the Internet transport protocol, a drop
from front policy results in better performance than is the case
under tail dropping and its variations [10]. Drop from Front a
partial solution to the problem of throughput collapse in
networks where TCP represents a sizeable part of the load.
Drop from Front can be used in conjunction with other
strategies such as Partial Packet Discard. In [10], showed that
moving to a drop from front strategy considerably improves
performance and allows use of smaller buffers than is possible
with tail drop. Drop from Front is also applicable to both the
switch and routers. During congestion episodes when buffers
are full, Drop from Front causes the destination to see missing
packets in its received stream approximately one buffer drain
time earlier than would be the case under tail drop. The sources
correspondingly receive earlier duplicate acknowledgements,
causing earlier reduction in window sizes.
However, drop from front has the advantage that the switch
and router does not need to maintain a table of drop
probabilities and does not have to know the traffic type being
carried. This is because drop from front also reduces latencies
for successfully transmitted packets and hence is a sensible
policy to use for delaying sensitive non-feedback controlled
traffic as well. This reduction in latency has been described by
[15], who considered a drop from front scheme for a very
different problem where none of the sources were feedback
controlled. They found that drop from Front resulted in shorter
average delay in the buffer for eventually transmitted packets
and recommended its use for time-constrained traffic.
III. METHODOLOGY AND RESULTS
Modified random early detection (MRED): a RED queue is
an important technique that aims to improve the utilisation of
the network and remove the synchronisation that tends to occur
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
72 | P a g e
www.ijacsa.thesai.org
with TCP flows when the network becomes congested. There
are two different methods that RED queues can use to
determine the queue utilisation. The first method is to simply
use the number of packets, while the second is to use the
number of bytes consumed in order to determine queue
utilisation. The second method has more overhead, however, it
allows for smaller packets to be favoured over larger packets.
This effectively gives priority (less chance to be dropped) to
smaller packets (eg.TCP acknowledgments). In jumbo frame
networks if RED is not modified in any way, jumbo frame will
be treated the same as any other packet. This behaviour is not
advantageous as a jumbo frame has the same percent chance to
be dropped as does any other packet. However, any time a
jumbo frame is dropped, all encapsulated packets are lost.
Because multiple packets are lost, this can result in poor TCP
performance, as packets from the flow can be dropped, thus
resulting in a greater than desired reduction of traffic.
MRED will start to calculate the new average queue size
and the time for the new flows that coming to the queue and
MRED will do compression between the number of jumbo
frames and the capacity of the queue and check if the capacity
of queue are enough to receive new flows or not. If the queue
has enough space for all flows then MRED will allow all
jumbo frames to queue up for forwarding out to the different
destinations. However if the capacity is not enough, there is a
congestion over flow problem that will happen in this queue,
all that will be calculated based on the MRED detection
mechanism. In this case MRED will do the calculation for each
jumbo frame for probabilities drop. From here MRED will
check the header of jumbo frame and will exactly check and
compare two of fields inside the header. It will check the
average length of each jumbo frame to calculate out the
percentage of jumbo frame packets to distinguish that jumbo
frame is not like any normal packet (this is because the average
length size is high), MRED will also check the number of
packets which encapsulated within the field header to verify
there are encapsulation packets inside. Here MRED will only
work with the average length and the number of packets that
encapsulated within jumbo frame and will not work with the
header of capsule frame.
After that, MRED will register out all of the information
from the header for each jumbo frame encapsulation; then
based on the percentage of packets that have been
encapsulated, MRED will mark jumbo frame for drop sub
packets inside, the percentage of packets that will drop are
different from jumbo frame to another that are in the same
flows. This calculation is based on the percentage of upper and
lower bounds for each jumbo frame with the packets
encapsulate, in which this calculation based on specific
mathematical formulas. MRED will compare the percentage of
packets inside each of jumbo frame with average queue size for
the queue of router. Then MRED will decide the percentage of
packets dropped from each of jumbo frame, to make the
average queue size stable between them and during the
congestion overflow time, and to reduce losing the whole
encapsulation of jumbo frame but for only subs of packets. The
marking operation of MRED for jumbo frame and the packets
inside are related with time that sets for each jumbo frame,
after that the probability marking drop will be set.
In this work MRED are combined with DDF, so based on
this mechanism it will only mark the jumbo frame at the head
of queue and the packets at the head of those jumbo frames.
MRED will distribute the drop marking operation with the
different jumbo frames to reduce the congestion and to let some
of the packets inside those encapsulations left within without
dropping it whole. This mechanism will reduce losing the
whole packets inside each jumbo frame at the same flow;
Figure 3 shows the diagram of MRED operation structure.

Figure 3. MRED operation structure
Developed drop front (DDF): development drop front
mechanism is combined with modified RED for the steps of the
packets drop in jumbo frame networks. After MRED has
marked the jumbo frame that needs to be drop inside it, by
calculating the upper and lower bounds for the encapsulations
based on the percentage of jumbo frame encapsulations. When
the MRED marked jumbo frames for dropping process, only
the sub packets inside the jumbo frame will be dropped; the
marked sub packets inside jumbo frame encapsulation will be
done in the head of encapsulation frame, based on the
mechanism of DDF which combines with MRED, so there are
no random packets dropped inside jumbo frame. DDF will wait
until the processing time finishes for the MRED with all the
flow packets, then the time for DDF operation will start; DDF
checks how many encapsulated for jumbo frame that marked
by MRED, based on the percentage for each jumbo frame
inside the queue. After checking the numbers of marked jumbo
frames, DDF calculates the sub packets that are marked for
drop by MRED inside each marked encapsulation.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
73 | P a g e
www.ijacsa.thesai.org
DDF will set new time differently with time that was set
before by MRED for each encapsulate frame that has been
marked by MRED to do drops operation. Each marked jumbo
frame has its own time drop packets. This time is set based on
how many packets that are marked to drop each sub packets
time that have been marked to drop for this operation. There is
a delay time for packet drop from one packet to another and
this time will be calculated and set for the total drop operation
time for the whole jumbo frame and each jumbo frame have
different time with other. The drop operation starts with the
first jumbo frame in the head of queue that was marked for
drop operation. Inside this marked encapsulation sub packets
drop operation will start with the first packet in the jumbo
frame encapsulation that has been marked to drop. The DDF
operation will start to drop packets by packets inside each
jumbo frame encapsulation, and the packets drop will set in
sequence number of router queue for each jumbo frames. Then
it will send notification to the source for retransmitting the loss
packets. In this operation, not all the encapsulation of jumbo
frame is lost and the drop operation for the sub packets did not
happen randomly but only from the front of jumbo frame,
Figure 4 shows the DDF operation.

Figure 4. DDF operation
DDF allows the possibility of dropping partial packets
without significant overhead. Firstly DDF looks at the number
of packets stored in the jumbo frame encapsulations. Once the
number of packets to be dropped is decided, the packets will
removed from the head of the jumbo frame. The length of
jumbo frame is shortened by the lengths of the packets that are
to be removed and their lengths in the jumbo frame header are
set to zero. The number of packets field for each jumbo frame
got marked to drop sub packets will not be modified, and also
the average length field in header will not be modified. This is
due to the need for correct parsing at the egress router and the
need for simplicity in modifying the packets in flight.
Removing the lengths that are zeroed out is not a desirable
option because multiple memory copies would have occur
before the packet could be forwarded. So here jumbo frame
will forward out without restructuring the sequencing of
packets that were encapsulated, only the number of packets and
average length fields in the jumbo frame header are not
modified, DDF will set zero at jumbo frame header instead
each packet has been dropped directly and one by one based on
the time has been set for each jumbo frame marked and for
each packets inside need to be dropped to remove the
restructure operation, Figure 5 shows the average length of
packets after drop operation inside jumbo frame header.. After
that jumbo frame will de-encapsulate the rest packets to the
destination address by the egress operation. DDF eliminates the
random marked jumbo frame and dropped the packets inside
encapsulation. However, if the packets are able to be removed
randomly by MRED in jumbo frame, the complexity of the
partial drop would substantially increase. The increase in
complexity is from performing an MRED calculation on each
encapsulated and from memory move operations needed to
close the gaps in the jumbo frame after drop sub packets in
different places in the encapsulation. DDF eliminates the
restructure operation for each jumbo frame; all that will
decrease the overhead in jumbo frame networks.

Figure 5. The average length of packets inside jumbo frame header after
packets drop
A. Simulation Setup
The simulations presented here illustrate MRED with DDF
well-understood dynamic of the average queue size varying
with the congestion level, resulting from MRED with DDF and
normal RED with tail drop fixed mapping from the average
queue size to the packet dropping probability and the
percentage of throughput. These simulations focus on the
transition period from one level of congestion to another.
These simulations used a simple dumbbell topology with 6
nodes, the congested link of 1.5Mbps. The buffer
accommodates 20 packets, which, for 3000 byte packet size
and MTU 3000 byte, corresponds to a queuing delay of 0.28
seconds. In all of the simulations, weight of queue is set as a
default in NS-2 to 0.0027, the choice of Wq determines the
queue weight of the averaging for the average queue size, if
Wq is too law, then the estimated average queue size is
probably responding too slowly to transient congestion, if
it too high, then the estimated average queue size is too
closely tracks the instantaneous queue size, MINth is set to 5
packets, the setting for MINth depends on exactly what the
desired tradeoffs is at that router between low average delay
and high link utilization. In the NS-2 MINth is set to a default
of 5 packets because if MINth is set as small as one or two
packets would only denied burstiness in the arrival process, and
MAXth is set to 15 packets; there times more than MINth.
Maximum value for the current marking of packet probability
MAXp is constrained to remain within the range [0.01, 0.5] (or
equivalently, [1%, 50%]), and the percentage of Jumbo Frame
packets is 0.025, the average size of encapsulated packet is read
from the Jumbo Frame header, not calculated at the router.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
74 | P a g e
www.ijacsa.thesai.org
B. Simulation Scenario
The first scenario is for the increased average queue size in
congestion, which used for testing the proposed MRED with
DDF and also for test normal RED with tail drop in jumbo
frame networks, this scenario is focus for the increase the
average queue size in router queue during the congestion over
flows at the transition period. The new flows are more than the
buffer size capacity, the over flows burst in the specific
simulation time, the average queue size has been increased
because this over flows and been near or over the MAXth, so
the congestion and packet drop happened, with decrease in
throughput. This simulation is test the efficiency and scalability
for the proposed MRED with DDF algorithm and compare the
results with normal RED with tail drop results, for reduce the
packet loss in and increase throughput with jumbo frames.
C. Results for MRED with DDF an Increased in Congestion
Scenario
For this simulation scenario, the forward traffic consists of
two long-lived TCP flows, and the reverse traffic consists of
one long-lived TCP flow. At 25 seconds time, there are 20 new
flows started, one every 0.1 seconds, each with a maximum
window of 20 jumbo frames. This is not intended to model a
realistic load, but simply to illustrate the effect of a sharp with
the average queue size changing as a function of the packet
drop rate. However after roughly 10 seconds, and because of
the new 20 flows of jumbo frames the congestion happened,
the algorithm of MRED detected the congestion and started to
calculate the average queue size in the overflow time, MRED
marked packets inside jumbo frames by put the drop
probability first, and then mark sub packets inside jumbo
frames at the head of queue and at the head of jumbo frames
to decrease the congestion and then the drop will be done at the
head of those jumbo frames by DDF without changing the
length of information inside the header for each jumbo frames
marked for drop. Here MRED with DDF has brought the
average queue size back down to the range, between (6 7
packets). That means the proposed algorithm makes the
average queue size away from the MAXth by making the
probability of the packet drop less (MINth avg < MAXth).
The simulations with MRED with DDF have a higher
throughput with smaller packet loss (drop), at the first half part
of simulation, the throughput percentage is 42.45% and the
packet drop is 0.69%. In the end of simulation scenario, the
throughput becomes 91.7% and packet drop 8.24%. Figure 6
shows the MRED with DDF an increase in average queue size
in congestion, the green trend represents the instantaneous
change of queue length and the red trend shows the average
queue size.
D. Result for Normal RED with Tail Drop an Increased in
Congestion Scenrio
For this scenario simulation, it used the same simulation
with MRED and DDF but with normal RED and tail drop
instead. There are also at 25 seconds time where there are 20
new flows start, one for every 0.1 seconds, and each with a
maximum window of 20 jumbo frames. In Figure 7 the graph
illustrates normal RED with tail drop, with the average queue
size changing as a function of the packet drop rate. With 20
new jumbo frames flow, congestion happened and packet
dropped, because RED detected congestion and the RED
algorithm dropped marked jumbo frames totally by tail drop at
the tail of queue only. The packet drop rate changes from
0.90% with throughput 41.06% over the first half of the
simulation, to 8.50% with the throughput 90.20% over the
second half of simulation That means the average queue size
here is become near to MAXth because of the algorithm for
normal RED with drop tail did not reduce the number of
packets that dropped during the congestion time. Due for that
reason, the average queue size has been increased and the
throughput has been decreased. Figure 7 shows the normal
RED with tail drop with an increase in congestion, here can be
noticed that at 25 second during the congestion the trend of
average queue size increases and almost near with MAXth
which means more packet drop happened.

Figure 6. MRED with DDF with an increase avg in congestion

Figure 7. Normal RED with tail drop with an increase avg in congestion
E. Results Comparision
Four scenarios were compared in this study, starting from
results for MRED with DDF with an increase of average queue
size in congestion compared with results for normal RED with
tail drop with an increase in congestion too; results for MRED
with DDF with a decrease of average queue size in congestion
compared with results for normal RED with tail drop with a
decrease in congestion also. All those comparisons are based
on the simulation metrics packet drop and throughput.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
75 | P a g e
www.ijacsa.thesai.org
It can be observe in Figure 8 and 9 the comparison between
the results for MRED with DDF and normal RED with tail
drop in the same scenario with an increased of average queue
size during the congestion. It has shown at the end of
simulation lower percentage packet drop decrement 26% when
used RED with DDF than in normal RED with tail drop, and
throughput increment 1.56% when used with MRED with DDF
than in normal RED with tail drop; it can be observed that
when there are over flow in the queue the MRED with DDF
makes the average queue size lower than MAXth by decreasing
drop of jumbo frame encapsulation and just drop packets inside
jumbo frame encapsulation during over flow in queue and
increases the throughput. This means the proposed MRED with
DDF technique achieved the objectives for decreasing the
packet drop and increases the throughput with jumbo frame,
which will be led to enhance the scalability and efficiency of
jumbo frame networks.

Figure 8. Packet drop rate between MRED with DDF and RED with
tail drop in increase of congestion

Figure 9. Throughput rate between MRED with DDF and RED wih
tail drop in increase of congestion
IV. CONCLUSION
This work has been proposed new scheme in AQM with
jumbo frame networks, by combined modified random early
detection MRED with developed drop front DDF. The
proposed algorithm help to reduce the packet loss in jumbo
frame networks, and increase the throughput, by reduce the
overhead and enhance the scalability and efficiency for jumbo
frame networks. The proposed algorithm has been
implemented by NS2 simulator, it have achieved the best
results for reducing the packet loss at queue and increase
throughput in jumbo frame environments when it compared
with a result for applying the normal RED combined with drop
tail technique in jumbo frame environments with the same
metrics.
REFERENCES
[1] Alteon (1999) Extended Frame Sizes for Next Generatinon Ethernets,
White paper, Lightwave technology journal, pages 66 -73.
[2] Balakrishnan, H. and Padmanabhan, V. N.,Seshan, S.,Stemm, S. and
Katz, R. H. (1998) TCP behavior of a busy interent server: Analysis
and improvements, INFOCOM'98. Seventeenth annual joint conference
of the computer and communication societies.
[3] Chelsio communication white paper (2007), Ethernet Jumbo Frames.
The Good, the Bad and the Ugly ITG fachbericht photonische netze
journal.
[4] Chung, J. and Claypool, M. (2003), Analysis Active Queue
Management, IEEE international symposium on network computing
and applications conference.
[5] Dykstra, P. (1999), Gigabit Ethernet Jumbo Frames, And why you
should care, White paper, WareOnEarth Communications and
Available at: http://sd.wareonearth.com/phil/jumbo.html,1999.
[6] Floyd, S. and Jacobson, V. (1993), Random Early Detection Gateways
for Congestion Avoidance, IEEE/ACM Transactions on Networking,
pages 397-413.
[7] Floyd, S., Gummadi, R. and Shenker, S. (2001) Adaptive RED: An
Algorithm for Increasing the Robustness of REDs Active Queue
Management, Preprint journal.
[8] Gass, R. (2004), Packet size distribution ACM SIGMETRICS
performance evaluation review journal, pages 373.
[9] Genkov, D. and Llarionov, R. (2006), Avoiding IP Fragmentation at the
Transport Layer of the OSI Reference Model, Proceedings of the
international conference on computer systems and technologies
CompSysTech, University of Veliko Tarnovo, Bulgaria.
[10] Lakshmant, T. V., Neidhardt, A., Teunis, J. (1996), The Drop from
Front Strategy in TCP and in TCP over ATM, Proceedings of the
fifteenth annual joint conference of the IEEE computer and
communications societies conference, pages 1242 - 1250.
[11] Ramakrishnan, K. and Floyd, S. (1999), A proposal to add Explicit
Congestion Notification (ECN) to IP, IETF RFC 2481.
[12] Rekhter, Y. and Li, T. (1995) A Border Gateway Protocol, IETF RFC
1771.
[13] Sauver, J. S. (2003), Practical Issues Associated with 9K
MTUs, University of Oregon computing center journal.
[14] Sathaye, S. (2009), Jumbo Framescommons design.
[15] Y n, N. and Hluchyj, M. G. (1990), Implication of Dropping Packets
from the Front of a Queue, IEEE transactions on communications.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
76 | P a g e
www.ijacsa.thesai.org
Cross Layer QoS Support Architecture with
Integrated CAC and Scheduling Algorithms for
WiMAX BWA Networks

Prasun Chowdhury, Iti Saha Misra, Salil K Sanyal
Department of Electronics and Telecommunication Engineering
Jadavpur University, Kolkata-700032, India


Abstract In this paper, a new technique for cross layer design,
based on present Eb/N0 (bit energy per noise density) ratio of the
connections and target values of the Quality of Service (QoS)
information parameters from MAC layer, is proposed to
dynamically select the Modulation and Coding Scheme (MCS) at
the PHY layer for WiMAX Broadband Wireless Access (BWA)
networks. The QoS information parameter includes New
Connection Blocking Probability (NCBP), Hand off Connection
Dropping Probability (HCDP) and Connection Outage
Probability (COP). In addition, a Signal to Interference plus
Noise Ratio (SINR) based Call Admission Control (CAC)
algorithm and Queue based Scheduling algorithm are integrated
for the cross layer design. An analytical model using the
Continuous Time Markov Chain (CTMC) is developed for
performance evaluation of the algorithms under various MCS.
The effect of Eb/No is observed for QoS information parameters
in order to determine its optimum range. Simulation results show
that the integrated CAC and packet Scheduling model maximizes
the bandwidth utilization and fair allocation of the system
resources for all types of MCS and guarantees the QoS to the
connections.
Keywords- Cross layer QoS support architecture; SINR based CAC
algorithm; Queue based Scheduling algorithm; Adaptive
Modulation and Coding; WiMAX BWA networks.
I. INTRODUCTION
Successful delivery of real-time multimedia traffic over the
IP based networks, like Internet is a challenging task because of
the strict requirement of Quality-of-Service (QoS). Traditional
IP based best effort service will be unable to fully meet all the
stringent requirements. The time-varying nature of channels
along with the resource constrained devices associated with
wireless networks make the problem even more complicated.
In wireless networks in contrast to wired networks, the channel
impairments vary rapidly due to the fluctuation of channel
conditions even when the transmitter and receiver are
stationary. It is, therefore, desirable to adjust the transmitted
power, type of modulation and data rate to match the channel
condition dynamically in the best possible manner in order to
support the highest possible channel capacity even under the
worst case condition of the channel. To make it feasible, the
cross layer design for wireless multimedia communication
networks has gained significant popularity in the research
community [1]. In cross layer design, the challenges from the
physical wireless medium and the associated QoS demands for
the relevant applications are taken into consideration so that the
rate, power, and coding at the physical layer can adapt
dynamically to meet the stringent requirements of the
applications under the current channel and network conditions.
The ever-increasing wireless traffic is a conglomeration of
various real-time traffic such as voice, multimedia
teleconferencing, games and data traffic such as WWW
browsing, messaging and file transfers etc. All these
applications require widely varying and diverse QoS
guarantees for different types of traffic. Next generation
wireless networks such as WiMAX (Worldwide
Interoperability for Microwave Access) or IEEE 802.16 [2]
have different types of services with varied QoS requirements,
but the IEEE 802.16 has not yet defined the standards for QoS
guarantees. Wireless channels suffer from bandwidth
limitation, fluctuations of the available bandwidth, packet loss,
delay and jitter. Real-time media such as video and audio is
highly delay sensitive but Non-real time media such as web
data is comparatively less delay sensitive but both types require
reliable data transmission. Some advanced techniques such as
Orthogonal Frequency Division Multiple Access (OFDMA)
[3], Adaptive Modulation and Coding (AMC) [4-6] and wide
variety of protocols and standards [7] are used to combat the
challenges for stringent QoS requirement in wireless networks.
Also, the mobile devices are power constrained. Maintaining
good channel quality in one side and minimizing average
power consumption for processing and communication on the
other are two very conflicting requirements. Receivers in
multimedia delivery systems are quite different in terms of
latency requirements, visual quality requirements, processing
capabilities, power limitations, and bandwidth constraints. In
view of the above constraints, a strict modularity and protocol
layer independence of the traditional TCP/IP or OSI stack will
lead to a sub-optimal performance of applications over IP
based wireless networks [8]. For optimization, we require
suitable protocol architectures that would modify the reference-
layered stack by allowing direct communication between the
protocols at non-adjacent layers or sharing of the state variables
across different layers to achieve better performance. The
objective of a cross layer design is to actively exploit this
possible dependence between the protocol layers to achieve
performance gains. Basically, a cross layer design involves
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
77 | P a g e
www.ijacsa.thesai.org
feedback received from other layers to optimize the relevant
parameters in the current layer to achieve the optimum
performance. Although the cross layer design is an evolving
area of research, considerable amount of work has already been
done in this area [1].
In this paper, information from the MAC layer is used to
optimize the parameter at PHY layer. Among several
alternatives, the IEEE 802.16 PHY layer standard realizes the
usage of Orthogonal Frequency Division Multiplexing
(OFDM) in order to mitigate the adverse effects of frequency
selective multi-path fading and to efficiently contrast Inter-
Symbol and Inter Carrier Interferences (ISI and ICI) [9]. In
addition, PHY layer also supports variable channel bandwidth,
different FFT sizes, multiple cyclic prefix times and different
modulation schemes such as QPSK, 16QAM and 64QAM with
convolutional encoding at various coding rates [10]. On the
other hand, Call Admission Control (CAC) mechanism [11-19]
and Scheduling mechanism [17-23] are the important wings of
WiMAX QoS framework at MAC layer. A CAC algorithm at
base station (BS) will admit a subscriber station (SS) into the
network if the BS ensures the minimum QoS requirement of
the SS without degrading the existing QoS of other SSs in the
network.
Providing CAC in IEEE 802.16 BWA networks guarantees
the necessary QoS of different types of connections and at the
same time decreases the Blocking Probability (BP) [13],
Dropping Probability (DP) [14] and Outage Probability (OP)
[24] for all types of connections. These QoS parameters can be
used as feedback information to the PHY layer in order to
select an appropriate Modulation and Coding Scheme (MCS)
for better network performance. Again, scheduling algorithm at
SS will distribute the bandwidth of the selected MCS among
the real-time and the non-real-time traffic, based on the type of
users in the network and their QoS requirements. Scheduling in
the uplink direction is tedious because of required bandwidth
request, bandwidth grant and transmission of data of each flow
which affect the real-time application due to large Round Trip
Delay (RTD) [25, 26].
The scheduling algorithm involves in fair allocation of the
bandwidth among the users, determining their transmission
order and enhancement of bandwidth utilization. Fairness refers
to the equal allocation of network resources among the various
users operating in both good and bad channel states. In this
paper, fairness has been quantified using Jains Fairness Index
[27].
A. Background literature
A significant number of proposals on cross layer designs
have been found in recent literature. Many of them refer to
feedback mechanisms between the PHY and MAC layers. In
[28] authors have introduced a cross layer approach for
calculating the QoS indicators (blocking rates, download time
and bit error rates) taking into account the PHY layer
conditions (modulation and coding, propagation and MIMO),
the MAC layer radio resource management algorithms and the
higher layer traffic characteristics. Authors have also
considered different admission control schemes and studied the
impact of adaptive modulation and coding on the performance
of elastic traffic. But no packet scheduling algorithms for fair
allocation of bandwidth among the traffic sources have been
considered. Also optimum range of Signal to Noise Ratio
(SNR) for adaptive MCS in the networks has not been
determined.
A framework of memory less scheduling policies based on
channel quality states, buffer occupancy states and
retransmission number states has been provided in [29]. The
scheduling policies combine AMC (Adaptive Modulation and
Coding) and ARQ (Automatic Repeat Request) in a cross layer
fashion that produces a good throughput performance with
reduced average delay. However, the authors have not
considered any CAC algorithm for connection management in
the networks.
Similar to [29], [30] proposes a scheduling algorithm only
at the MAC layer for multiple connections with diverse QoS
requirements, where each connection employs AMC scheme at
the PHY layer over wireless fading channels. A priority
function is defined for each connection admitted in the system,
which is updated dynamically depending on the wireless
channel quality, QoS satisfaction, and services across all layers.
The connection with highest priority is scheduled at each time.
At the MAC layer, each connection belongs to a single service
class and is associated with a set of QoS parameters that
quantify its characteristics.
A cross layer optimization mechanism for multimedia
traffic over IEEE 802.16 BWA networks has been proposed in
[31, 32].The main functionality of the cross layer optimization
mechanism resides at the BS part. The authors have used a
decision algorithm at the BS part that relies on the values of
two major QoS parameters, i.e., the packet loss rate and the
mean delay. These QoS parameters activate proper adjustment
of the modulation and/or the media coding rate, aiming at
improving QoS and system throughput. However, this paper
has not considered proper CAC and scheduling mechanisms for
the delay sensitive real-time traffic flow at MAC layer. The
authors have also not evaluated the optimum range of SNR in
order to uniformly control modulation and data encoding rates.
A cross layer scheduler that employs AMC scheme at the
PHY layer, according to the SNR on wireless fading channels
has been described in [33]. The authors have defined cost
function for each kind of multimedia connections based on its
service status, throughput or deadline in MAC layer to achieve
an optimum tradeoff between the throughput and fairness.
Though the authors have included a suitable scheduling
algorithm at the MAC layer but the authors have not considered
any CAC algorithm.
[34] proposes a cross layer optimization architecture for
WiMAX system. It consists of a Cross layer Optimizer (CLO),
which acts as an interface between MAC and PHY layers. The
CLO gathers and optimizes the parameters from both layers to
achieve optimum performance gain. The CLO gets the channel
condition information from the PHY layer and bandwidth
requests and queue length information from the MAC layer so
that it can switch to different burst profile for different
modulation and coding schemes. In this case also, the authors
have not specified proper CAC and scheduling mechanism at
the MAC layer. Moreover, no optimum range of SNR has been
taken into account in CLO while switching to different MCS.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
78 | P a g e
www.ijacsa.thesai.org
Another cross layer approach as provided in [3], analyses
several QoS parameters that include the bit rate and the bit
error rate (BER) in the PHY layer, and packet average
throughput/delay and packet maximum delay in the link layer.
Authors show that dynamic OFDMA has a stronger potential to
support multimedia transmission than dynamic OFDM-TDMA.
Authors have taken care of proper scheduling, throughput and
delay guarantee of traffic flows ignoring the CAC mechanism
at MAC layer.
[4] describes a polling based uplink scheduling schemes for
TCP based applications in a multipoint to-point fixed
broadband IEEE 802.16 BWA network. This scheme adapts
the transmission rates between the SSs and the BS dynamically
using adaptive modulation technique. With adaptive
modulation, the uplink scheduling algorithm helps to achieve
higher data transmission rate, fairness in slot assignment and in
the amount of data transmission.
B. Contribution
The novelty of our work focuses on the dynamic selection
of MCS based on Eb/N0 ratio of the concerned connections and
relevant feedback information obtained from the MAC layer in
terms of New Connection Blocking Probability (NCBP), Hand
off Connection Dropping Probability (HCDP) and Connection
Outage Probability (COP) to employ the most appropriate
MCS in the network. The target values of the feedback
information changes the selection of modulation and coding
rate dynamically based on optimum range of Eb/N0 ratio of the
connections to enable a flexible use of the network resources.
The effect of Eb/N0 ratio has been observed on NCBP, HCDP
and COP for all the relevant connection types under various
MCS by means of exhaustive simulation. The simulation
results help to determine the optimum range of Eb/N0 ratio for
all the connections with appropriate MCS. It has been observed
that the integrated CAC and Scheduling algorithm maximizes
the utilization and fair allocation of the system resources for all
types of MCS and provide better QoS support with reduced
NCBP, HCDP and COP in the IEEE 802.16 BWA networks.
Since there is also a possibility that the wireless channel
deteriorates or becomes unavailable at certain instant of time,
the CAC and scheduling algorithms at the MAC layer, which
has been left undefined in the standard, can get proper
knowledge of the channel condition, in terms of SINR of the
connections. Thereby, only the connection with a good channel
condition is admitted and scheduled for transmission to achieve
more gain in bandwidth utilization and fairness index.
An analytical model based on Continuous Time Markov
Chain (CTMC) [35] is developed for the more realistic analysis
of QoS parameters for the performance evaluation of IEEE
802.16 Broadband Wireless Access (BWA) Networks. As the
Markov Chain basically involves state transition analysis, the
probabilities for getting admission without degradation of QoS
for both real as well as non-real time services is accurately
analysed at the state level.
The remainder of the paper is organized as follows: Section
II discusses System models. In addition to it, cross layer
architecture and a joint algorithm for SINR based CAC and
Queue based scheduling are proposed. Detailed description of
the analytical model of proposed SINR based CAC algorithm is
given in Section III. Section IV sets the parameter for
performance evaluation. Section V shows QoS performance
results numerically for different MCS. Finally, Section VI
concludes the paper.
II. SYSTEM MODEL AND CROSS LAYER ARCHITECTURE
To accommodate a wide variety of applications, WiMAX
defines five scheduling services that should be supported by the
BS MAC scheduler for data transport. Four service types are
defined in IEEE 802.16d-2004 (Fixed) standard [12], which
includes UGS (Unsolicited Grant Service), rtPS (Real-time
Polling Service), nrtPS (Non Real-time Polling Service), and
BE (Best Effort). In addition, one more service i.e. ertPS
(Extended Real-time Polling Service) is defined in IEEE
802.16e-2005 (Mobile) standard [14]. The UGS is designed to
support real-time service flow that generates fixed-size data
periodically, such as T1/E1 and VoIP without silence
suppression. On the other hand, the rtPS supports the same with
variable data size, such as video streaming services. Similarly,
the nrtPS deals with FTP. The BE and ertPS perform tasks
related to e-mail and VoIP respectively. The guaranteed delay
aspect is taken utmost care in video streaming and VoIP. In the
mobile WiMAX environment, the handover procedure begins
as soon as the mobile SS moves into the service range of
another BS. The relevant QoS parameters specified in the
standard are Maximum Sustained Traffic Rate (MSTR),
Minimum Reserved Traffic Rate (MRTR), Maximum Latency
(ML), Tolerated Jitter (TJ) and Request/Transmission Policy.
The Service Flow Identifier (SFID), Connection Identifier
(CID) and traffic priority are mandatory for all QoS classes
[10].
A. PHY Layer Model
IEEE 802.16 specifies multiple PHY specifications
including Single Carrier (SC), SCa, OFDM and OFDMA.
Among the several alternatives, we consider only OFDMA in
the PHY802.16 layer for the cross layer design because of its
limited interference in the network [9]. OFDMA is similar to
OFDM using multiple sub-carriers to transmit data. However,
while OFDM uses all available sub-carriers in each
transmission, different sub-carriers could be arranged to
different subscribers in downlink and each transmission could
use the available sub-carriers in uplink in OFDMA.
We consider 802.16 PHY layer to be supported with
variable channel bandwidth, different FFT sizes, multiple
cyclic prefix time and different modulation schemes such as
QPSK, 16QAM and 64QAM with convolutional encoding at
various coding rates. The raw data rates of the OFDMA are
functions of several parameters such as channel bandwidths,
FFT size, sampling factor, cyclic prefix time, modulation
scheme, encoding scheme and coding rate. It can be up to 70
Mbps by using high-grade modulation scheme with other
suitable parameters [10, 36]. For proper design of PHY layer
with MCS, we first observe the effect of all the said parameters
in terms of raw data rate and spectrum efficiency.
In [36], a method of calculating raw data rate and OFDM
symbol duration for different MCS is given. In this paper, we
also derive a relevant term, spectrum efficiency (the alternative
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
79 | P a g e
www.ijacsa.thesai.org
term bandwidth efficiency is also frequently used), being
defined [37] as the transmitted bit per second per Hertz
(b/s/Hz). This normalized quantity is a valuable system
parameter. For instance, if data are transmitted at a rate of 1
Mbps in a 0.6 MHz wide baseband system, the spectral
efficiency is 1 Mb/s/0.6 MHz, or 1.67 b/s/Hz. In this paper,
spectral efficiency is evaluated as the ratio of raw data rate to
the channel bandwidth as determined by the particular MCS in
IEEE 802.16 BWA network. The calculation for the above
parameters is given below. Let, R be the raw data rate;

c/T * b * N R =
(1)
where:
N = Number of used sub-carriers
b = Number of bits per modulation symbol
c = Coding rate
T = OFDM symbol duration
The value of N is a function of the FFT size. Table I lists
the values of N for different FFT sizes.
TABLE I. NUMBER OF USED SUB-CARRIERS AS A FUNCTION OF FFT SIZE
[36]






The OFDM symbol duration is obtained by using the
following formulae [36]:
Bandwidth Channel * factor Sampling = Fs
(2)

Fs / FFT_size = Tb
(3)

Tb * G = Tg
(4)

Tg + Tb = T
(5)

where:
Fs = Sampling frequency
Tb = Useful symbol time
Tg = Cyclic prefix time
G = Cyclic prefix
T = OFDM symbol duration
Sampling-Factor is set to 8/7, G is varied as 1/32, 1/16,
1/8 and 1/4 and coding rates used are 1/2, 2/3, and 3/4. These
are the standard values specified in WiMAX [36].
Based on the given values of Sampling-Factor, Channel
bandwidth and G, the OFDM symbol duration T is
computed by using equations (2) to (5). Raw data rates can be
computed using equation (1). We compute raw data rate and
spectrum efficiency for different Cyclic Prefix, modulation
schemes and coding rates, using FFT_size = 2048 and channel
bandwidth= 20 MHz and the computed results are shown in
Table II. From Table II, it is observed that raw data rate and
spectrum efficiency are maximum when cyclic prefix G =1/32
(i.e. Minimum value of G) which is obvious because this value
of G causes lower time spread of the signal which results in the
larger data rate and hence provides maximum spectrum
efficiency.
B. Medium Access Control (MAC) Layer Model
We consider SINR based CAC algorithm for BS and Queue
based Scheduling algorithm for SS in the WiMAX MAC layer
for the design of cross layer model. The terminologies used in
this section are given in Table III. We make the following
assumptions in our proposed algorithm.
1) ertPS connection requests are considered to be same
as rtPS connections, because both connections have
the same QoS parameters and differ only by the way
of Request/Transmission policy.
2) BE connections are not considered in our CAC
scheme, because they are designed to support best
effort flows which do not need any QoS guarantees.
3) The connections of the similar service types have
same QoS parameter values.
TABLE II. LIST OF DATA RATES AND SPECTRUM EFFICIENCY FOR DIFFERENT MCS AND CYCLIC PREFIX
Modulation
Type
Coding
Rate
Cyclic Prefix=1/32 Cyclic Prefix=1/16 Cyclic Prefix=1/8 Cyclic Prefix=1/4
Raw data
rate
(Mbps)
Spectrum
Efficiency
Raw data
rate
(Mbps)
Spectrum
Efficiency
Raw data
rate
(Mbps)
Spectrum
Efficiency
Raw data
rate
(Mbps)
Spectrum
Efficiency
QPSK 1/2 15.5844 0.7792 15.1261 0.7563 14.2857 0.7143 12.8571 0.6429
QPSK 3/4 23.3766 1.1688 22.6891 1.1345 21.4286 1.0714 19.2857 0.9643
16QAM 1/2 31.1688 1.5584 30.2521 1.5126 28.5714 1.4286 25.7143 1.2857
16QAM 3/4 46.7532 2.3377 45.3782 2.2689 42.8571 2.1429 38.5714 1.9286
64QAM 1/2 46.7532 2.3377 45.3782 2.2689 42.8571 2.1429 38.5714 1.9286
64QAM 2/3 62.3377 3.1169 60.5942 3.0252 57.1429 2.8571 51.4286 2.5714
64QAM 3/4 70.1299 3.5065 68.0672 3.4034 64.2857 3.2143 57.8571 2.8929

FFT Size Number of Used Sub-carriers (N)
2048 1440
1024 720
512 360
128 72
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
80 | P a g e
www.ijacsa.thesai.org
GLOSSARY

C. Description of the SINR based CAC algorithm
In the proposed CAC mechanism, the QoS requirements
under consideration are indicated by SINR value, transmission
delay and bandwidth availability of the connections. The
proposed CAC algorithm uses the Signal to Interference plus
Noise Ratio (SINR) to check for the availability of enough
OFDMA sub-carriers for the new connection request. When a
new connection request arrives, the BS calculates the SINR
threshold (SINRth) for the requesting connection. The BS then
updates the measured SINRi for each connection type, with the
assumption that requesting connection of service type i is in
the system by occupying an OFDMA sub-carrier. If the
measured SINRi is greater than or equal to the corresponding
SINRth, and the transmission delay requirement is guaranteed
for every real time connection in the system and also the
required bandwidth is available for the requesting connection,
the connection is admitted into the system. However, if the
former two conditions are satisfied but bandwidth availability
is not met, the requesting connection is admitted in the system
through adaptive bandwidth degradation. Otherwise, the
connection request is rejected. The detail of CAC mechanism is
shown in the Figure 1.
D. Calculation of SINRth and SINRi
The SINRth for the requesting connection should be
determined in such a way so that the wireless transmission
error remains below the acceptable threshold for the connection
type i. To find SINRth, the effective bit rate of the requesting
connection type is first determined. The effective bit rate is
defined as the minimum bandwidth required by traffic source
to meet the transmission error threshold. In this paper, we have
considered that the effective bit rate of a connection type i, is
its MRTR, as defined in the WiMAX standard. Hence, SINR
threshold of a connection type i can be calculated as given in
equation (6).

) ( * ) ( = SINR
W
MRTR
i N
E
i th,
i
0
b
(6)
Let the requesting connection be the Cpth connection and
the requesting connections traffic type be i and the
connections which are already admitted in the particular
WiMAX cell is denoted by Cj, where 1 j p-1 and j= p. To
find the effects of the requesting connection on all other
admitted connections, the BS measures the total power
received from all existing connections Cj in the system. The
received power from all existing connection acts as interference
to the requesting connection. Let Sj be the received power
from the existing connection which acts as interference to the
OFDMA sub-carrier of the requesting connection and Si be
the signal power of the ith type of the Cpth requesting
connection in an OFDMA sub-carrier as shown in figure 2.
Hence,

=
=
1 p
1 j
j
S
(7)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
81 | P a g e
www.ijacsa.thesai.org

Figure 1. Flow chart of proposed SINR based CAC algorithm for WiMAX BWA Networks


Figure 2. OFDMA sub-carriers in time and frequency domain
Let SINRi be the new measured SINR for a connection
when a new connection type i wants to get admitted in the
network and be the thermal noise in the network. SINRi is
updated by the BS using equation (10) as derived below [10].

q +
i
i
S
= SINR

1
S
= SINR
i
i
+
q

1
S
S
= SINR
1 p
1 j
j
i
i
+
q
q

=
(8)
Again

S
i
q
for a connection can be calculated as given in
[37],

W
r
* )
N
E
(
S
i
i
0
b i
=
q
(9)


, get we ) 8 ( in
W
r
* )
N
E
(
S
Replacing
i
i
0
b i
=
q



1
S
)
W
r
( * )
N
E
(
= SINR
1 p
1 j
j
i
i
0
b
i
+
q

=
(10)

Where r
i
is the transmission data rate of i
th
connection type.
Again

S
1 p
1 j
j
q

=
can be calculated using equation (9) for
all connection type,
(11) )
W
r
* )
N
E
( * (n + )
W
r
* )
N
E
( * (n + )
W
r
* )
N
E
( * (n =
S
n
n
0
b
n
r
r
0
b
r
U
U
0
b
u
1 p
1 j
j
q

=

Where r
u
, r
r

and r
n

are the bit rates and n
u
, n
r
, n
n
are
the number of admitted connection of UGS, rtPS and nrtPS
services respectively.

E. Description of delay guarantee condition
A condition has been provided in [18] to satisfy the Delay
Guarantee required by the rtPS connection type i. Since token
bucket mechanism is used to schedule the packets, maximum
number of packets (in terms of arriving bits) that arrive in a
time frame of duration f is bi+ri*f. These arriving bits must be
scheduled in next (mi-1) time frame (see Table III) to avoid
delay violation. The required condition is given below, (for
proof of the condition kindly see [18])

f * r * 1] ) C C (1 * 1) [(m b
i rtPS,i NRT,i i i
+ s

(12)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
82 | P a g e
www.ijacsa.thesai.org
Where
) B * (n ) B * (n B C
r r U U i NRT,
=

and
r r i rtPS,
B * n C =
.

Here
i , NRT
C and
i , rtPS
C are calculated taking the new
connection type i to be admitted. Here, i refers to the rtPS
connection type.
F. Adaptive bandwidth degradation of rtPS and nrtPS
connections
In this scheme, bandwidths of lower priority connections
are degraded as per minimum bandwidth requirement of the
newly admitted or handoff connections. To admit more new
UGS connections, the bandwidth is degraded only from nrtPS
connections. In addition, degradation is performed on both rtPS
and nrtPS connections to allow more UGS, rtPS and nrtPS
handoff calls. The advantages of adaptive bandwidth
degradation [38] over fixed bandwidth degradation [12- 16]
are:
No fixed step size degradation.
No need to assign initial arbitrary step size.
Instead, the minimum required bandwidth is
calculated and then degraded.
The degradation is adaptive to the required
bandwidth.
Bandwidth utilization of the system is greatly
improved.
G. Need for rescheduling of bandwidth at SS
According to IEEE 802.16 standard, BS is responsible for
allocating the uplink bandwidth based on the request from SSs.
As the SS may have multiple connections, the bandwidth
request message should report the bandwidth requirement of
each connection to BS. All packets from application layer in
the SS are classified by the connection classifier based on CID
and are forwarded to the appropriate queue. Scheduler at SS
will retrieve the packets from the queues and transmit them to
the network in the appropriate time slots, as defined by the
uplink map massage (UL-MAP) [18] sent by the BS. The UL-
MAP is determined by the uplink packet scheduling module
based on the BW-request messages that report queue size of
each connection in SS. But the scheduler inside the BS may
have only limited or even outdated information about the
current state of each uplink connection due to large Round Trip
Delay (RTD) [25] as shown in Figure 3.
So a scheduling algorithm is needed in each SS to reassign
the received transmission bandwidth among different
connections. Since the uplink traffic is generated at SS, the SS
scheduler is able to arrange the transmission based on the up-
to-date information and provide tight QoS guarantee for its
connections.
Hereunder, we integrate the distributed scheduling
algorithm [26], with our proposed SINR based CAC scheme.
The scheduling algorithm is based on current queue size and
queue delay at SS.

Figure 3. RTD for Uplink transmission in IEEE 802.16
H. Scheduling algorithm at SS
Since UGS service has a critical delay and delay jitter
requirement and its transmission cannot be deferred or
interrupted by other flows, SS scheduler firstly guarantees the
bandwidth for UGS queue. This generally takes a fixed chunk
of bandwidth. The remaining uplink bandwidth Bpoll that is
allotted for the polling service will be

U U poll
B n - B = B -
(13)
A parameter , proposed in [26], is defined as the ratio of
the maximum time a rtPS or nrtPS MAC Protocol Data Unit
(MPDU) can wait in the queue (i.e. max_mpdu_delay) to the
maximum latency specification of the real-time flows.

flow _ rtPS _ of _ latency max_
delay _ mpdu max_
= o
(14)

This can be considered as a design parameter which
provides the present status of the queue. It also controls the
QoS of the real and non-real time services. The remaining
uplink bandwidth after allocation to UGS, are divided among
nr and nn numbers of real-time and non real-time flows as
follows [1, 26].
The real-time traffic flows are allotted an uplink bandwidth
of


n r
r
poll
r
tot
n * n
* n
* B B
+ o
o
=
(15)

The non real-time traffic flows are allotted an uplink
bandwidth of

r
tot poll
n r
r
poll
n
tot
B B
n * n
n
* B B =
+ o
=
(16)

Every SS repeats this process at the beginning of every
uplink. Since the bandwidth request/grant will take some time
due to RTD, the rtPS traffic can actually vary within this
period. This algorithm is not affected by the rtPS delay bound
traffic because the granted bandwidth is always redistributed.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
83 | P a g e
www.ijacsa.thesai.org
In this way this algorithm provides fair distribution of the
bandwidth.
I. Cross layer Architecture
Figure 4 shows the cross layer architecture proposed in this
paper. At PHY layer FFT Size =2048 and Cyclic Prefix=1/32
have been selected because these parameters provide highest
raw data rate and spectrum efficiency for all types of MCS as
shown in Table II. In addition, various MCS like QPSK-1/2,
QPSK-3/4, 16QAM-1/2, 16QAM-3/4, 64QAM-1/2, 64QAM-
2/3 and 64QAM-3/4 have been selected at PHY layer to
evaluate their individual performance in IEEE 802.16
networks.
Further, our proposed CAC algorithm and the Scheduling
algorithm have been implemented at MAC layer of BS and SS
respectively. The input parameters for the CAC algorithm are
the SINR for all the connections, a condition for providing
delay guarantee to the rtPS connection and availability of
bandwidth. Based on these input parameters, CAC algorithm
takes the proper decision to admit or reject an incoming
connection request along with its required bandwidth. The
output performance parameters obtained from the CAC
algorithm are NCBP, HCDP and COP. These output
performance parameters act as feedback information, in order
to select appropriate MCS dynamically at PHY layer to support
the required QoS guarantee.
This dynamic allocation of MCS at PHY layer based on
feedback information from MAC layer employs Adaptive
Modulation and Coding (AMC) technique in the network.
Scheduling algorithm takes current queue size and queuing
delay requirement as the input parameters to reschedule the
bandwidth shared between real-time and non real-time traffic.
It can be noticed that bandwidth which has already been
granted or scheduled by the CAC algorithm at BS is again
rescheduled by the Scheduling algorithm at SS among the
traffic sources. Bandwidth utilization and fairness index are the
output performance parameters obtained from the Scheduling
algorithm which reveals efficient utilization and fair allocation
of the system resources for all kind of MCS. Thereby, when
AMC technique selects a MCS in the network, the scheduling
algorithm distribute the bandwidth provided by that MCS
among the various traffic sources in the network. In this way,
Scheduling algorithm helps to improve utilization and fair
allocation of the system resources for all MCS.
III. ANALYTICAL MODEL
In this paper, the performance evaluation of the CAC
mechanisms is obtained by using the Continuous Time Markov
Chain (CTMC) Model [35]. The Markov model has been opted
because it examines the probability of being in a given state at
a given point of time, the amount of time a system is expected
to spend in a given state, as well as the expected number of
transitions between states.
A single WiMAX BS in isolation is considered. This BS
will receive the bandwidth requests from the SS within the
coverage area of a BS. Three types of services UGS, rtPS,
nrtPS need QoS guarantees and request for connection
admission. BS changes state from one to another upon the
admission or rejection of a connection. Further, it is assumed
that the BS either admits or rejects only one connection at a
particular instant of time. So the next state of the BS depends
only on the present state of the BS but does not depend on the
previous states of the BS.

Figure 4. Cross layer architecture for IEEE 802.16 BWA network
Therefore, the states of the BS form a Markov Chain, and
accordingly the BS can analytically be modeled as shown in
Figure 5. In this scenario, the BS can uniquely be represented
in the form of a five dimensional Markov Chain
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n
based on the number of admitted
connections of each type.

Figure 5. State Transition Diagram of the Markov Chain model for proposed
CAC Algorithm
State s =
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n
represents that the BS
has currently admitted n
u
, n
r
, and n
n
number of UGS, rtPS
and nrtPS connections respectively into the network. In Figure
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
84 | P a g e
www.ijacsa.thesai.org
5,
'
r
d
and
'
n
d
are the degraded bandwidth of rtPS and
nrtPS connections respectively after state transition and may
have a different value from that of
r
d and
n
d . The BS will
be in a particular State s =
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n
until a new
connection of one of them i.e. UGS, rtPS, nrtPS is admitted
into the network or an ongoing connection is terminated. The
arrival process of the handoff and newly originated UGS, rtPS,
and nrtPS connections is Poisson with rates ,
hu
,
hr
,
hn

,
ou
,
or
and
on
respectively. Therefore Total arrival rate
tu

,
tr

,
tn

corresponding to UGS, rtPS and nrtPS

connections have been evaluated as given below,

ou

hu
+ , if + +
U
1)B
u
(n
B
min
n
B
n
n
max
r
B
r
n s +






hu
, otherwise (17)


or

hr
+ , if
max
r
1)B
r
n (
U
B
u
n + + B
min
n
B
n
n s +


hr

, otherwise (18)


on

hn
+ , if
max
r
B
r
n
U
B
u
n + B
max
n
1)B
n
(n s + +





hn
, otherwise (19)
The service times of UGS, rtPS and nrtPS connections are
exponentially distributed with mean ,
u
1/
r
1/ and
n
1/
respectively.
The state space S for our proposed CAC scheme is obtained
based on the following equation.
S={
| )
n
d ,
n
n ,
r
d ,
r
n ,
u
(n s =
B )
n
d
n
n
r
.d
r
n
U
.B
u
(n s + +
)
max
n
B
n
d
min
n
(B )
max
r
B
r
d
min
r
(B s s . s s .
} (20)
From Figure 5, it is observed that every state s =
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n
in state space S is reachable from
every other state i.e. each state communicates with other states
in the state space S. As for example, state
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n

communicates with state
)
'
n
d ,
n
n ,
'
r
d 1,
r
n ,
u
(n +
and also communicates with
state
)
'
n
d 1, -
n
n ,
'
r
d ,
r
n ,
u
(n
. Hence state
)
'
n
d ,
n
n ,
'
r
d 1,
r
n ,
u
(n +
and
)
'
n
d 1, -
n
n ,
'
r
d ,
r
n ,
u
(n
can
also communicate to each other. In this way, each state can
communicate with other state in the state space S. Therefore,
the state space S forms a closed set and the Markov chain
obtained is irreducible [35].
Let the steady state probability of the state s =
)
n
d ,
n
n ,
r
d ,
r
n ,
u
(n
is represented by
) d , n , d , n , (n
n n r r u
t
.

As the Markov chain is irreducible, thereby observing the
outgoing and incoming states for a given state s, the state
balance equation of state s is shown in equation (21).
) 21 (
) z 1, y , x w, 1, (v

) z 1, y , x w, (v, n
1) (y
) z y, , x 1, w (v,

) z y, , x 1, w (v, r
1) (w
) z y, , x w, 1, (v

) z y, , x w, 1, (v u
1) (v
) z 1, y , x w, (v,

) z 1, y , x w, (v, tn

) z y, , x 1, w (v,

) z y, , x 1, w (v, tr

) z y, , x w, 1, (v

) z y, , x w, 1, (v tu

z) y, x, w, (v,
)
z) 1, y x, w, (v, n
y
z) y, x, 1, w (v, r
w
z) y, x, w, 1, (v u
v
z) 1, y x, w, (v, tn

z) y, x, 1, w (v, tr

w.x.y.z) 1, (v tu
(
' + ' + ' + '
+ +
' ' + ' ' +
+ +
' ' + ' ' +
+ +
' ' ' '
+
' ' ' '
+
' ' ' '

+
+
+
+
+
+

Where v, w, x, y and z represent , n

u
, n
r
, d
r
and n
n n
d
respectively.

z) y, x, w, (v,

represents the characteristic equation as
shown below.
1,

e z) y, x, w, (v, S

z) y, x, w, (v,

0, otherwise
By using equation (21), the state balance equations of each
state in the state space S are obtained. Solutions of these
equations provide the steady state probabilities of all states in
the space S with the normalized condition imposed by equation
(22).

e
=
S s
) d , n , d , n , (n
1 (s)
n n r r U
(22)

From the steady state probabilities we can determine
various QoS parameters of the system as given under.
A. New Connection Blocking Probability (NCBP)
The new connection blocking probability is the probability
of rejecting a new connection request for admission into the
network. Conditions for blocking a new connection request
have been included in Table IV.
- Estimation of NCBP for UGS, rtPS, nrtPS connections
While admitting a new connection, if the next state is not
allowable Markov chain state in the state space S due to the
conditions as given in Table IV, then the next state is


=
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
85 | P a g e
www.ijacsa.thesai.org
considered to be a blocked state for that connection i.e. the new
connection is blocked.
Let S
UB
, S
rB
and S
nB
form a state space for the states
whose next state are not allowed in the Markov chain due to
connection blocking.
The summation of the steady state probabilities of the states
in the state space S
UB
, S
rB
and S
nB

give the NCBP of UGS,
rtPS and nrtPS connections respectively.

Hence,

NCBP-UGS=
) s (
S s
UB
) d , n , d , n , (n
n n r r u

e
t
(23)


NCBP-rtPS= ) s (
S s
rB
) ,d n , ,d n , (n
n n r r u

e
t (24)
NCBP-nrtPS= ) s (
S s
nB
) d , n , d , n , (n
n n r r u

e
t

(25)
B. Handoff Connection Dropping Probability (HCDP):
The handoff connection dropping probability is the
probability of rejecting a handoff connection request for
admission into the network. Conditions for dropping a handoff
connection request are summarized in Table IV.
- Estimation of HCDP for UGS, rtPS and nrtPS
connections
Again, for the handoff connection if the conditions as given
in Table IV occur, then the handoff connection is dropped and
the next state is not allowed in the Markov chain due to
connection dropping.
Similar to NCBP, HCDP can also be estimated as below,
where S
UD
, S
rD
and S
nD
form a state space corresponding
to UGS, rtPS and nrtPS connection for the states whose next
state is not allowed in the Markov chain due to connection
dropping. Hence,
HCDP-UGS=
) s (
S s
UD
) d , n , d , n , (n
n n r r u

e
t
(26)
HCDP-rtPS=
) s (
S s
rD
) d , n , d , n , (n
n n r r u

e
t
(27)
HCDP-nrtPS= ) s (
S s
nD
) d , n , d , n , (n
n n r r u

e
t (28)
C. Connection Outage Probability (COP)
The connection outage probability is the probability that the
SINR of the connections in the network drops below a certain
threshold when a new connection wants to get admitted in the
network. Condition for connection outage in the network is
summarized in Table IV.
- Estimation of COP for UGS, rtPS and nrtPS
connection
A new or handoff connection is not admitted in the network
if SINR of the connection falls below its threshold value as
given in Table IV.
Similar to NCBP and HCDP, COP can be estimated as
shown below.
COP-UGS=
) s (
S s
UO
) d , n , d , n , (n
n n r r u

e
t

(29)

COP-rtPS = ) s (
S s
rO
) ,d n , ,d n , (n
n n r r u

e
t
(30)
COP-nrtPS =
) s (
S s
nO
) d , n , d , n , (n
n n r r u

e
t

(31)
As E
b
/N
0
ratio of the connections has the direct relationship
(e.g. Equation (6), (10) and (11)) to the SINR and SINR
th ,
so
from the proposed CAC mechanism and above defined
parameters it is observed that the E
b
/N
0
ratio provides
significant influence on NCBP, HCDP and COP for all the
connection types. The effect of E
b
/N
0
ratio on NCBP, HCDP
and COP will help to determine the optimum range of E
b
/N
0

ratio for all the connections, in order to select an appropriate
MCS to improve the channel condition. This dynamic selection
of MCS based on E
b
/N
0
ratio of the connections, employs
Adaptive Modulation and Coding (AMC) technique in the
network.
D. Bandwidth Utilization (BU):
The Bandwidth Utilization (BU) is defined [16] as the ratio
of total used bandwidth to the available bandwidth of the
system. BU of the system is encountered to estimate whether
SS lying in a bad channel state wastes precious bandwidth. BU
can be obtained as follows.
BU=
B
z) y, x, w, (v,
)
n
.d
S s
n
n
r
.d
r
n
U
.B
u
(n
e
+ +
(32)
Where, B is the total bandwidth.

Fairness Index
Fairness refers to the equal allocation of network resources
among the various users operating in both good and bad
channel states. Fairness is quantified using Jains Fairness
Index (JFI) [27] as given below.

=
=
-
=
n
1 i
2
i
2
n
1 i
i
r n
) r (
JFI
(33)
Where, r
i
is the data rate of connection type i.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
86 | P a g e
www.ijacsa.thesai.org
TABLE III. CONDITION FOR CONNECTION BLOCKING, DROPPING AND OUTAGE IN THE NETWORK
IV. PARAMETERS FOR PERFORMANCE EVALUATION
The frame duration (f) is taken as 1ms because it provides
more bandwidth utilization of the system compared to higher
length of frame duration as analyzed in our earlier work [19].
Total channel bandwidth (W) is taken as 20 MHz because this
is the maximum bandwidth supported by the WiMAX network
[36]. The FFT size is taken as 2048 with cyclic prefix 1/32 as
this combination provides highest raw data rate and spectrum
efficiency for all types of MCS as calculated in Table II.
As the IEEE 802.16 standards have not specified values for
the QoS parameters, we have considered the system parameters
for performance evaluation as given in Table V. The arrival
rates of all the connections are assumed to be same i.e.
on hn or hr oU hU
= = = = =
. The service time for
UGS, rtPS and nrtPS connections is exponentially distributed
with mean
n r U
1
,
1
,
1

respectively and we
assume
2 . 0
n r U
= = =
. That is, the mean service
time of the connections is taken as 5 sec. Service time of the
admitted connections should be as minimal as possible [16] in
order to provide access into the network by other users.
Degradation of rtPS and nrtPS connections is adaptive to the
required bandwidth.
TABLE IV. QOS PARAMETERS
Service
Max.Sustained
traffic rate
(kbps)
Min. reserved
traffic rate
(kbps)
Bucket
Size
(Bits)
Delay
(ms)
Eb/N0
(dB)

UGS 256 256 64 -
3.6
rtPS 1024 512 10240 21
6.3
nrtPS 1024 256 10240 -
8.1
V. NUMERICAL RESULTS
To evaluate the effectiveness and efficiency of the
proposed cross layer architecture, the IEEE 802.16 PHY and
MAC layer protocols are analyzed using MATLAB under
version 7.3. Exhaustive simulations are carried out.
Firstly, the performance of the CAC algorithm is verified
by using the CTMC model under various MCS. The effect of
E
b
/N
0
ratio on our proposed CAC mechanism is observed. The
optimum range of E
b
/N
0
ratio is determined to select an
appropriate MCS in order to employ AMC in the network.
Next, the impact of queue based scheduling algorithm is
examined on bandwidth utilization and Jains fairness index for
different connection arrival rates under various MCS.
Justifications behind all the numerical results have been
provided.
A. Impact of CAC Mechanism on QoS informationParameters
Figure 6 shows the blocking probability of UGS, rtPS and
nrtPS connections for different MCS. Total arrival rate of the
connections have been taken from equation (17), (18) and (19).
It is observed that 64QAM-3/4 provides least blocking
probability and QPSK-1/2 provides highest blocking
probability to the new connections to be admitted in the
network as compared to the other MCS. This is because of the
highest raw data rate (70.1299 Mbps) and spectrum efficiency
(3.5065) obtained by the 64QAM-3/4 among all other MCS
(Reference Table II). Again, it is observed that performances of
16QAM-3/4 and 64QAM-1/2 have been merged together
because of their same raw data rate.
Figure 7 shows the dropping probability of UGS, rtPS and
nrtPS connections for different MCS. Here, it is also observed
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
87 | P a g e
www.ijacsa.thesai.org
that, 64QAM-3/4 provides least dropping probability and
QPSK-1/2 provides highest dropping probability to the ongoing
hand off connections as compared to the other MCS because of
the same reasons as mentioned previously.
Next, connection outage probability of UGS, rtPS and
nrtPS are examined under different MCS and results are shown
in Figure 8. Outage probability is the probability that the SINR
of the connections in the network drops below a certain
threshold when a new connection wants to get admitted in the
network and thereby required QoS guarantee is not met to the
admitted connections. It is observed that 64QAM-3/4 provides
highest outage probability and QPSK-1/2 provides least outage
probability as compared to other MCS. Since QAM scheme has
the higher data rate and spectrum efficiency than the other
MCS, so it has the capability of accommodating more number
of connections. Thereby the SINR value of each admitted
connections in the network gets decreased and falls below the
threshold level, which results in the higher outage probability
of the connections in the system.
Therefore, QAM schemes provide good QoS in terms of
blocking probability of the newly admitted connections and
dropping probability of the hand off connections but fail to
provide good QoS in terms of connection outage probability of
the network. On the other hand, QPSK schemes provide good
QoS in terms of connection outage probability but fail to
provide good QoS in terms of blocking and dropping
probability.
Hence, there is a need for Adaptive Modulation and Coding
(AMC) technique whose performance is based on feedback
information of the MAC layer QoS parameters like connection
blocking probability, connection dropping probability and
connection outage probability. By dynamically changing the
allocation of modulation and coding rate based on Eb/N0 ratio
as well as SINR of the connections, AMC enables a flexible
use of the network resources that can support nomadic or
mobile operations.
B. Impact of Eb/N0 ratio on QoS information Parameters
To comprehend how modulation and coding schemes could
be made adaptive in a region under WiMAX cell boundary, the
proposed SINR based CAC scheme is again simulated under
various MCS. We consider Eb/N0 ratio of all the connections
because it has the influence on both SINR and SINRth of the
admitted connections in the network. Eb/N0 ratio is varied
between 1 dB to 20 dB because within this span we are able to
find out the adaptive range of all MCS. Connection arrival rate
is kept constant at 10 calls per second.
Figure 9 shows the blocking probability under various MCS
with different Eb/N0 ratio. Table VI shows the suitable range
of Eb/N0 ratio where different MCS show zero blocking
probability for all the connections. As for example, 64QAM-
3/4 provides zero blocking probability in the range 4 dB to 20
dB for UGS and rtPS connection but for nrtPS connection the
range is 9dB to 20 dB. So the optimum range of Eb/N0 ratio in
which 64QAM-3/4 provides zero blocking probability is 9dB
to 20 dB for all connections types. Again, we have taken 20 dB
as the maximum value of the Eb/N0 ratio in our simulation so it
is regardless to mention the maximum value of the Eb/N0 ratio
because the MCS can provide zero blocking probability beyond
this range.
Next, Figure 10 shows the dropping probability under
various MCS with different Eb/N0 ratio, whereas Table VI also
shows the range of Eb/N0 ratio where different MCS show zero
dropping probability for all the connections in our proposed
SINR based CAC scheme.
Figure 11 shows the outage probability under various MCS
with different Eb/N0 ratio and also Table VI shows suitable
range of Eb/N0 ratio in which different MCS show zero outage
probability for all the connections.
TABLE V. RANGE OF EB/N0 RATIO
Modulation and
Coding Scheme
Minimum
Eb/N0
ratio(dB) for
zero blocking
probability
obtained from
figure 9
Minimum
Eb/N0
ratio(dB) for
zero dropping
probability
obtained from
figure 10
Maximum
Eb/N0
ratio(dB) for
zero outage
probability
obtained from
figure 11
64QAM-3/4 9 4 1
64QAM-2/3 10 4 1
64QAM-1/2 14 6 2
16QAM-3/4 14 6 2
16QAM-1/2 20 9 3
QPSK-3/4 - 14 4
QPSK-1/2 - - 7

By analyzing Table VI, it can be observed that, QAM
schemes perform well at higher E
b
/N
0
ratio in terms of zero
blocking probability (range is above 9 dB ) and zero dropping
probability (range is above 4 dB) and QPSK schemes perform
well at lower E
b
/N
0
ratio (range is below 7 dB) in terms of zero
outage probability. However, zero probability is not achievable
in the practical scenario owing to high intense traffic arrival
rate in comparison to 10 calls per second which we have taken
for our simulation. Despite of it, the evaluated adaptive range
of E
b
/N
0
ratio will provide minimum blocking, dropping and
outage probability in the practical scenario.
Hence, it is better to select higher order MCS like QAM
near to the WiMAX base station where SINR of the
connections as well as population density of the users are more
and lower order MCS like QPSK far away from the base
station where SINR of the connections as well as population
density of the users are less. This adaptive allocation of MCS
makes the user to experience less blocking, less dropping and
less outage in their connection request. Thus, AMC is
employed in the network that enables dynamic use of the
network resources to support nomadic or mobile operation.
Therefore, cross layer adaptation of the different modulation
capability at the PHY layer with the QoS requirement at the
MAC layer is obtained.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
88 | P a g e
www.ijacsa.thesai.org

Impact of CAC Mechanism on QoS Information Parameters

(a) (b) (c)
Figure 1. Blocking probability of UGS connections, (b) Blocking probability of rtPS connections and (c) Blocking probability of nrtPS connections

(a) (b) (c)
Figure 7(a) Dropping probability of UGS connections, (b) Dropping probability of rtPS connections and (c) Dropping probability of nrtPS connections

(a) (b) (c)
Figure 8(a) Outage probability of UGS connection, (b) Outage probability of rtPS connections and (c) Outage probability of nr tPS connection
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
89 | P a g e
www.ijacsa.thesai.org
Impact of E
b
/N
0
ratio on QoS Information Parameters


(a) (b) (c)
Figure 9 (a) Blocking probability of UGS connections, (b) Blocking probability of rtPS connections and (c) Blocking probability of nrtPS connection

(a) (b) (c)
Figure 10(a) Dropping probability of UGS connections, (b) Dropping probability of rtPS connections and (c) Dropping probability of nrtPS connections

(a) (b) (c)
Figure 11(a) Outage probability of UGS connections, (b) Outage probability of rtPS connections and (c) Outage probability of nrtPS connections
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
90 | P a g e
www.ijacsa.thesai.org
C. Impact of Queue Based Scheduling on QoS Parameters
To demonstrate the advantage of rescheduling of bandwidth
in our proposed cross layer architecture, further simulation has
been carried out for different connection arrival rates under
various MCS.

Figure 12(a). Average Bandwidth utilization before queue based rescheduling

Figure 12 (b). Average Bandwidth utilization after queue based rescheduling

Figure 12(a) and 12(b) show the bandwidth utilization
before and after queue based scheduling scheme respectively
under different MCS. It shows the significant performance
improvement in the bandwidth utilization due to rescheduling
of the resource sharing between real-time and non real- time
traffic based on their current queue size and latency
requirements. From Figure 12(b), it can be observed that
bandwidth utilization seems to be independent of MCS i.e.
High bandwidth utilization for all MCS. It is because of the
influence of scheduling algorithm at SS. As bandwidth is
considered to be a limited resource in the network,
rescheduling of bandwidth at SS will automatically improve the
revenues of the service providers.
On the other hand, Figure 13(a) and 13(b) show the Jains
fairness index before and after the queue based scheduling
scheme respectively under different MCS. It is also observed
that substantial performance improvement in the fair allocation
of resources between real-time and non real-time traffic is
achieved because of the rescheduling of the resources. Hence,
joint performance of the CAC and rescheduling algorithms
also provides better resource utilization at lower traffic arrival
rate.

Figure 13(a). Jains fairness index before queue based rescheduling

Figure 13(b). Jains fairness index after queue based rescheduling

Percentage improvements of the performances after
rescheduling of bandwidth for all MCS at traffic arrival rate =
1 are shown in Table VII. Highest performance improvement
is observed in case of 64QAM-3/4 as compared to the other
MCS. This is because 64QAM-3/4 has highest data rate and
spectrum efficiency as calculated in Table II.
TABLE VI. PERCENTAGE IMPROVEMENT IN QOS AFTER RESCHEDULING
OF BANDWIDTH WHEN TOTAL ARRIVAL RATE = 1
Modulation and
Coding Scheme
Percentage
improvement in the BU
Percentage
improvement in the
JFI
QPSK-1/2
2.3813
7.4288
QPSK-3/4 10.0498 20.1068
16QAM-1/2 18.8653 29.4374
16QAM-3/4 43.2356 43.3220
64QAM-1/2 42.2646 45.8475
64QAM-2/3 76.5587 46.6461
64QAM-3/4 96.6757 49.1656
When AMC technique selects an MCS in the network based
on the feedback information obtained from the CAC algorithm,
the scheduling algorithm redistribute the bandwidth of that
selected MCS among the various traffic sources in the network.
In this way, scheduling algorithm helps to improve utilization
and fair allocation of the system resources for all MCS in IEEE
802.16 BWA networks.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
91 | P a g e
www.ijacsa.thesai.org
VI. CONCLUSION
Cross layer adaptations are essential for guaranteeing QoS
supports in real-time multimedia traffic over wireless
networks. In this paper, a cross layer architecture for adapting
different MCS in PHY layer has been considered by
incorporating MAC layer information in terms of New
Connection Blocking Probability (NCBP), Hand off
Connection Dropping Probability (HCDP) and Connection
Outage Probability (COP) for WiMAX BWA systems. In
literatures, many researchers have proposed various cross
layer mechanisms for providing better QoS support to the
system, but so far no such comprehensive cross layer design
considering the parameters stated above, has yet been reported
in the literature. In this work, SINR based CAC integrated
with the Queue based Scheduling has been analysed with
Markov Chain model. The effect of E
b
/N
0
ratio is observed on
NCBP, HCDP and COP for all the connection types under
various MCS by means of exhaustive simulation. Also
optimum range of E
b
/N
0
ratio is determined, in order to select
an appropriate MCS which may be used as the threshold
parameters for SINR in any adaptive modulation scheme.
Moreover, the joint performance of the CAC and Scheduling
algorithms has been proved to be good enough to meet the
QoS requirements in terms of bandwidth utilization and Jains
fairness index.
ACKNOWLEDGEMENT
The authors deeply acknowledge the support from DST,
Govt. of India for this work in the form of FIST 2007 Project
on Broadband Wireless Communications in the Department
of ETCE, Jadavpur University.
REFERENCES
[1] Y. Zhang, WiMAX Network Planning and Optimization, CRC Press,
Taylor and Francis Group, 2009.
[2] WiMAX Forum, http://www.wimaxforum.org/home/.
[3] Y.J. Chang, F.T. Chien, C.C.J. Kuo, Cross layer QoS Analysis of
Opportunistic OFDM-TDMA and OFDMA Networks, IEEE Journal on
Selected Areas in Commun. Vol. 25(4) pp. 657 666, 2007.
[4] H.K. Rath, A. Karandikar, V. Sharma, Adaptive Modulation-Based
TCP-Aware Uplink Scheduling in IEEE 802.16 Networks, IEEE
International Conference on Commun., pp. 3230-3236, 2008.
[5] D. Marabissi, D.Tarchi, R. Fantacci, F. Balleri, Efficient Adaptive
Modulation and Coding Techniques for WiMAX Systems IEEE
International Conference on Communications, ICC '08, pp. 3383 3387,
2008.
[6] V. Nitnaware, S. Lande, S. Balpande, Comparative Study of Various
Modulation Techniques used in Point to Multipoint Mode for WiMAX
Published in International Journal of Advanced Engineering &
Applications, pp. 46-56,Vol.1, Jan. 2010.
[7] B. Rong, Y. Qian, H. H. Chen, Adaptive power allocation and call
admission control in multiservice WiMAX access networks IEEE
Wireless Commun., Vol. 14(1), pp. 14 19, 2007.
[8] F. D. Priscoli, T. Inzerilli, L. Munoz, QoS Provisioning in Wireless IP
Networks, Wireless Personal Communication, Springer, pp. 23-39,
2006.
[9] L. Litwin, M. Pugel, The Principles of OFDM, www.Rfdesign.com,
Jan 2001.
[10] J. G. Andrews, A. Ghosh, R. Muhamed, Fundamentals of WiMAX-
Understanding Broadband Wireless Networking, PEARSON
Education, March 2007.
[11] M. H. Ahmed, Call Admission Control in Wireless Networks: A
comprehensive Survey, IEEE Communications Surveys & Tutorials,
First Quarter, Vol.7 (1) pp. 50-69, 2005.
[12] H. Wang, W. Li and D. P. Agrawal, Dynamic admission control and
QoS for IEEE 802.16 Wireless MAN, Proc. of Wireless
Telecommunications Symposium (WTS 2005), pp. 60-66 April 2005.
[13] F. Hou, P. H. Ho, X. Shen, Performance Analysis of Reservation Based
Connection Admission Scheme in IEEE 802.16 Networks, Global
Telecommunications Conference, GLOBECOM06, pp. 1-5, November
2006.
[14] L. Wang, F. Liu, Y. Ji, N. Ruangchaijatupon, Admission Control for
Non-preprovisioned Service Flow in Wireless Metropolitan Area
Networks, Universal Multiservice Networks, (ECUMN-07). Fourth
European Conference, pp. 243 249, Feb. 2007.
[15] Y. Ge, G.S. Kuo, An Efficient Admission Control Scheme for Adaptive
Multimedia Services in IEEE 802.16e Networks, IEEE 64th Vehicular
Technology Conf., pp.1 5, Sept. 2006.
[16] K. Suresh, I. S. Misra and K. Saha (Roy), Bandwidth and Delay
Guaranteed Call Admission Control Scheme for QOS Provisioning in
IEEE 802.16e Mobile WiMAX Proceedings of IEEE GLOBECOM,
pp.1245-1250, December 2008.
[17] C. H. Jiang, T.C. Tsai, CAC and Packet Scheduling Using Token
bucket for IEEE 802.16 Networks, Consumer Communications and
Networking Conf., 2006. CCNC 2006. 3rd IEEE Vol. 1, pp. 183-187,
Jan. 2006.
[18] K. Wongthavarawat, A. Ganz, Packet scheduling for QoS support in
IEEE 802.16 broadband wirelessaccess systems, International Journal
of Communication Systems, Vol. 16(1), 81-96, 2003.
[19] P. Chowdhury, I. S. Misra, S. K. Sanyal, An Integrated Call Admission
Control and Uplink Packet Scheduling Mechanism for QoS Evaluation
of IEEE 802.16 BWA Networks, Canadian Journal on Multimedia and
Wireless Networks, Vol.1 (3), 2010.
[20] P. Chowdhury, I. S. Misra, A Comparative Study of Different Packet
Scheduling Algorithms with Varied Network Service Load in IEEE
802.16 Broadband Wireless Access Systems IEEE Proc. Int. Conf.
Advanced Computing & Communications, Bangalore, Dec. 2009.
[21] J.Lin, H.Sirisena, Quality of Service Scheduling in IEEE 802.16
Broadband Wireless Networks, Proceedings of First Int. Conf. on
Industrial and Information Systems, pp.396-401, August 2006.
[22] C.Cicconetti, A.Erta, L.Lenzini, E.Mingozzi, Performance Evaluation
of the IEEE 802.16 MAC for QoS Support, IEEE Transactions on
Mobile Computing, Vol. 6(1), pp. 26-38, 2007.
[23] P. Chowdhury, I. S. Misra, An Efficient Quality of Service Scheduling
Strategy for IEEE 802.16 Broadband Wireless Access Systems,
proceedings of Int. conf. NeCoM-2010, Recent Trends in Netwoks and
communication, springer, Vol. 90, pp. 306-315, July 2010.
[24] W. Ye, A. M. Haimovich, Outage Probability of Cellular CDMA
Systems with Space Diversity, Rayleigh Fading, and Power Control
Error, IEEE Communications Letters, Vol.2 (8), pp. 220-222, 1998.
[25] C. S. Chu, D. Wang, S. Mei, A QoS architecture for the MAC protocol
of IEEE 802.16 BWA system, IEEE communications, Vol.1, pp. 435-
439, 2002.
[26] K. R. Raghu, S. K. Bose, M.Ma, Queue based scheduling for IEEE
802.16 wireless broadband, proceedings of ICICS, Vol. 5(6), pp.1-5,
2007.
[27] A. Haider, R. Harris, A novel proportional fair scheduling algorithm for
HSDPA in UMTS networks, Proceedings of 2nd IEEE AusWireless,
pp.43-50, 2007.
[28] S. E. Elayoubi, B. Fourestie, Performance Evaluation of Admission
Control and Adaptive Modulation in OFDMA WiMax Systems
IEEE/ACM Transaction on Networking, Vol. 16 (5), pp. 1200-1211,
2008.
[29] M. Poggioni, L. Rugini, P. Banelli, QoS Analysis of a Scheduling
Policy for Heterogeneous Users Employing AMC Jointly with ARQ
IEEE Transaction on Communication, Vol. 58(9), pp. 2639-2652, 2010.
[30] Q. Liu, X. Wang, G.B. Giannakis, A cross layer scheduling algorithm
with QoS support in wireless networks. IEEE Transactions on Vehicular
Technology, Vol. 55(3), pp. 839-84 ,2006.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
92 | P a g e
www.ijacsa.thesai.org
[31] D. Triantafyllopoulou, N. Passas, A.K. Salkintzis, A. Kaloxylos, A
cross layer optimization mechanism for multimedia traffic over IEEE
802.16 networks. International Journal of Network Management, Vol.
17(5), pp. 347-361, 2007.
[32] D. Triantafyllopoulou, N. Passas, A.K. Salkintzis, A. Kaloxylos, A
Heuristic Cross layer Mechanism for Real-Time Traffic in IEEE 802.16
Networks , IEEE 18th International Symposium on Personal, Indoor
and Mobile Radio Communications, pp. 1-5, September 2007.
[33] X. Li, X. Wu, W. Li, X. Wang, An Adaptive Cross layer Scheduling
Algorithm for Multimedia Networks International Conference on
Intelligent Information Hiding and Multimedia Signal Processing, , pp.
52-55, 2008.
[34] K. A. Noordin, G. Markarian, Cross layer Optimization Architecture for
Wimax Systems, IEEE 18th International Symposium on Personal,
Indoor and Mobile Radio Communications, pp. 1-4, 2007.
[35] S. M. Ross. Probability Models for Computer Science, Elseveir, June
2001.
[36] QualNet 4.5 Advanced Wireless Model Library , Scalable Network
Technologies, Inc., http://www.scalable-networks.com.
[37] K. Feher, Wireless Digital Communications: Modulation and Spread
Spectrum Application, Prentice-Hall, 1995.
[38] C. T. Chou, K.G. Shin, Analysis of adaptive bandwidth allocation in
wireless networks with multilevel degradable quality of service IEEE
Transactions on Mobile Computing, Vol. 3, pp. 5 17, 2004.
AUTHORS PROFILE
1. Mr. Prasun Chowdhury (prasun.jucal@gmail .com) has completed his
Masters in Electronics and Telecommunication Engineering from
Jadavpur University, Kolkata, India in 2009. Presently he is working as
Senior Research Fellow (SRF) in the Department of Electronics and
Telecommunication Engineering, Jadavpur University, Kolkata, India.
His current research interests are in the areas of Call Admission control
and packet scheduling in IEEE 802.16 BWA Networks. He has authored
some journals and international conference papers.
2. Dr. Iti Saha Misra ([email protected]) received her PhD in
Microstrip Antennas from Jadavpur University (1997). She is presently a
Professor in the Department of Electronics and Telecommunication
Engineering, Jadavpur University, India. Her current research interests
include Mobility and Location Management, Next Generation Wireless
Network Architecture and protocols, Call Admission control and packet
scheduling in cellular and WiMAX networks. She has authored more
than 100 research papers in refereed Journal and International
Conference and a book on Wireless Communications and Networks. She
is an IEEE Senior Member and founder Chair of the Women In
Engineering, Affinity Group, IEEE Calcutta Section.
3. Dr. Salil K. Sanyal ([email protected]) received his Ph.D from
Jadavpur University, India (1990). He is currently a Professor in the
Department of Electronics and Telecommunication Engineering,
Jadavpur University. He has authored more than 130 Research Papers in
refereed Journals and International/National Conference Proceedings and
also co-authored the Chapter Architecture of Embedded Tunable
Circular Microstrip Antenna in the book entitled Large Scale
Computations, Embedded Systems and Computer Security. He is a
Senior Member of IEEE and past Chair of IEEE Calcutta Section. His
current research interests include Analog/Digital Signal Processing,
VLSI Circuit Design, Wireless Communication and Tunable Microstrip
Antenna.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
93 | P a g e
www.ijacsa.thesai.org
A Conceptual Design Model for High Performance
Hotspot Network Infrastructure (GRID WLAN)
Udeze Chidiebele. C, Okafor Kennedy .C
R & D Department, Electronics Development Institute
(FMST-NASENI), Awka, Nigeria.
Prof. H. C Inyiama, Dr C. C. Okezie,
Electronics and Computer Engineering Department,
Nnamdi Azikiwe University, Awka, Nigeria

AbstractThe emergence of wireless networking technologies
for large enterprises, operators (service providers), small-medium
organizations, has made hotspot solutions for metropolitan area
networks (MAN), last mile wireless connectivity, mobile
broadband solutions, IP-based cellular phones (VOIP) and other
event-based wireless solutions in very high demand. Wireless radio
gateways (routers and access points) with its wide-spread
deployment has made Wi-Fi an integral part of todays hotspot
access technology in organizational models. Despite its role in
affecting performance for mobility market segments, previous
research has focused on media access control (MAC) protocol
techniques , carrier sense multiple access with collision avoidance
(CSMA/CA), fair scheduling, and other traffic improvement
methodologies without detailed consideration for virtual switch
partitioning for load balancing, reliable fragmentation capacities,
buffer size dependencies, load effects, queuing disciplines as well
as hotspot access control framework. This paper proposes a
conceptual high performance hotspot solution (GRID WLAN) and
through simulations with OPNET modeler, presents efficient
performance metrics for its deployment. Considering the GRID
WLAN access points in context, the results shows that with the
design model, a careful selection of buffer sizes, fragmentation
threshold, network management framework with network load
intensity will guarantee an efficient hotspot solution.
Keywords: Hotspot solutions, IP-based cellular phones, Buffer size
dependencies, GRID WLAN.
I. INTRODUCTION
The mobility market segments has made modern computing
very attractive as such creating a platform for application
developers INTEGRATE solutions in laptops, cloud computing
environments as well as mobile devices. In this research, a
conceptual high performance hotspot solution termed GRID
WLAN is shown in Fig. 1. Anyone with a PC, notebook,
mobile phone or PDA that is Wi-Fi-enabled can use the
proposed GRID WLAN hotspot services.
From the user perspective, the network can be selected,
browser launched and the user signs up by entering an
authentication ID or by using a signed credit card for access
control. The network comprises the infrastructure gateway
(ADSL modem and GRID WLAN Gateway), Hypervisor layer
(Virtualization firmware), GRID WLAN switch/load balancer,
GRID WLAN Access points (AP1.APn), Management
framework, Wimax base station, VOIP, and hotspot nodes.
The network model adopted in this research is an extended
service set GRID WLAN mode. The APs establishes
connections to other users who are directly connected to the
hotspots. Basically, FTP traffic and HTTP traffic take
place in the GRID WLAN hotspot zone sites while the GRID
WLAN gateway access point creates a gateway link to the IP
cloud (internet). The scope of this paper is majorly on the
capacity of the various buffer sizes, loads as well as
fragmentation threshold of the GRID WLAN APs as the
principal metrics of interest in the proposed GRID WLAN
design. This paper defines capacity in this context to be the
maximum number of simultaneous, bidirectional traffic flow
that can be supported in the hotspot environment.
The paper is organized as follows: In section II, the
literature review was discussed, the general system model and
assumptions for GRID WLAN hotspot was presented in III. In
IV, GRID WLAN mechanism is presented. Section V gives the
simulation results to support our propositions. The paper ends
with the conclusions and future directions.
II. LITERATURE REVIEW
A detailed study on various works on hotspot WLAN
solutions was carried out. However, a close attention was given
to the work in [1] which is the most related work. The
authors focused on WLAN performance issues viz: tuning the
physical layer related parameters, MAC protocol techniques,
tuning the IEEE 802.11 parameters, fair scheduling, and
Carrier Sense Multiple Access with Collision Avoidance
(CSMA/CA) and other traffic improvement methodologies.
A good introduction to the 802.11 standard, followed by
a performance study of both the Distributed Coordination
Function (DCF) and Point Coordination Function (PCF) is
presented in [2]; this study suggests that an IEEE 802.11
network may be able to carry traffic with time-bounded
requirements using the PCF. In Point Coordination Function
(PCF) mode, polling occurs with a point coordinator
determining which station has the right to transmit [3]. In the
DCF mode, a BSS operates as an ad-hoc network in which any
station can communicate with any other station in the BSA
without the intervention of a centralized access point (AP) by
contending for a shared channel. All stations have equal
priorities and hence an equal chance of getting the channel [3].
Basically, the current use of WLANs for Internet access
from wireless stations is dominated by downlink TCP traffic;
however, the optimal performance of WLAN setups have
not been sufficiently taken into account for data transport over
WLANs devices like APs. WLAN parameters can be fine
tuned to optimize performance. This has not been addressed in
the previous literature reviews [4], [5].This paper shall focus on
metrics that will enhance performance in a hotspot setup as
shown in our system model in Fig.1.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
94 | P a g e
www.ijacsa.thesai.org
III. SYSTEM DESIGN AND ASSUMPTIONS
A. System Model
The high performance GRID WLAN hotspot solution is
an extended service set wireless configuration shown in Fig. 1.
It assumed that the GRID WLAN gateway is a composite
powerful gateway combined with radio which acts as a
complete hotspot. At the core of the proposed GRID WLAN is
a switch load balancer (bridge network) which acts as a
distribution medium for all indoor access points, wired LAN,
outdoor hotspots nodes, wimax base station, back office server
(management framework). This paper practically details the
implementation methodology as well as discusses the GRID
WLAN mechanism in the section IV.
B. Design Goals
In designing our proposed GRID WLAN hotspot solution,
the main goals are to maintain throughput and to create a
wireless hotspot solution that is based on robust network
infrastructure, hence scalable, flexible and quick to deploy at
anytime with less administrative overhead. Fig. 1 shows the
proposed system model.
As shown in Fig 1, the core LAN switch with virtual OS
connects the ADSL modem, management framework ,GRID
WLAN access points (for indoor coverage site), the Wimax
base station (outdoor coverage site), local hotspot nodes
(hsn1.....hns6) , and VOIP. For our large hotspots including
mesh nodes, the high powered GRID WLAN gateway will be
capable of supporting over 500 simultaneous user sessions. In
this context, we assume the GRID WLAN hotspot to be a
combination of wired access points, Wimax base station, and
indoor and outdoor mesh access point nodes dependent on the
local site.

Figure 1. The proposed GRID WLAN Hotspot system model
GRID WLAN Access Point Array- Indoor Coverage
Management Framework
hn1 hn2 hn3 Wimax hn4 hn5 hn6
ADSL MODEM
Grid WLAN Gateway
BRIDGE NETWORK
SWITCH/LOAD
BALANCER

*PORTAL *MANAGMNT *VPN
*GRID WLAN SERVER
AP
1
AP2 AP
3
Vsw1
Virtualization OS
AP
4
AP
5
AP
7
AP
6
VOIP
Vsw2
Vsw3
IP CLOUD
Virtualization Firmware
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
95 | P a g e
www.ijacsa.thesai.org
C. Experimental Test Bed
In this paper, to measure the system performance of the
proposed GRID WLAN hotspot network infrastructure,
OPNET Modeler [6] was used to achieve the objective.
OPNET Modeler is a graphical network simulator mainly used
for simulation of both wired and wireless communication
networks and information systems. For our proposed test bed,
the infrastructure components include:
i. 5 Hotspots nodes for outdoor Coverage
ii. Bridge Switch/Load balancer with virtual OS
(Extended service mode).
iii. 5 Access Points for indoor coverage
iv. 40 WLAN client Stations
v. Management Server (Enterprise Red hart Server)
vi. Back office Applications (Network Monitoring,
Configuration, Authentication, Authorization and
Accounting ) :
- Grid Control Management server for central
network configuration & Monitoring.
- Portal Server for Network Services and
management.
- VPN tunnel Terminator server
In Fig 1, the proposed GRID WLAN Hotspot system
model utilizes the infrastructure components outline above. A
management server has the role authenticating and monitoring
the overall network for efficient service delivery. The
extended service mode of the bridge switch enables both indoor
and outdoor connectivity. The virtual switching by the
switch bridge interfaces the management server, hotspot
nodes, access points and client stations in our context as shown
in Fig. 2a. The Access points were setup in the various subnets
with the client base stations (Fig. 2b). The bridge switch is a
speed redundancy layer supporting virtualization and load
balancing. Access to the IP internet cloud is completed by
the gateway and modem. The network model, node model and
process models were accomplished in our test bed using
OPNET modeller. After setting up the model, a simulation
run was carried out to generate our graphical plots shown in
this work (Fig 3). Also, a consistency test carried out shows
that the design model is stable and consistent before the
simulation execution.
IV. DISCUSSION
From Fig. 1, the GRID WLAN hotspot solution employs
the concept of mesh routing to create a self configurable,
centrally controlled hotspot solution. Every GRID WLAN
access point (AP) is capable of relaying traffic coming from
any of its neighbours. The mesh network reconfigures itself
when a node is added or removed. By adding a GRID WLAN
access point (AP), the wireless network range is expanded.
Owing to the self-reconfigurable mesh routing algorithms, the
network has no point of failure, hence, a reliable wireless
platform for connectivity
The GRID WLAN gateway router supports the following
functionalities: Automatic Configuration, Bandwidth
Control/Optimization, Firewall, Multiple Services, and Virus
Checking with Extensive Access list control.

Figure 2. GRID WLAN testbed Implementation with subnetted clients

Figure 2b: GRID WLAN nodes

Figure 3. shows the GRID WLAN Simulation run
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
96 | P a g e
www.ijacsa.thesai.org
V. SIMULATION PARAMETERS
In this section, we provide simulation results that support
section III. The GRID WLAN APs including the gateway
router connects wireless network (nodes) to wired network to
wired networks via the bridge switch as depicted in our model.
A. Simulation Configurations
We used OPNET modeler [6] was used to generate the
parameters for various case scenarios in the simulations.
Traffic attributes are listed in Table 1, while Table 2 show the
GRID WLAN parameters used and the various load intensities
from OPNET Modeler.
TABLE 1. OPNET TRAFFIC ATTRIBUTES
CONFIGURATIONS Values
1. Simulation Duration for Each
Scenario
60mins
2. Link Propagation Delays 0.5usecs
3. Switch Output Buffer 100 packets
4. Simulation Seed 128
5. Update Interval 50000 Events
6. Simulation Kernel Optimized

TABLE 2. GRID WLAN SIMULATION PARAMETER
WLAN
PARAMETERS

5 Sources
WLAN
BUFFER-
64K
10 Sources
WLAN
BUFFER-
128K
15 Sources
WLAN
BUFFER-
256K
20 Sources
WLAN
BUFFER-
512K
30 Sources
WLAN
BUFFER-
1024K
40 Sources
WLAN
BUFFER-
2048K
RTS-Threshold
(Bytes)
256 Bytes 256Bytes 256Bytes 256Bytes 256Bytes 256Bytes
Fragmentation
Threshold (Bytes)
256 Bytes 512Bytes 1024Bytes 2048Bytes 4096Bytes 8192Bytes
Data Rate (bps) 54 54 54 54 54 54

Physical
Characteristics
DSSS (D3S) DSSS DSSS DSSS DSSS (D3S) DSSS (D3S)
Packet-Reception
Power
Threshold(W)
7.33e-14 7.33e-14 7.33e-14 7.33e-14 7.33e-14 7.33e-14
Short Retry Limit 7 7 7 7 7 7
Long Retry Limit 4 4 4 4 4 4
AP Functionality Active Active Active Active Active Active
Buffer Size (bits) 64K 128k 256k 512k 1024K 2048K
Max-Receive
Lifetime (Sec)
0.5 0.5 0.5 0.5 0.5 0.5
Large-Packet
Processing
Drop Drop Drop Drop Drop Drop

TABLE 3. GRID WLAN SCENARIO TABLE

Fragmentation
Threshold (Bytes)
Load
Intensities/Sources
Buffer Sizes
(Bytes)
Scenarios/Cases
256 5 64K CASE-1
512 10 128k CASE-2
1024 15 256k CASE-3
2048 20 512k CASE-4
4096 30 1024k CASE-5
8192 40 2048k CASE-6


B. Performance Evaluation
The simulation was run accordingly and the statistics
collected for various cases and analyzed. As claimed before,
the proposed GRID WLAN hotspot is a high performance



solution that is robust and scalable. The paper used a case-
based scenario for various load intensities as well as
Subnet1
Subnet2
Subnet3
Subnet4
Subnet5
Subnet6
Subnet7





(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
97 | P a g e
www.ijacsa.thesai.org
fragmentation threshold and buffer sizes as depicted in Table
3.0
To demonstrate the effects of Fragmentation Threshold, we
employed the 5 node simulation scenarios with combinations
of values for Fragmentation Threshold and RTS as well as
other parameters. After the first case simulations, the
throughput was collected for analysis. The throughput is the bit
rate sent to the higher layer. It represents the rate of data
successfully received from other stations. At the offered load
from one node source to five node sources, it was shown that
for a buffer size of 64 bytes for the FTP and HTTP traffic, the
throughput had a slight deviation at 89% showing a significant
network performance.
Fig 7: shows the media access delay for offered Load time
against fragmentation threshold Fragmentation threshold is an
important parameter that affects WLAN performance. It is used
to improve the WLAN performance when the media error rate
is high.

Figure 6: A plot of Offered Load Versus Processed Throughput for a
64k buffer size

Figure 7: Offered Load time against fragmentation threshold

From Fig. 7, the effect of Fragmentation threshold on the
WLAN performance was seen to be insignificant considering
the offered load times. This could result from the selected
threshold in the APs and client machines. For the first five
simulation scenarios, parameters for Fragmentation Threshold
are listed in Table 2. The simulation results indicate that for
low bit error rates (2x10-5), various fragmentation thresholds
(256 bytes, 512 bytes, or no fragmentation limit ) have no
significant effect on the WLAN performance.

Figure 8: A plot of Offered Load Versus Processed Throughput for a
512k buffer size

Figure 9: A plot of WLAN Load Against fragmentation threshold for a
512k buffer size.
Fig. 8 shows the throughput effect on offered load intensity
for a 512k buffer size. From one to five workstations, there was
a gradual exponential response until at a saturation point of
about 87% , then it now begin to maintain a constant
throughput with additional loads. For, every network, traffic
optimization can only be achieved when the network
throughput is significantly acceptable or high.
The simulation results in Fig. 9 illustrates that with an
adaptive back-off mechanism, the offered load between WLAN
stations can be greatly decreased with increase in the
0
36
72
108
144
180
216
252
288 324 360 396 432 468 504 540 576 612 648 684 720 756 792 828 864 900 936 972 1008 1044 1080 1116 1152 1188 1224 1260 1296 1332 1368 1404 1440 1476 1512 1548 1584 1620 1656 1692 1728 1764 1800 1836 1872 1908 1944 1980 2016 2052 2088 2124 2160 2196 2232 2268 2304 2340 2376 2412 2448 2484 2520 2556 2592 2628 2664 2700 2736 2772 2808 2844 2880 2916 2952 2988 3024 3060 3096 3132 3168 3204 3240 3276 3312 3348 3384 3420 3456 3492 3528 3564 3600 160 80
53.333333
33
40 32
26.666666
67
22.857142
86
20
17.777777
78
16
14.545454
55
13.333333
33
12.307692
31
11.428571
43
10.666666
67
10
9.4117647
06
8.8888888
89
8.4210526
32
8
7.6190476
19
7.2727272
73
6.9565217
39
6.6666666
67
6.4
6.1538461
54
5.9259259
26
5.7142857
14
5.5172413
79
5.3333333
33
5.1612903
23
5
4.8484848
48
4.7058823
53
4.5714285
71
4.4444444
44
4.3243243
24
4.2105263
16
4.1025641
03
4
3.9024390
24
3.8095238
1
3.7209302
33
3.6363636
36
3.5555555
56
3.4782608
7
3.4042553
19
3.3333333
33
3.2653061
22
3.2
3.1372549
02
3.0769230
77
3.0188679
25
2.9629629
63
2.9090909
09
2.8571428
57
2.8070175
44
2.7586206
9
2.7118644
07
2.6666666
67
2.6229508
2
2.5806451
61
2.5396825
4
2.5
2.4615384
62
2.4242424
24
2.3880597
01
2.3529411
76
2.3188405
8
2.2857142
86
2.2535211
27
2.2222222
22
2.1917808
22
2.1621621
62
2.1333333
33
2.1052631
58
2.0779220
78
2.0512820
51
2.0253164
56
2
1.9753086
42
1.9512195
12
1.9277108
43
1.9047619
05
1.8823529
41
1.8604651
16
1.8390804
6
1.8181818
18
1.7977528
09
1.7777777
78
1.7582417
58
1.7391304
35
1.7204301
08
1.7021276
6
1.6842105
26
1.6666666
67
1.6494845
36
1.6326530
61
1.6161616
16
1.6
0%
20%
40%
60%
80%
100%
120%
1
1
3
2
5
3
7
4
9
6
1
7
3
8
5
9
7
Offered Load intensity
Throughput Effect for a 64k buffer
size
Wireless
LAN.Throu
ghput
(bits/sec).n
one
FTP & HTTP traffic
in BSS Mode
5 Nodes
0
500
1000
1500
2000
2500
3000
3500
4000
0 1 2 3
R
T
S

t
h
r
e
s
h
o
l
d

Offered Load Time
Wireless LAN.Media Access Delay (sec)
Series
1
Series
2
Series
3
Series
4
Series
5
Series
6
TCP & HTTP Traffic
5 Nodes
0%
20%
40%
60%
80%
100%
120%
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
1
1
0
1
Offered load intensity
Throughput Effect for a 512k buffer
size
Series1

WLAN
throughput
(bits/sec)
0
10
20
30
40
0 1000 2000 3000 4000
O
f
f
e
r
e
d
L
o
a
d



Wireless LAN.Load (bits/sec) fragmentation
Effect for 512k buffer sizes
Wireless
LAN.Load
(bits/sec).
none
Fragmentation
sizes
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
98 | P a g e
www.ijacsa.thesai.org
fragmentation sizes while throughput can still maintain the
same or achieve a slightly higher value. The reduction of
WLAN load is important for power reduction in wireless
devices. The Fragmentation threshold parameters used for
the simulation run indicate that when the bit error rate.

Figure 10: A Plot Of WLAN Delay Effects For 64k,512k,1024k,&2048k

Figure 11: A plot of WLAN Data Dropped for 512K & 2048K buffer Sizes
During the simulations, the media access delay i n the four
cases above was collected. The media access delay is the sum
of queue and contention delays of data packets received by the
WLAN MAC layer from the higher layer. For each packet, the
delay is recorded when the packet is sent to the physical layer.
The simulation results (Fig.10) indicate that smaller
buffer size values can decrease the average media access delay,
while larger buffer sizes increases average media delay. Also
the average media access delay is increased with increase in
load intensities. This shows that performance of a wireless
LAN is reduced with an increase in load intensity even when
the buffer size is as well increased.
Fig. 11 shows simulated DCF throughput as a function of
offered load for 540 wireless stations. Offered load in this
context depicts the WLAN client nodes. First, we note that for
the 10 and 20 stations, throughput is higher for 2048k buffer
size compared with 512k buffer size. As offered load is
increased, throughput grows linearly until a saturation point at
which throughput ceases to increase; in fact, it decline. When
the number of stations is large, throughput decrease and hence
packet drop is inevitable. The drop in peak and saturation
throughput as a function of the number of wireless stations is
shown in Fig. 11.
At a moderate load of 16-20 wireless nodes, peak
throughput decreases for buffer sizes 512k-2048k. Hence, for
efficient network performance, the number of workstations
should be carefully chosen and the AP device properly selected
for high throughput in all cases.
VI. CONCLUSION
This research has presented a conceptual GRID WLAN
hotspot model with performance analysis which can be further
expanded to address the security challenges of the Wimax base
station. The work investigates the effects of GRID WLAN APs
buffer size distributions with respect to load intensities as well
as fragmentation capacities. The analysis and measurement
results from OPNET simulation shows that a careful selection
of buffer sizes, fragmentation thresholds in GRID WLAN
hotspot model can help to optimize network performance.
Although some of these results are characteristic of the
designed network, I can conclude that the GRID WLAN
hotspot network performance is mainly based on the
simplification of the traffic in addition to mobility and the
buffer sizes.
REFERENCES
[1] Jiaqing Song and Ljiljana Trajkovic: Enhancements and performance
evaluation of wireless local area networks.
[2] B. P. Crow, I.Widjaja, J. G. Kim, and P. T. Sakai. IEEE 802.11 wireless
local area networks.IEEE communications Magazine, September 1997.
[3] S. Khurana, A. Kahol, S. K. S. Gupta and P. K. Srimani :Performance
evaluation of distributed co-ordination function for IEEE 802.11
Wireless LAN protocol in presence of mobile and hidden terminals
(Unpublished).
[4] LAN MAN Standards Committee of the IEE Computer Society, Part
11: wireless LAN medium access control (MAC) and physical layer
(PHY) specifications, ANSI/ IEEE Standard 802.11, 1999 Edition.
[5] L. bononi, m. conti, and L. donatiello, design and performance
evaluation of a distributed contention control (dcc) mechanism for ieee
802.11 wireless local area networks, in proceedings of first Acm
international workshop.
[6] C.-H. Ng, J. Chow, and Lj. Trajkovic, Performance evaluation of the
TCP over WLAN 802.11 with the snoop performance enhancing proxy,
OPNETWORK 2002, Washington, DC, Aug. 2002.
[7] .S. A. Bawazir, S. H. Al-Sharaeh, Performance of infrastructure mode
wireless LAN access network based on OPNET Simulator Department
of Computer Science, Normal, AL 35762,USA
[8] P. Jarmo, OPNET - Network Simulator, seminar presented at VTT
Technical Research Centre of Finland, March, 2006
[9] J.Singh, Quality Of service in wireless Lan Using OPNET Modeller,
Available at http://dspace.thapar.edu.
[10] M. S. Gast, 802.11 Wireless networks: the definitive guides, OReilly,
2002.
0
5E-09
1E-08
1.5E-08
2E-08
0 0.1 0.2 0.3 0.4
D
e
l
a
y

x

1
0
0
0
0
0

Offered Load Time (Sec.)
x 10000
WLAN DELAY EFFECTS
*2048k wlan_buffer
*1024K Wlan_buffer
*512K Wlan_buffer
*64K Wlan_buffer
TCP &
HTTP Traffic
-1000
0
1000
2000
3000
4000
5000
6000
7000
0 1000 2000 3000 4000
T
h
r
o
u
g
h
p
u
t

Buffer sizes
WLAN Data Dropped for Buffer sizes
512k & 2048k
Wlan_buffer_
512K: Wireless
LAN.Data
Dropped
(bits/sec).non
e
TCP &
HTTP
Traffic
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
99 | P a g e
www.ijacsa.thesai.org
[11] A. Banchs, Analysis of the delay distribution in 802.11 DCF: A step
towards delay guarantees in WLANs, LNCS, vol. 3266, Sept. 2004, pp.
6473.
[12] IEEE 802.11e/D13.0, Part 11, Wireless LAN medium access control
(MAC) and physical layer (PHY) specifications: medium access Control
(MAC) enhancements for Quality of Service (QoS), draft supp. to IEEE
802.11 std., Jan. 2005.
[13] R. Litjens et al., Performance analysis of wireless LANs: an integrated
packet/flow level approach, Proc. 18th Intl. Teletraffic Cong., Sept.
2003, pp. 93140.
[14] B. Liu, Z. Liu, and D. Towsley. On the capacity of hybrid wireless
networks. In Proc. IEEE INFOCOM 03, pages 15431552, 2003.
[15] T. Hansen, P. Yalamanchili and H-W. Braun, Wireless measurement
and analysis on HPWREN, Proceedings of Passive and Active
Measurement Workshop, Fort Collins, Co, pp. 222-229, March 2002.
[16] M. H.Manshaei and T. T. INRIA, Simulation-based performance
analysis of 802.11a Wireless LAN. Route des Lucioles, BP-93, 06902
Sophia-Antipolis Cedex, France ,2004.
[17] J. Song and L. Trajkovic, Enhancements and performance evaluation
of wireless LAN. Communication Networks Laboratory Simon Fraser
University, Burnaby, BC, Canada.
[18] Kaur1, S. Vijay, S.C. Gupta, Performance analysis and enhancement of
IEEE 802.11 wireless LAN. Global Journal of Computer Science and
Technology Vol. 9 Issue 5 (Ver 2.0),130 ,January 2010 .
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
100 | P a g e
www.ijacsa.thesai.org
An enhanced Scheme for Reducing Vertical handover
latency

Mohammad Faisal, Muhammad Nawaz Khan

Department of Computing, Shaheed Zulfikar Ali Bhutto Institute of Science & Technology (SZABIST),
Islamabad, Pakistan.


Abstract- Authentication in vertical Hand over is a demanding
research problem. Countless methods are commenced but all of
them have insufficiencies in term of latency and packet loss.
Standard handover schemes (MIPv4, MIPv6, FMIPv6, and
HMIPv6) also practice these shortages when a quick handover is
desirable in several genuine circumstances like MANETs,
VANETs etc. This paper will evaluate the literature of the work
done in past and present for undertaking the authentication
concerns in vertical handover and will put down a basis for
building the latency and packet loss more effective in such a huge
shared situation that can produce to an extremely bulky level.
This effort will mostly focus on the existing tendency in vertical
handover mostly with the authentication, latency and packet loss
issues.
Keywords: Vertical hand over, Authentication, FMIPv6, HMIPv6,
reactive, proactive, latency and packet loss.
I. INTRODUCTION
Future generation heterogeneous wireless networks
(FGHWNs) are facing with service continuity challenge for
which vertical handoff is indispensible. So its assurance by
means of rapid and efficient vertical handover that recognize
service connection and seamless mobility maintaining the
security and QoS is demand of the day [4]. Administered by
diverse operators like WiMax,WiFi ,UMTS,GSM etc each in
the run to achieve the highest quality and data rates,
dynamically to choose for the always best connected network.
Same technology networks switching are called horizontal
while different technology networks switching are called
vertical handover.
As shown:

Figure 1. Vertical and Horizontal Handoff
[K.Savitha et al:1]
Four stages are accomplished during handover mechanism
Handoff Initiation stage: Signal strength, link quality
etc initiated the handover procedure.
System discovery stage: Discovering the neighbor
networks, sharing Quality of Service (QoS)
information offered by these networks.
Handover Decision stage: Comparing QoS of all
available networks leads the user toward best network
selection.
Handoff Execution stage: Relinquishing old and
establishing new connection and security services is
done in this phase [1].
FGHWN are the result of overlapping of diverse wireless
networks. The heterogeneity of FGHWN needs the realization
of a vertical authentication method that trim down the handover
delay whereas safeguarding the security perspective of the
enduring communication. To uphold the continuing
connections still after the home network is no longer
accessible, vertical handover methods must be put into
practice. Today, subscribers turn into additional challenges in
conditions of omnipresent services availability and security.
Furthermore, the installed fourth generation (4G) networks are
illustrated by several wireless access networks expertise. To
assure availability at the same time as preserving the security
level, operators and service providers be supposed to
harmonize and facilitate a mobile subscriber steadily using the
services he is subscribed to regardless of his position and the
access network expertise that is obtainable in his region, for
which novel inter-networks roaming and vertical handover
means ought to be cleared. To provide effective mobility with
security all generation technologies (2G, 3G, and 4G) should
be interoperable to facilitating the subscribers with ubiquitous
wireless communication with high data rates. In case of loose
coupling between WiMAX and WiFi networks operators have
to describe supplementary mechanisms to implement network
intelligence be capable to execute vertical handovers, as WiFi
networks are incapable of management. In case of tight
coupling, WiMax apparatus will be utilized to tackle
subscribers mobility and seamless connection shift, as they are
capable of intelligent management [5].
II. LITERATURE REVIEW
In this paper the authors K.Savitha et al illustrated three
different designs namely centralized, distributed and trusted
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
101 | P a g e
www.ijacsa.thesai.org
distributed for vertical handoff decision, by means of three
parameters; Throughput, Decision Processing and End-End
delay [1]. The strength of the paper is that, it validates that
among the each outcome of all the three designs the trusted
distributed for vertical handoff decision having best
performance among all. The limitation of the paper is that, a lot
of delay time both in decision processing and End-End
transmission. Secondly, there are many inappropriate
judgments attempted by the mobile nodes to decide their target
network due to data flooding which leads to network
congestion.
In this paper the authors Bao Guo et al compared and
contrast two schemes (Encrypt-and-MAC, Encrypt- then-
MAC) to achieve the confidentiality and integrity for the Short
Message Services Networks [2]. The Encrypt-and-MAC
scheme using CTR and CMAC modes while Encrypt-then-
MAC scheme using EAX mode, both characterize by AES.
Strength of the paper is that, output of each scheme in terms of
operation time is almost similar but the Encrypt-then-MAC
scheme offer more assurance, because appending MAC of the
encrypted plaintext. Secondly Encrypt-then-MAC scheme also
demonstrating best performance during online processings.
The limitation of paper is that, implementation of transmitter
flow (SE/ES) will create the I/O obstruction if input data is
increased. Secondly, Python 2.6 executables codes
functionality is platform restricted.
In this work the authors Jakub Szefer et al suggested a
flexible hardware mechanism (FPGA) that facilitates in time
operations for encryption and hashing algorithms [3]. Strength
of the paper is that, the Design is compatible with any portable
devices and sensor nodes. Secondly, its implementation doesnt
compromise its performance at the cost of its area and time [3].
Thirdly, the design can be used for other security services not
just ciphering and hashing. Limitations of the paper are that,
the mechanism of the FPGA policy is fixedly predefined.
Secondly mechanism is adjustable with a limited and fixed
quantity of Parameters. Thirdly, its efficiency is variable for
cipher-to-cipher or hash-to-hash; maximum for whirlpool and
block ciphers.
In this paper the authors wafaa bou diab et al proposed a
new Seamless vertical handover solution for real-time data
transfer with fast authentication. while analyzing the results
with conventional schemes in terms of signaling cost (handover
and authentication) and packet loss [4] It proved better results
than its predecessor (IMS, MIPv6, FMIPv6) although quality
and security services both maintained. Strength of the paper is
that, it decreased the number of authentication messages,
because both handover and authentication amalgamated in a
single message, which reduced handoff latency as well.
Resultantly the chances of packet loss also diminished. The
limitation of the paper is that, this model is very sensitive to
packet loss. Any loss of packet will lead to miss-
synchronization, resultantly either to reinstate a new security
association or roll back to the initial state.
In this paper the authors Neila KRICHENE et al discussed
a vertical authentication method among the GSM, UMTS,
WiFi and WiMAX technologies, regardless of any previous
subscription to the visited network [5]. Strength of the paper is
that, the authors proposed a global authentication protocol
authorizing vertical handover between 4G architecture
networks and proved the strength of the protocol against man-
in-the-middle and denial of service attacks during vertical
handover. The limitation of the paper is that, this protocol is
only compatible with mesh topology 4G networks and does not
bother for Quality of Service.
In this paper the authors Ahmed H. Zahran et al described
a broker-based design for integrated heterogeneous networks to
enhance the vertical handoff management, utilizing novel
resource query method for Media Independent Handover [6].
The strength of the paper is that, it focuses on the decline in
signaling load, MT configuration time, power consumption,
user authentication delay, VHO delay, and the probability of
VHO interruption, although guarantying seamless transitions as
well. The limitation of the paper is that, almost all of its work is
least cited which degrade its authenticity.
In this paper the authors Ali Al Shidhani et al put forward
two re-authentication protocols (FUAR,LFR) comparing them
to existing protocols(EAP-AKA, UMTS-AKA) in the 3G
Home Networks [7]. The strength of the paper is that, It
considerably spawning less signaling traffic and enduring less
delay, accomplishing secured key management and mutual
authentication, demanding no adjustments to the interworking
architecture. The limitations of the paper is that,FAURs
security and performance depends upon the lifetime of security
keys (TK, nCK, nIK).while security keys neither can be shared
nor it can be reused, one time the session expired. The paper
only focused on 3G networks
In this paper the authors Hoyeon Lee et al, have done
efforts to enhance SIP, for vertical handover design (WiFi-to-
UMTS) to reduce the IMS delay, comparing the results with
conventional SIP [8].The strength of the paper is that, it
enabling make-before-break vertical handover for IMS,
successfully defined and operated two new headers in SIP, as
sustaining for delay-sensitive real time applications, presenting
compatibility with traditional. The limitation of the paper is
that, it does not bother about the Layer 2 and 3 handover
latency, secondly neglecting the forecasted troubles due to
increase in header size.
In this paper the authors Jaeho Jo et al have focused on
vertical handover via layer 2 and layer 3 signaling messages
connecting mobile WiMAX and 3G networks [9].
The strength of the paper is that, it sort out L2 and L3
signaling messages as merging them resultantly limiting
handover latency and UDP packet loss ratio while enhancing
the TCP throughput. The limitation of the paper is that, it
implemented the idea of predictive mode fast handover in
which chances of failure are always open.
In this paper the authors chan-Kyu Han et al, assessed the
signaling load, authentication procedures and compatibility of
vertical handover in the EPS architecture networks [10]. The
strength of the paper is that, it introduces the EPS network
release 8 concentrating on security, secondly allowing random
processes in authentication arrival, thirdly discovering new
authentication generation processes other than routine, fourthly
enhancing authentication signals as required. Limitations of the
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
102 | P a g e
www.ijacsa.thesai.org
paper are that, mandatory parameters like mobility
organization, security strategies, and a variety of haphazard
arrival processes are ignored in the current model.
In this paper the authors Liming Hou et al, introduced a
pre-authentication design, based on the EAPTLS protocol,
between WiFi and WiMAX hybrid networks, comprises of two
stages; pre-authentication and re-authentication [11]. Strength
of the paper is that, authentication delay is considerably
reduced, while introducing pre-authentication stage. The design
can be used for real time services as well. Limitation of the
paper,EAPTLS is based on a public key infrastructure for its
authentication, which reduces its portability.
In this paper the author Mario Marchese, integrate the
interconnection and (Delay Tolerant Networks) DTN gateways
architectures for wireless ubiquitous networks [12]. The
strength of the paper is that, it shared successfully the
functionalities of QoS mapping, and resource control from
interconnection Gateways while extended delays and
momentary link unavailability from DTN Gateways. The
limitation of the paper is that, essential functionalities like well-
organized mobility operates on broad territory coverage and
guarantee of end-to-end data deliverance for elongated delay
corridor and momentary link distraction are ignored.
In this paper the author Shih-Jung Wu specify (Host
Identity Protocol) HIP-based vertical handover scheme via
integrated architecture for seamless mobility of diverse
wireless networks [13]. The strength of the paper is that, it used
effectively HIP to tackle the problems of mobility and multi-
homing while Diameter Protocol for registered users
authentication. Limitations of the paper is that, as the local
scope identifier (LSI) is a 32-bit identifier of host identity
which is incompatible for IPv6 so collision probability is
enviable, and its scope will be also controlled. Lastly results are
not simulated or proved.
In this paper the authors Gabriele Tamea et al have
provided an algorithm which makes its vertical handover
decision on the basis of probability to avoid ping-pong effect
[14]. Strength of the paper is that, the suggested algorithm
evaluating the (WDP) wrong decision probability on the basis
of: improving the collective goodput, and dropping of
redundant and needless vertical handovers (PPE), was
empowering the Mobile Terminal (MT) for the commencement
and restriction of VHO that is why it is distinct as a Mobile
Terminal-Controlled Handover scheme. Limitations of
the paper is that, it didnt bother other estimations of good put ,
and analysis of its correlation at different intervals.
In this paper the authors Pedro J. Fernandez Ruiz et al,
portrayed a testbed setup installed over inter and intra
technology vertical handover (WiFi,WiMAX and UMTS) to
testify the secure mobility on vehicular networks [15]. Strength
of the paper is that, it successfully experienced the
authentication process on vehicular networks using IKEv2 and
EAP3 for dynamic IP allocation and authentication
respectively, exclusive of losing connectivity. Limitation of the
paper is that, due to vehicular network there must be
overlapping and no overlapping zones for which the selection
and authentication procedures are not discussed.
III. CRITICAL ANALYSIS
TABLE 1.
Author Working Problems Solution
K.Savitha et al
Validates that the trusted distributed for vertical
handoff decision having best performance
among all.
Delay time both in decision processing,
End-End transmission, many
inappropriate judgments which lead to
network congestion.
By replacing the reactive routing protocol, Ad
hoc On Demand Distance Vector (AODV)
with a proactive routing protocol like, Highly
Dynamic Destination-Sequenced Distance
Vector routing protocol (DSDV), to reduce
the delay time.
Bao Guo et al
Encrypt-then-MAC scheme offer more
assurance, because appending MAC of the
encrypted plaintext.
Implementation of transmitter flow
(SE/ES) will create the I/O obstruction if
input data is increased. Python 2.6
executables codes functionality is
platform restricted.
If we implement MAC-then-encrypt scheme,
will enhance security services. And if python
2.6 is replaced by PERL or ruby
programming language then the platform
independence issue will also be resolved
Jakub Szefer et al
The Design is compatible with any portable
devices and sensor nodes. Its implementation
doesnt compromise its performance at the cost
of its area and time. The design can be used for
other security services not just ciphering and
hashing.
The mechanism of the FPGA policy is
fixedly predefined. Mechanism is
adjustable with a limited and fixed
quantity of Parameters. Its efficiency is
variable for cipher-to-cipher or hash-to-
hash; maximum for whirlpool and block
ciphers.
We can design FPGA user defined while
assigning unlimited parameters according to
the need and situation dynamically.
wafaa bou diab et al
It decreased the number of authentication
messages, because both handover and
authentication amalgamated in a single message,
which reduced handoff latency as well.
Resultantly the chances of packet loss also
diminished.
This model is very sensitive to packet
loss. Any loss of packet will lead to miss-
synchronization, resultantly either to
reinstate a new security association or roll
back to the initial state.
A mechanism can be introduced so that, on
each and every mobile node SCID (Session
Context ID) routing table should be fully
updated with full header packets rather only
destination IP addresses.
Ali Al Shidhani et al
It considerably spawning less signaling traffic
and enduring less delay, accomplishing secured
key management and mutual authentication,
FAURs security and performance
depends upon the lifetime of security keys
(TK,nCK,,nIK).while security keys
Consequence of additional security keys for
each session is required to be investigated.
Proper confirmation of the security and
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
103 | P a g e
www.ijacsa.thesai.org
demanding no adjustments to the interworking
architecture.
neither can be shared nor it can be reused,
one time the session expired. The paper
only focused on 3G networks
performance of FUAR will be carried out
with constraint of life time.
Hoyeon Lee et al
It enabling make-before-break vertical handover
for IMS, successfully defined and operated two
new headers slots in SIP.
It does not bother about the Layer 2 and 3
handover latency, secondly neglecting the
forecasted troubles due to increase in
header size.
Large bandwidth allocation to solve the
header size problem.
Jaeho Jo et al
It sort out L2 and L3 signaling messages as
merging them resultantly limiting handover
latency and UDP packet loss ratio while
enhancing the TCP throughput.
It implemented the idea of predictive
mode fast handover in which chances of
failure are always open.

predictive mode can be replaced with reactive
mode fast handover


chan-Kyu Han et al
Allowing random processes in authentication
arrival, discovering new authentication
generation processes other than routine,
enhancing authentication signals as required.
Parameters like mobility organization,
security strategies, and a variety of
haphazard arrival processes are ignored in
the current model.
Several features like haphazard walk mobility
model, associations with every authentication
activation and different random process
generation are to be examined for the
compatibility with the real world scenarios.
Liming Hou et al
Authentication delay is considerably reduced,
while introducing pre-authentication stage. The
design can be used for real time services as well.
EAPTLS is based on a public key
infrastructure for its authentication, which
reduces its portability.
we can replace the protocol by any other of
EAP family member like EAP-MD5, EAP-
OTP, EAP-GTC, EAPTLS and EAP-SIM etc.
Mario Marchese
It shared successfully the functionalities of QoS
mapping, and resource control from
interconnection Gateways while extended delays
and momentary link unavailability from DTN
Gateways.
Essential functionalities like well-
organized mobility operates on broad
territory coverage and guarantee of end-
to-end data deliverance for elongated
delay corridor and momentary link
distraction are ignored.
A novel illumination relating to architectures
and protocols is obligatory
ih-Jung Wu
It used effectively HIP to tackle the problems of
mobility and multi-homing while Diameter
Protocol for registered users authentication.

As the local scope identifier (LSI) is a 32-
bit identifier of host identity which is
incompatible for IPv6 so collision
probability is enviable, and its scope will
be also controlled. Lastly results are not
simulated or proved.
Its results should be simulated so that to
assess its presentation and should also be
compared with its predecessor work.

Gabriele Tamea et
althe
Improving the collective good put, and dropping
of redundant and needless vertical handovers
(PPE),empowering the Mobile Terminal (MT)
for the commencement and restriction of VHO
It didnt bother other estimations of good
put, and analysis of its correlation at
different intervals.

To include the other goodput parameters and
enhance correlation among them.

Pedro J. Fernandez
Ruiz et al
it successfully experienced the authentication
process on vehicular networks using IKEv2 and
EAP3 for dynamic IP allocation and
authentication respectively, exclusive of losing
connectivity.
As it is vehicular network there must be
overlapping and no overlapping zones, for
which the selection and Authentication
procedures are not discussed.
To set pre-established security and mobility
policy that will allow the user to opt the
appropriate interface in each and every
situation.
IV. PROPOSED SOLUTION
The above observations reveal that many schemes were
introduced but all of them have deficiencies. Hand over process
takes time in address acquisition and on time calculations
which increases latency and hence leads to packet loss. For fast
and smooth hand over between homogeneous and
heterogeneous networks, a modified and fast hand over
mechanism always needed. Hand over is more critical in some
conditions like in ad hoc networks such as MANETs and
VANETs. In ad hoc networks the hand off need very few time
for changing point of attachment to the access point and with
routers as mobile nodes move very fast. The horizontal
handover between different network such as WiFi, WIMax,
UMTS also leads latency and packet loss. Latency created in
the hand over process because the difference in the band width.
When a node leaves from one network and inters into the
premises of another network such as from WiFi to WiMax, the
difference between band widths leads latency. The queue
becomes full at the bridge of the network and waiting packets
feel delay in term of latency and as a result all this lead to
packet loss. In horizontal hand over on time calculations also
takes timing to change from point to point.
The Standard IEEE 802.11 Handover process of consist
into following six phases: triggering, discovery, authentication,
association, IP address acquisition, and home agent (HA)
registration [16]. The first four phases related to the data link
layer and called layer two (L2) while the last two phases done
it the network layer and called layer three (L3) handover [17].
In the whole process there is on time calculation and awake the
moment of the mobile node. The moment of the mobile node
from one network premises to another continuously monitor
from its Received Signal Strength Indicator (RSSI) value.
When RSSI value decrease in area the mobile node start
scanning for other network having better RSSI value. IEEE
802.11 had over two types of scan modes are define: active
scan and passive scan. In active scanning, the MN transmits a
probe request massage on a channel and then waits for a while.
If no response from other side, mobile node continues to sends
a probe massage on the next channel and wait for response
from other access router. While in passive scanning, the mobile
node sequentially listens to beacons on different channels
instead of transmitting massages for request. When mobile
node listen all the channel, it choose a channel with better RSSI
value and start massages to shift from one access router to
another.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
104 | P a g e
www.ijacsa.thesai.org
In our proposed scheme passive scanning mode, the mobile
node start hand over procedure rather network. Nearly all
parameters are set prior starting the connection establishment
of new connection with another access router. While some very
necessary calculation still performed at time of changing access
routers.
In the proposed scheme total focus is on proactive
mechanism rather than reactive all time. Proactive means much
of the work been done by the system before actually hand over
been starting. First, when mobile node experience weak RSSI
value, it starts searching for better RSSI network value. The
mobile node starts discovery and section for new access
routers. Once the mobile is performed authentication to
entering at premises of the new network. Then no need for re-
authentication, when changing from one access point to other
(horizontal handover). At the association process the mobile
node define itself prior to association and almost all massages
shared between new router and mobile node.
Therefore for a mobile node tack very little time to
associate with access router. After association the mobile node
trying to get new IP address. Two variations either Router
Advertisement messages are periodically broadcasted by new
router or the mobile node sends Router Solicitation messages to
new router to obtain the Router Advertisement messages. Care
of Address (CoA) has been assign to mobile node for time
being, duplication address detection (DAD) process ensure
unique addresses in the same network premises. When a
mobile node obtains its CoA in new network then mobile node
informs its Home Agent about the new CoA by sending a
binding update message to home agent. The home agent also
sends a binding acknowledgement massages to the mobile node
to complete the process [18].
The standard scheme shows that latency turn outs because
of the on board calculation. Nearly all schemes needs on time
calculations for completing the hand over procedure. But these
calculations can be decrease by dividing the whole process into
two parameters. Some necessary parameters are still calculated
on time but most of the parameters should set before the actual
process being start.
The proactive mechanism reduces the delay in term of
reducing latency. Reactive mechanisms are applied when some
parameters need some on line calculating values otherwise the
proactive mechanism. Proactive mechanisms reduce the hand
over latency by applying pre-define parameters values.
Proactive parameters including pre-define calculations by the
mobile node for smooth roaming in the same network with
different access points or roaming between different routers of
the different networks, pre-define pool of address for mobile
nodes which decrease on time calculation for address
acquisition and duplication address detection.
V. CONCLUSION AND FUTURE WORK
While evaluating all the above mentioned work it can be
concluded that almost all of the time reactive protocols are used
when a node shift from one entrance point to a new one, while
executing estimations in terms of handover and authentication.
As handover massages are exchanged, the actual data packets
experiences extensive average delay, which amplify latency
and lastly leads to packet loss in few cases. In our proposed
scheme, a few parameters will be pre-defined, which will assist
in calculation at hand over process. The proposed scheme will
be helpful as it is based on preceding standard approaches with
improvements. The proposed elucidation will be a hybrid
approach of reactive and proactive.
AKNOWLEDGMENT
We are very thankful to Almighty Allah; whose grace
and blessed mercy enabled us to complete this work with
full devotion and legitimacy. We are grateful to Mr.
Mohammad Ibrahim, Lecturer in Computer Science, Virtual
University of Pakistan, for their invaluable support and
guidance throughout this research work. We also want to
thank our friends and family for their encouragement;
without whose support we could not have lived through this
dream of ours.
REFERENCES
[1] K.Savitha and Dr.C.Chandrasekar VERTICAL HANDOFF
DECISION SCHEMES IN HETEROGENEOUS WIRELESS
NETWORK USING MADM, JGRCS 1(5), December 2010, 16-20.
[2] Bao Guo and William Emmanuel Yu Comparison between Encrypt-
and-MAC Composite (CMAC CTR) and Encrypt-then-MAC Composite
(AES EAX) Modes of Operation in Cryptography Systems for Use in
SMS-based Secure Transmission. Proceedings of the international multi
conference of Engineers and Computer Scientists 2011 Vol I, IMECS
2011 March 16-18, 2011, Hong Kong.
[3] Jakub Szefer, Yu-Yuan Chen and Ruby B. Lee General-Purpose FPGA
Platforms for Efficient Encryption and Hashing
[4] Wafaa Bou Diab and Samir TohmeSeamless Handover and Security
Solution for Real-Time Services 2009 11th IEEE International
Symposium on Multimedia 978-0-7695-3890-7/09 $26.00 2009 IEEE
[5] Neila KRICHENE and Noureddine BOUDRIGA Securing roaming and
vertical handover in fourth generation networks Third International
Conference on Network and System Security 978-0-7695-3838-9/09
$26.00 2009 IEEE
[6] Ahmed H. Zahran and Cormac J. Sreenan Extended Handover Keying
and Modified IEEE 802.21 Resource Query Approach for Improving
Vertical Handoff Performance 978-1-4244-8704-2/11/$26.00 2011
IEEE
[7] Ali Al Shidhani and Victor C. M. Leung Reducing Re-authentication
Delays during UMTS-WLAN Vertical Handovers 978-1-4244-2644-
7/08/$25.00 2008 IEEE
[8] Hoyeon Lee, Bongkyo Moon, and A. H. Aghvami Enhanced SIP for
Reducing IMS Delay under WiFi-to-UMTS Handover Scenario The
Second International Conference on Next Generation Mobile
Applications, Services, and Technologies 978-0-7695-3333-9 /08
$25.00 2008 IEEE

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
105 | P a g e
www.ijacsa.thesai.org
[9] Jaeho Jo and Jinsung Cho A Cross-layer Vertical Handover between
Mobile WiMAX and 3G Networks978-1-4244-2202-9/08/$25.00
2008 IEEE
[10] Chan-Kyu Han, Hyoung-Kee Choi,Jung Woo Baek, Ho Woo Lee
Evaluation of Authentication Signaling Loads in 3GPP LTE/SAE
Networks2009 IEEE 34th Conference on Local Computer Networks
(LCN 2009) Zrich, Switzerland; 20-23 October 2009 978-1-4244-4487-
8/09/$25.00 2009 IEEE [11] LimingHou and Kai X.
[11] Miao A Pre-authentication Architecture in WiFi & WiMAX Integrated
System.
[12] Mario Marchese, Wireless Pervasive Networks for Safety Operations
and Secure Transportations IEEE 2010 International Symposium on
Wireless Pervasive Computing (ISWPC)
[13] Shih-Jung Wu A New Integrated Mobile Architecture for
Heterogeneous Wireless Networks 2010 Sixth International Conference
on Intelligent Information Hiding and Multimedia Signal Processing
2010 IEEE.
[14] Gabriele Tamea, Anna Maria Vegni, Tiziano Inzerilli, Roberto Cusani
A Probability based Vertical Handover Approach to Prevent Ping-Pong
Effect ISWCS 2009.
[15] Pedro J. Fernandez Ruiz Cristian A. Nieto Guerra Antonio F. Gomez
Skarmeta Deployment of a Secure Wireless Infrastructure oriented to
Vehicular Networks 2010 24th IEEE International Conference on
Advanced Information Networking and Applications
[16] Daehan Kwak, Jeonghoon Mo, Moonsoo Kang, Investigation of
Handoffs for IEEE 802.11 Networks in Vehicular Environment.
[17] Yao-Tung Chang, Jen-Wen Ding, Jen-Wen Ding, Ing-Yi Chen. A
Survey of Handoff Schemes for Vehicular Ad-Hoc Networks.
[18] R. Koodli, Ed. Mobile IPv6 Fast Handovers, Request for Comments:
5268 Starent Networks Obsoletes: 4068 June 2008.
AUTHORS PROFILE
Mohammad Faisal received his B.S. degree in Computer
Science (Hons) from University of Malakand at Chakdara,
Dir lower, KPK, Pakistan. He is currently pursuing his MS
in Information Security management system at Shaheed
Zulfiqar Ali Bhutto Institute of Science & Technology
(ZABIST) Islamabad Pakistan. Since November 2006, he
is working as Network Engineer in Pakistan Revenue
Automation Limited (PRAL), Motorway Department.
His research areas focuses on Handoff, forensics, cryptography and networks
(SENSOR, WIRELSS, MANETS) security. He intended to precede his studies
(PhD) in any of the above mentioned fields.

Muhammad Nawaz Khan is lecturer in Computer
Science in Govt. College of Management Science. In
2008, he received Silver Medal in B.S. (Hons) degree in
Computer Science from University of Malakand, K.P.K.
Pakistan. He partially completed MS in Computer
Communication Security at School of Electrical
Engineering & Computer Science (SEECS), National
University of Science & Technology (NUST) Islamabad,
Pakistan. In 2010, he worked as a Research Assistant in a project on
Distributed Computing supported by Higher Education Commission of
Pakistan. Currently he is working as Research Assistant at Shaheed Zulfikar
Ali Bhutto Institute of Science & Technology Islamabad. His research is
focused on Computer Information Security especially Computer
Communication Security. He has also showed keen interest in Ad-hoc
networks (MANETs, VANETs), wireless communications security and
security related issues in distributed computing. He intended to proceed his
studies (PhD) in any of the above mentioned fields.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
106 | P a g e
www.ijacsa.thesai.org
A Feasible Rural Education System

Lincy Meera Mathews
Assistant Prof, Department of ISE,MSRIT
M S R Nagar, Bangalore,
Karnataka, India
Dr Bandaru Rama Krishna Rao
Senior Professor, School of Information Technology and
Engineering, Vellore Institute of Technology
University, Vellore, Tamil Nadu


Abstract The education system in rural and semi-rural areas of
developing and underdeveloped countries are facing many
challenges. The limited accessibility and challenges to the
education are attributed mainly to political, economic and social
issues of these underdeveloped countries. We propose a Feasible
Rural Education System (FRES) based on Ontology and
supported by Cloud to enhance the accessibility to education in
rural areas. The system has been proposed incorporating the
FOSS approach.
Keywords- FRES; Education; FOSS; Semantic Web; Ontology;
Natural language processing; Cloud computing.
I. INTRODUCTION
In developing countries like India, the rural and village
schools form 87% of the total schools, of which 90% are run
by the government with financial aid and rest of the schools
are unaided [1]. These schools offer education in four phases:
Primary, Upper-primary, Secondary and the Higher
Secondary.
Though the literacy rates in developing countries have
been steadily increasing in higher education system (which
includes educational and vocational training), the same is not
noticed in the Primary and Upper-primary education in rural
and semi-urban areas. Various initiatives had been taken up by
the Government of India to support Education in rural areas by
bringing in various technologies and through the usage of
internet. The role of Information and Communications
Technologies in education can be dispersed into manners [6].
Alternative instructional delivery systems such as
radio, educational TV and audio visual
communication.
Computers and computer based system for
instructional delivery and management; use of
multimedia and internet/web based education.
As an initiative for the former mode of communication,
EDUSAT was launched in September 2004, at a cost of USD
20 million. Indias first Education satellite was dedicated.
However it has failed to provide any speak able impact in rural
areas. The problems attributed to this were previously
recorded lectures, non-interactive nature of sessions, not
formed for as per students learning capability.
As for the latter, the government has initiated an Internet
Based Continuing Education. The components are comprised
of online learning materials, online Academic counselling,
online assignments and projects and online collaboration. To
help back all the state government ICT education initiatives, in
May 20, 2006, the government of India, Ministry of HRD,
Department of Secondary and Higher Education had issued an
order for the implementation of Broadband Connectivity in all
secondary schools. To accelerate the progress of education in
rural areas, we take the help of the free and open source
software (FOSS) movement. The FOSS has originally begun
as an intellectual movement for development, modification,
ownership and redistribution of software managed by a
community of programmers. We suggest that the principles of
FOSS namely transparency, collaboration, openness and co-
ownership can be introduced in our education system in
providing accurate, accessible and free education resources.
II. LACUANES IN RURAL EDUCATION SYSTEM
A typical rural education system contains three main
components namely: Teacher, Student and Infrastructure [2].
A. Student
In rural scenario, student speaks his native language more
fluently, demanding communication for learning in local
language. A student in rural area faces more challenges in
daily life as compared to his counterpart in urban areas, to
name few (i) lack of commutation to schools situated far off
from villages, (ii) playing in streets, want of setup area meant
for play (iii) studying under street lamps, want of electricity at
home, (iv) lack of good curriculum or educational material, (v)
lack of full-pledged laboratory or library facilities etc. In
essence, priority for rural student is to survive for the day
rather than look and work for a better future. The governments
of developing nations aim to bring quality education to these
rural students so that they could compete with their urban
counterpart. However, this becomes difficult, since, the urban
schools have access to top-class teachers, laboratory and
library facilities. Therefore, the needs to be addressed for rural
students are as follows:
An interesting and challenging representation of free
education material, so that attention of the student
holds.
Communication in learners language.
Limited dependence on teachers instruction and on
non-availability of other modes of learning.
More interaction with students of other schools of the
same curriculum.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
107 | P a g e
www.ijacsa.thesai.org
Up to date and accurate representation of curriculum
B. Teacher
A teacher plays the most impacting role in the rural
education system. A teacher forms the root of the system on
which students, learning resources and technologies are
dependent. Without the teacher, the student stands to gain a
very limited knowledge. The learning resources and the new
technologies cease to play any role if teachers do not use the
facilities. The learning resources will not be reviewed if the
teacher fails play an active role in the education system. The
more effective the teacher becomes, the better the quality of
learning by students. However, the challenges faced by
teachers of rural and semi urban in terms where technology
can play roles are as follows:
The large gap between rural and urban technology in
facilitating teaching and learning process.
Limited availability of facilities like study materials,
language friendly multimedia.
Very less or non- availability of resources for
accelerated and quality learning.
Huge cost involved in procuring and maintenance of
infrastructure- mainly software and hardware.
Lack of communication between schools handling the
same curriculum, hence shortage of updated
information and mentoring programs.
C. Infrastructure
The basic infrastructure required can be divided into
consumables products and non-consumable products. The
consumable products mainly are stationery items such as
exercise books, pens, pencils etc and the non-consumable
materials are the curriculum, learning aids, visual aids and
technological equipments. The dearth of non-consumable
products can be met by present day technologies. The various
factors that need to be addressed in this context are as follows:
Free availability of resource materials in the form of
text, video or audio classes.
An effective teaching guide to support the teachers
and learners.
Challenge based or game based learning curriculum
that can hold the attention of learners.
An effective educator system that can take care of
mentoring and supporting of rural educators.
Availability of technology (software and hardware)
without the maintenance and cost issues.
A user friendly interface preferably in the local
language of the user
Low cost systems and lab facilities.
Various other factors like political, male to female ratio,
economic reasons exist [3]. However factors where technology
can play a role and help alleviate the hurdles should be
addressed in this current era.
III. FOSS IMPACT IN EDUCATION
The Free and Open source movement was begun by a
community of programmers to remove or decrease restrictions
on collaborative development, distribution, ownership and
redistribution of software [5]. This however got noticed and
the principle is being applied by various researchers and
academic communities in various domains. The values that the
FOSS movement promotes are openness, transparent
collaboration and co-ownership. All these values need to be
adapted into the education scenario.
A. Openness
Openness in FOSS environment is to have open access to
the source code. In education, openness leads to free access to
curriculum and pedagogy. Currently information related to
curriculum and pedagogy is available only to the selected few.
Though governments provide may Open-Courseware
resources such as Indira Gandhi National Open University
(IGNOU), NPTEL (National Programme on Technology
Enhanced Learning ) and the CEC ( Centre for Education and
Communication ) repository in India, they are not entitled for
changes by others to suit their curriculum, and faces the
following challenges:
Information retrieval works on the basis of queries.
The relevance of data is directly proportional to the
completeness of a query. The information retrieved is
dependent on the learners query.
Trust factor of information retrieved is yet a widely
researched area. The user has to decide on the
accuracy and completeness of information.
There is no guidance in the flow of information or in
the method of learning process. The user or reader
must decide on the flow of information to be read as
there is no connection between the information. If the
student begins on a certain topic, he must be guided to
its prerequisite information.
No guidance is available when the student scours
through lengthy material or tutorial. The online tutor is
absent. She/he does not have any interaction with
subject matter expert or any answering system for his
queries & doubts.
The maturity of the user is also of utmost importance.
Her/his understandability of the instruction, level of
language proficiency, and the grasping power of an
individual all plays a role in effectiveness of the
information.
B. Collaborative development
This in terms of software relates to the development of
code by several professionals interested in the tool or software.
They decide to do changes so that it helps in their application.
With education, collaborative development has been
interpreted as a transparent process for formulation of
educational policy, syllabi and content related to the
curriculum or course. The hurdle here lies with the
transparency process. The limitations of applicability of
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
108 | P a g e
www.ijacsa.thesai.org
transparent collaboration in education may be attributed to the
following reasons.
There exist no common policy and decision makers.
There are different methods of learning tools. Some
schools believe in sending work home. Some work on
completing at school premises
There is always a change in curriculum delivery for
secondary education. Textbook variation, selection of
portions and time of completion causes difference in
the delivery of curriculum.
Many universities have created their digital
repositories for the benefit of students. This however
has been a good initiative for the students of the
respective schools. All different kinds of tools are
being used for updating like PPT, text, audio, video or
word. Updating of resources requires understand
ability of the underlying software. It poses restriction
to authors of various domains.
C. Co-ownership
In FOSS, Software can be freely distributed without
licensing fees. In education the main stakeholders are the
students, teachers, parents and the institution. Ownership
relates to the freedom of access and an equal influence over all
intellectual artefacts used in the education system. Here again
the access and empowerment of curriculum lies in the hands of
the select few, who are eligible for the resources and not those
few who are in search of knowledge. Faculties do not have
access to better materials from various guides. They lack
support from domain proficient members and no
knowledgeable tools. To bring co ownership to education
would mean able to modify and reuse the repositories.
Creativity also has been the other important aspect where
the users who have laid the FOSS foundation have encouraged.
In all its forms, creativity is the spirit in FOSS. This also must
be followed in areas of education too. Tools that enable
creativity must be created and used to provide various
challenged based learning. The user plays a very important
role in creativity. Incorporating the Blooms taxonomy is one
of the major hurdles that are faced by curriculum developers.
IV. SEMANTIC WEB BASED EDUCATION SYSTEM
A. Use Case Scenario
Understanding the shortcomings and the difficulties of a
student from a rural background is important aspect to
consider in the teaching learning process. Since the system can
be categorized as an e-learning space, all the characteristics of
classroom learning if not more must be incorporated into the
system. The rural education system has basically three entities:
Domain model, student model and Facilitators model. Three
requirements of the Semantic Web Education System for three
different entities are discussed below.
1) Student requirement:
Student needs a guide as he sits through the learning
system. She needs an interactive system that guides
him through his learning process
Her preference of the learning style :Audio, Video,
Text or Incorporation of all formats
The dynamic flow of learning material that suits his
understanding capability. She does not have to go
through the entire material but that suits his need of
knowledge
Learning material presented depending upon his
cognitive learning style
A good and friendly user interface
Student data to be stored with respect to his interest,
his learning mode etc
High quality information
Student community
2) Domain requirement
Storing of contents metadata or its relevance in the
learning repository
Structure definition: Organization of each content
material or the learning sequence of the learning
material
Format of the resource material like video. Website ,
Audio etc
Reusability of the content
Dynamic updating of the content
Quality assurance
Automatic of display of Content depending on users
input
3) Facilitators requirement:
Uploading of resource material in any format
Updating of content
Questionnaire in relevance with the content
Evaluation procedures
Guiding of students online or offline
A secured and an user friendly interface
B. Ontology Support
Ontology can be used for knowledge representation with
respect to each domain [4]. Domain ontology has its
advantages. Firstly it allows reuse, updating and sharing of
ontology between ontology friendly environments. Secondly it
has its formal structure which makes it easy to extract and
obtain knowledge representations. The courses or resource
material can be divided into declarative or procedural
knowledge. The declarative knowledge represents the factual
and conceptual knowledge while the procedural knowledge
represents action sequences and problem solving procedures.
The objective behind using ontology is to facilitate a
structured knowledge which helps facilitate reasoning and
help provide the best learning session as per learners need.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
109 | P a g e
www.ijacsa.thesai.org
We present here the architecture (Figure 1) of ontology
supported education system. The ontology supported web
based education system has basically three components. They
are the facilitator, learner and the data repository. The data
repository forms the ontology supported data for the student
and the learning documents. The facilitators role is to support
in content design and providing student support. The learners
form the student as well as faculties of rural schools. They are
the actual beneficiaries of the system. The web based system
is then deployed on to a dedicated cloud for education purpose
only. The web based application is in the form of centralized
web-based applications. One instance of an application hosted
in the cloud is cheaper and easier to manage. Upgrading of the
application which is to be done often due to development of
educational content has to be done often.
To define the ontology, OWL-DL (Protg Editor has been
used. The curriculum developed was included for a single
course. The proof of concept will be deployed in Amazon,
where it has its Elastic Compute Cloud (EC2), a web service
that provides cloud-based resizable computing capacity for
application developers.
















Figure 1: Architecture of Rural Education System
1) Domain Model
The courses are being represented in the form of ontology.
The ontology as shown in figure 2 has been organized as
follows. The definition class has been subdivided into mainly
three sub classes. The definition class is being defined by the
source, Content author, last_modified_date and the learning
objectives. The definition class exists for contents updated into
the domain model. The learning objectives compulsorily
follow the objectives as per Washington accord. The source of
content will be updated as a pair value of Source and
contented. The content_ID acts as the sequence no which
denotes the sequence no in which the content is to be
presented.
The format class is linked by its various subclasses by the
relationship of for_a_particular. The individual contents
could be an image, audio, video or text. The relationship
between these various subclasses follows an OR relationship.
The contents can be of more than one format but not
necessarily all formats. The text data can be further
categorized as PPT, .doc or Pdf formats. The PPT might be
supported with audio also.
The structure denotes the actual content and details for
rendering of content. The subclasses are the content, FAQs,
and test. The content is further divided by its purpose and aim.
It can be classified as definition, principals, advantages,
application etc. The content is also rated depending on its
difficulty level. The content can be rated as topics that come
under surface learning and deep learning. The surface learning
can be described as the using low level cognitive skills and
minimum effort to complete the course requirements. The
deep learning involves understanding, engaging in higher level
cognitive skills. The student should be able to think
conceptually about a topic. The content is being identified by
its content _id which will be given as input for the sequence
_no subclass. The sequence _no subclass will be related to the
prerequisite subclass. The prerequisite states that content to be
mastered before the actual content is to be read. The test
subclass relates to the content that is to be used to evaluate the
students. The evaluation content could be objective or
subjective type of questions. The structure class also maintains
a FAQs. The FAQs will be referred when a student poses a
question. The question will be matched against the FAQ pool.
OWL-DL (Protg Editor) has been used for defining the
ontology .The document ontology is being represented as an
Onto Graph. Figure 3 states the object properties between
various classes. The relationships indicate the functional
dependence between various classes. All classes have direct
dependence to its subclasses. However certain classes have
dependencies to other disjoint classes. One of the functional
dependence indicated is the prerequisite. This states that
content will have a prerequisite. The inverse relationship is
built automatically. In the above example, a prerequisite exists
for all content.
2) Student Ontology
The lack of physical educator must be compensated by the
system. Said in other words, the system must understand the
need of the students and present the course that suits his
learning style. The student background information is stored to
help navigate and select the right learning style for the student.
The profile includes his pre schooling, his language
competency, his reading skills, his attention span etc. The
preference of his learning style is also recorded for the benefit
of the student. The different learning style can be categorized
into visual learners, auditory learners, reading /writing learners,
and kinaesthetic learners.
The cognitive learning style defines the cognitive
characteristics that the student possesses. The Cognitive
characteristics can be categorized as brainstorming followed
by session or vice versa. It is believed that certain students
need facts as to why certain content must be mastered. The
evaluation procedure is recorded in the subclass in evaluation
metric.
The metric store scores as against to each content for each
student. As they complete the test in accordance to the content
referred, the metric is updated. The description of the
individual can be specified by using the individual description
view.
D
a
t
a

R
e
p
o
s
i
t
o
r
y

Web Based Semantic Education System
Facilitator
Student
Cloud
Environment
Student and
ContentOntology
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
110 | P a g e
www.ijacsa.thesai.org

Fig. 2 OWL Viz representation of Document ontology
Domain Object Property Range
Definition Depends_on Content
Content Requires Prerequisite
Sequence no Is_based_on Content
Format For_a_particular Content
Test Evaluation Content
FAQs For_each Content
Fig. 3 Relationship between various classes
The year to which he belongs can be added. Restrictions
can be added by selecting the property assertion view. For
example, he belongs to any year but value for year must be
less than four. Using the equivalent class tab, conditions can
be asserted. For example: Student and belongsto some integer
[,=4^^integer] hasdetail some profile. The subclass feedback
provides the mentors or facilitators a timely statement so that
the facilitators can improve the course. The student ontology
has been shown in Figure 4. The relationship defined between
the various subclasses has been shown in figure 5.

Fig. 4 Onto Graph representation of Student ontology
Domain Object Property Range
Profile Has _skill Cognitive Skills
Profile Has_style Learning Style
Profile Has_some feedback
Fig. 5 Onto Graph representation of Student ontology
3) Facilitators Model
Facilitators should be able to enter any of the content in
any formats as per convenience. The tutorials can be used by
faculties also for their benefit. The tutor when online will have
the capability of clearing the doubts raised by students. The
questions with its answers will enter the GAQs pool. The
domain specialized author will be authorized to verify and
rank the content. The evaluation metric can be for assessment
of students performance. It also indicates to which level of
learning the student has progressed.
A domain expert is involved to develop the course into
various modules. With those various modules, ontology has to
be constructed. Each module or content must be mapped with
the learning objectives. The FAQs must be evaluated for it
content by the domain expert.
C. Design of Flow
Student:
As a student logs onto his system, his profile will be
considered.
As he request for certain content, his preferred
learning style and cognitive style will be taken into
account.
Depending on his completion of previous sessions,
topics will be presented to the user.
Brainstorming session will be followed by content if
his preferred style of cognitive learning is concrete
generic type
If questions are asked, he can directly pose the
question to the facilitator
If facilitators are offline, the question will be matched to
the existing FAQs
Facilitator
The author can upload and modify the existing content
Answers to questions of users should be completed
Domain experts must be able to rank the content and
verify the same.
Provide the right sequence of rendering of content
V. CLOUD COMPUTING FOR RURAL EDUCATION
Once the system has been developed, it should be
accessible to the rural community. For this, the Web based
system must be deployed on the Cloud. The lacunae in
infrastructure, student and teachers in the rural community can
be offset by the benefits of deploying on Cloud [7].
As the application is run in the Cloud and not on the
desktop PC, the desktop PC doesnt need the
processing power or hard disk space demanded by
traditional desktop software. Hence the client
computers, i.e. the rural schools systems can be lower
priced, with smaller hard disks, less memory, more
efficient processors, and the like. There is no
requirement of a CD or DVD drive, because no
software programs have to be loaded and no document
files need to be saved at the user side.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
111 | P a g e
www.ijacsa.thesai.org
Computers in a Cloud computing system will boot up
faster and run faster, because theyll have fewer
programs and processes loaded into memory
Cloud computing greatly reduces both hardware and
software maintenance. With less hardware (fewer
servers) necessary in the schools, maintenance costs
are immediately lowered. As to software maintenance,
as cloud applications are based elsewhere, theres no
software on the schools computers for the IT staff to
maintain.
The need to access any particular software in the cloud
now becomes less expensive. Software licenses need
not be purchased for each system. Even if it costs the
same to use web-based applications as it does similar
desktop software (which it probably wont), the
maintenance staff have saved the cost of installing and
maintaining those programs on every desktop.
When the app is web-based, updates happen
automatically and are available the next time the user
logs in to the cloud. Whenever the student accesses a
web-based application, the updated version of the
application is being downloaded.
All contents are instantly available from whichever
community schools have been registered with the
education system
The documents can be accessed simultaneously; the
modifications done by one user is automatically
reflected by which the other users can see onscreen.
Fig. 6 Life Cycle of Deployment of system on Cloud
The web based education system will be implemented
through various stages as shown in figure 6.
In the requirements phase, the feasibility study is to be
carried out. The number of schools, users including the
students and faculty should be assessed. The key processes
such as content loading and updating; student updating of
assignment downloading of contents, availability of software
pertaining to the format of the content to be downloaded
should be assessed. The scope of the project should also be
analyzed.
This includes the content will be formed for how many
courses, the formats available for a particular content, the no
of users and the number of schools accessing the web based
application. The risks also has to be accessed which can occur
due to disagreement in the content of a course. It can also
occur due to loss of information, change of author etc .In the
analysis and design phase, the various components for
students, faculties and authors have to be tested. Phase wise
testing will take place. The web based education system will
be uploaded on a server however the capacity required as per
users, data should be considered and decided.
The implementation phase will be for deciding the scope
of proof of concept. The border of environment will be
formed. Here the number of users, schools, and the content
will be worked for the initial POC. The entire POC will be
revisited depending on the feedback and issues that arises in
the design and planning phase.
VI. CONCLUSION
The paper aims to present in details the components of a
rural education system. The drawbacks and the challenges
have briefly discussed. The technological solution to fill the
lacunas of the present rural education scenario was briefly
explained.
The ontology has been used for representing the
knowledge repositories with respect to students and content.
By deploying the web based semantic education system on to
cloud help address several other issues regarding the rural
education scenario.
REFERENCES
[1] T.Coladarce, Improving the yield of Rural Education Research: An
Editors Swan Song, Journal of Research in Rural Education, 2007, 22(3)
[2] S.Chakraborty and Bhatacharya, Shishak: An Intelligent tutoring System
Authoring tool for Rural Education, Information and Communication
Technologies & Development, 2007 ICTD 2007,pp1-10
[3] M. Hall , The Challenges for Indias Education System, Chatham House,
Asia Programme, April 2005, ASP BP 05/05
[4] BiswanathDutta and Devika P Madalli, Ontology Supported
Personalized E-learning repositories, University of Trento, August 2009,
Technical report, #DISI-09-052
[5] S.Babu, FOSS Movement and its impact education, CSI
Communications, Knowledge Digest for IT community, Issue No 6, Sept
2011, Pp 19-22
[6] A Kumar and R Rajendra, Perspectives on ICT Education in India,
[7] Micheal Miller, Cloud Computing, 2008, Edition
Phase Activity
Requirements Define Objectives
Feasibility study
Validate scope of the project
Understand Key processes
Study the Application Landscape
(Environment)
Understand the key requirements
List the risks, dependencies &
impacted components
Planning &
Analysis
Design a Rollout plan
Define test cases
Design Document
Validate the capacity planning of
the servers
Implementation Setup POC environment for pilot
application and build platform
Test the Upgrade & amend the
target design as per results
Assess any performance and
availability related issues.
Revisit the design and update as
needed
Prepare the Master rollout
schedule
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
112 | P a g e
www.ijacsa.thesai.org
Efficient Threshold Signature Scheme

Sattar J Aboud
Department of Information Technology
Iraqi Council of Representatives
Baghdad-Iraq


Mohammad AL-Fayoumi
Faculty of Computer and Information Systems,
Umm Al-Qura University
Saudi Arabia


Abstract In this paper, we introduce a new threshold signature
RSA-typed scheme. The proposed scheme has the characteristics
of un-forgeable and robustness in random oracle model. Also,
signature generation and verification is entirely non-interactive.
In addition, the length of the entity signature participate is
restricted by a steady times of the length of the RSA signature
modulus. Also, the signing process of the proposed scheme is
more efficient in terms of time complexity and interaction.
Keywords- Shamir secret sharing; threshold signature; random
oracle model.
I. INTRODUCTION
Disclosure of a private key for non-cryptography purposes
for example a compromise of the basic system, human mistake
or insider attacks, is actually the highest threat to many
cryptography schemes. The most generally suggested solution
is distribution of the private key over multiple servers by secret
sharing. For digital signature, the primitive we deal with in this
paper is the main direction of this thought threshold signature
scheme.
However, the interesting type of secret sharing scheme
contains threshold scheme with a set of
n
participants. Their
access structure contains all subgroup of
t
or more participants.
Such schemes are called
t
out of
n
threshold schemes or just
) , ( n t
schemes. Threshold scheme was independently presented
by Shamir [1]. This scheme is relied on polynomial
interpolation over a finite field. Suppose
) (q GF K =
is a finite
field with
q
elements. To build a
) , ( n t
threshold scheme a
dealer
D
selects
n
distinct nonzero numbers of
) (q GF
indicated
by
n
x x ,...,
1
, and passes
i
x
to
i
P
upon a public key channel
) ,..., 1 ( n i =
. For a secret
) (q GF Ke
,
D
arbitrarily selects
1 t
set
1 1
,...,
t
a a
from
) (q GF
and builds a polynomial
i
t
i
i
x a K x f * ) (
1
1

=
+ =
.The share for participant
i
P
is
) (
i i
x f s =
.The degree of
) (x f
is at most
1 t
. It is documented that
Shamir scheme is perfect. That is, when a collection of fewer
that
t
participants work together, their original doubt about
K
is
not reduced. Suppose that any subset of
r
players out of
R

generate a signature, but reject the generation of a valid
signature when less than
r
players involve in the scheme. This
unforgeability characteristic must keep even when certain
subgroup of fewer than
r
players are cheated and act mutually.
For a threshold scheme to be practical if certain players are
cheated, it must also be strong, meaning that cheated players
must not be capable to stop honest players from creating
signature. In this paper, we will consider suggested scheme
which face at least one of the following difficulties:
a) With no accurate security proof, even with a random
oracle model.
b) Signature generation and verification is not
interactive.
c) The length of an entity signature explodes linearly in
the number of players.
To enhance this, we will introduce a new threshold RSA-
based signature scheme which faces these difficulties. We will
highlight that the signature outcome is an entirely invert RSA
signature, meaning that the generation and verification
algorithms are the same as for common RSA signature. But,
there are certain limitations on the public key which should be
a prime and the modulus should be the result of two strong
prime numbers. The suggested scheme is easy to calculate, and
has not previously suggested. However, preceding schemes of
threshold signature have that
1 + = w r
. This generalization is
practical in situations where the honest players is not necessity
choose what they are signing, but capable to verify that a big
number of them have authorized a specific signature. In
specific, threshold signatures with
w R r =
and
3 / R w<
is
used to decrease the lengths of the messages pass in
coordinated network agreement scheme [1]. The use to
coordinated network agreement was in fact the original purpose
for this study. Almost all preceding work on threshold
signatures supposes with a coordinated network, and any
players in some way simultaneously agree to commence the
signing scheme on a known document. Obviously, we cannot
act in such a system when we desire to employ coordinated
network agreement protocol.
We also highlight that the idea of a twin parameter
threshold scheme gives robust security than one parameter
threshold scheme; such scheme is actually more challenging to
build and to discuss. The proposed idea of a twin parameter
threshold scheme must not be confused with a vulnerable idea
that from time to time seems in a threshold cryptosystem
research [2]. For this vulnerable idea, there is a parameter
w r >
'
where the rebuilding algorithm needs
'
r
shares, but the
security is lost when only one truthful player discloses a share.
In proposed idea, no security is lost unless
w r
truthful
players disclose their shares. We work with a static cheating
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
113 | P a g e
www.ijacsa.thesai.org
system; the opponent should select which players to cheat at
the start of the attack. This is in line with preceding studies into
threshold signatures, which also suppose static cheating. The
proposed system can be verified if
1 + = w r
in the random
oracle model using the RSA signature.
II. RELATED WORK
In 1989, Desmedt and Frankel [3] describe the difficulty
with threshold signature scheme. This appear from the truth
that polynomial interpolation by a coefficient ring
) (n
Z
u
such
that
n
the RSA modulus and
u
is the Euler phi. Also, Desmedt
and Frankel in 1991[4] return again to the difficulty of
threshold, and introduce a non-robust threshold scheme that is
non-interactive but with small share length and without security
discussion. Frankel and Desmedt in 1992 [5] introduce
approach that providing a proof of security for a non-robust
threshold scheme with small share length, but which needs
coordinated interaction. Harn in 1994 [6] introduces a robust
threshold scheme with small share length that also needs
coordinated interaction. Gennaro et al. in 1996 [7] describe a
robust threshold scheme with small share length, but again
needs coordinated interaction. Actually, Gennaro et al. scheme
can be examined with no reconstruction of random oracle. But
this will have some practical disadvantages, demanding a
particular relationship between the sender and receiver about
the share of a signature. It appears that the security of these
systems needs carefully examination by an acceptable
approach. However, the above schemes are interactive and any
threshold signature scheme relied on integer factoring seems
inevitable to be interactive, because such signature schemes are
randomized, and thus the signers have to create random values,
which actually needs coordinated interaction.
But, in 1996 De Santis et al. [8] introduce a variant scheme
that uses interaction for large share length. This scheme
prevents the difficulties of polynomial interpolation over
) (n
Z
u
by working with
)) ( /( ) (
) (
i i Z
q n
u
u
, such that
) (i
q
u
is the
th
q
polynomial taken
) ( mod n u
, and
q
is a prime larger than l. This
is suitable, as standard secret sharing method can be directly
used, but guides to a more difficult scheme that need
coordinated interaction. In 1998, Rabin [9] suggests a strictly
robust threshold scheme that has small share length, but need
coordinated interaction. This scheme takes a diverse line of the
interpolation over
) (n
Z
u
problem, avoiding it by presenting an
additional layer of secret sharing and a lot more interaction. In
2006, Jun et al [10] described a non-interactive verifiable secret
sharing scheme built by Shamir secret sharing scheme for
secure multi-party communication scheme in distributed
networks. In 2007, Li et al. [11] they introduce a secure
threshold signature scheme without trusted dealer. In the
meantime, the signature share generation and verification
algorithms are non-interactive. In 2010 Gu, et al. [12] discuss
the security of Jun et al. scheme and show that their scheme
cannot withstand the misleading performance as they claimed.
' '
A A A A
w w z z =
.
III. SCHEME REQUIRMENTS
There are three entities the player
R
, the dealer and an
opponent. There are also a signature verification phase, a share
verification phase and a share combination phase. In addition,
there are two other variables,
w
represent number of cheated
players; and
r
denote the number of signatures required to get
a signature. The only restrictions are that
1 + > w r
, and
r w R >
.
The opponent chooses a subset of
w
players to cheat. In
the dealing phase, the dealer establishes a public key
e
and
private key shares
R
sk sk ...
1
, and verification keys
R
vk vk ...
1
.
The opponent gets the private key shares of the cheated players
and the public key and verification keys. Following the dealing
phase, the opponent passes signing demand to the honest
players for document of his choice. Upon such a demand, a
player results a signature share for the known document. The
signature verification phase obtains a document, a signature
and the public key, then verifies whether the signature is valid
or not. The signature share verification phase obtains a
document, a signature share on that document from players
i
,
with
x x
vk sk pk , ,
, and verifies whether the signature share is
valid or not.
IV. THE PROPOSED SCHEME
In this section, we describe the proposed scheme.
The dealer. The dealer must do the following:
1. Selects arbitrarily two primes p and q , such that
1 * 2
'
+ = p p , 1 * 2
'
+ = q q with
' '
, q p are also primes.
2. Finds the modulus q p n * = .
3. Selects the message
' '
*q p m= .
4. Selects the public key e as a prime 1 > e .
5. The public key is ) , ( n e .
6. Finds m e d mod
1
= .
7. Let d v =
0
.
8. Selects
x
v arbitrarily from ) 1 ,..., 0 ( m for

1 1 s s r x .
9. The vector
1 0
,...,
r
v v determine the polynomial
x
r
x
x
i v i f * ) (
1
0

=
= .
10. Finds m x f s
x
mod ) ( = for R x s s 1 . (1)
11. This element
x
s is a secret key share of player x .
This indicate by
n
D the subgroup of squares in
*
n
Z .
12. Selects an arbitrary
n
D ue .
13. Finds
n
s
x
D u u e = for R x s s 1 .
14. These statements determine the verification keys,
u vk = and
x x
u vk = .
Remarks. We will ensure that all set computations are
performed in
n
D
, and equivalent exponent arithmetic in
m
Z
.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
114 | P a g e
www.ijacsa.thesai.org
This is suitable, because
' '
*q p m=
has no small prime factors.
Because the dealer selects
n
D ue
arbitrarily, suppose that
u

creates
n
D
, because this occurs with all but small probability.
Since of this, the number
x
u
entirely find out the result of
m s
x
mod
. For each subgroup of
r
points in
) ,..., 0 ( R
, the
result of
m i f mod ) (
at these points uniquely finds out the
coefficients of
m i f mod ) (
, and since the result of
m i f mod ) (

at every other point mod in
) ,..., 0 ( R
. This follows from the
information that the equivalent Vandermonde vector is
invertible mod
m
, because its determinant is co-prime to
m
.
From this, it ensures that for each subset of
1 r
points in
) ,..., 1 ( R
, the distributions of the result of
) (i f
mod
m
at these
points are standardized and equally independent. Suppose
! R a =
for each subset
s
of
k
points in
) ,..., 0 ( R
and for each
s R x \ ) ,..., 0 ( e
and
s ye
, we can describe:

[
[

e
e
=
) ) ( \
) ) ( \
,
'
( '
'
( '
*
y y
y x
y s y
y s y
s
y x
a H
(2)
These results are resulting from the standard Lagrange
interpolation equation. They are obviously integers; hence the
denominator divides
)! ( ! y R y
which in divides
! R
. It is also
obvious that these results are easy to calculate. From the
Lagrange interpolation equation, we have:
m y f H x f a
S y
S
y x
mod ) ( ) ( *
,
e
(3)
Valid signature. We require a hash function
h
to elements
of
*
n
Z
. If
) (m h i =
, thus the valid signature on
m
is
*
n
Z j e

where
i j
e
=
. This is only a common RSA signature.
Generating signature share: In order to generate a
signature share on a document
m
we should do the following.
1. Choose ) (m h i = .
2. The signature share of player x
is
x
s a
x
i i
* * 2
=
(4)
Correctness. The verification of correctness is just a proof
of the discrete logarithm of
2
x
i
to the base
a
i i
* 4 '
=
. (5)
However, we can simply adjust a well-known interactive
scheme of Chaum and Pedersen [13]. We collapse the scheme,
making it non-interactive, by employing a hash function to
generate the challenge such that a random oracle model is
required. We also have to handle the actuality that we are using
a group
n
D
whose order is not known. So, this is unimportantly
managed by just using adequately big integer. Suppose
) (n L
is
the bit-size of
n
. Assume that
'
h
is the hash function, whose
result is
1
L
bit integer, such that
1
L
is a security parameter. To
build the verification of correctness player
x
select a random
number
) 1 2 ,..., 0 (
1
2 ) (
e
+ L n L
r
, then finds:
-
r
u u =
'

-
r
i i
' '
=
- ) , , , , , (
' ' 2 '
i u x u i u h c
x x
=
- k c s z
x
+ = *
The verification of correctness is ) , ( c z .
Correctness. one verifies that
) * , * , , , , (
* 2 ' 2 ' ' c
x
c
x
z
x x
x i u u x u i u h c
z

= . The cause for using

2
x
i
instead of
x
i is that because
x
i is assumed to be a square, this
is not simply checked. This means, we are certain to be using
n
D , so we want to ensure soundness.
Combining shares. Assume that we have valid shares from
a group s of players, such that ) ,..., 1 ( ) ,..., (
1
R x x s
r
c = .
Assume ) (m h i = and suppose that
xy
s
a
xy
i i
* 4
2
= . Then to
rearranged shares, we find
s
xy
s
x
H
xy
H
x
i i t
,, 0
1
, 0
1
* 2 * 2
... = Such
that H is the integers described in (2). From (3), we hold
'
e e
i t = , thus
2 '
* 4 a e = (6)
as 1 ) , gcd(
'
= e e , it is simple to find j where i j
e
= , employing
a standard method
b q
i t j * = such that v and b are integers
where 1 * *
'
= + b e v e ; that can be got from the extended
Euclidean method on
'
e , ande
V. SECURITY DISCUSSION
Theorem 1: the proposed scheme is a secure threshold
signature protocol if the common RSA signature scheme is
secure. We illustrate that to simulate the opponent vision, if the
opponent requests for a signature share from the honest player.
Assume
1 1
,...,
r
x x
is the set of cheated players. Consider
m x f s
x
mod ) (
for all
R x s s 1
, and
m f d mod ) 0 (
. To
simulate the opponent vision, we just select the
xy
s
belonging
to the group of cheated players randomly from the vector

) 1 4 / ,..., 0 ( n
.
We have by now discussed that the cheated player private
key shares are arbitrary numbers in the vector
) 1 ,..., 0 ( m
. We
hold
) ( 4 / 1 2 / ) ( 4 /
2 / 1 ' '
n O q p m n = + + =
; and from this
easy computation illustrates that the statistical distance between
the regular distribution on

) 1 4 / ,..., 0 ( n
and the regular
distribution on
) 1 ,..., 0 ( m
is
) (
2 / 1
n O
. When these
xy
s

members are selected, the values
x
s
for the honest players are
also entirely fixed mod
m
, although cannot be simply
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
115 | P a g e
www.ijacsa.thesai.org
calculated. Though, provided
j i,
and
i j
e
=
, we can simply
find
x
s a
x
i i
* * 2
=
for the honest player
x
as:
)) * ... * ( ( 2
1 1
,
1 1
, 0 ,

+ + +
=
r
x
s
r
x x x
s
x x
s
x
s H s H e H
x
j i .
Such that
) ,..., , (
1 1
=
r
x x o s
, this results from (3). Employing
this method, we can create the vector
R
u u u ,..., ,
1
, and also
create some share
x
i
of the signature, provided the common
RSA signature. This case illustrates that we described the share
x
i
to be
x
s a
i
* * 2
and not
x
s
i
* 2
. This thought was employed by
Feldman [14] in the situation of where another associated
problem of provable secret sharing.
Proofs of correctness: entity can use the random oracle
model for the hash value
'
h
to obtain soundness and
arithmetical zero-knowledge. This is very simple, but we
drawing the information.
Now, we study soundness. We need to illustrate that the
opponent cannot build, except with insignificant probability,
the proof of correctness for an inaccurate share. Assume
i
and
x
i
is provided, and a valid proof of correctness
) , ( c z
. We hold
) , , , , , (
' ' 2 ' '
i u i u i u h c
x x
=
such that:
-
a
i i
* 4 '
=
-
c
x
z
u u u

= *
'

-
c
x
z
i i i
* 2 ' '
*

=
Right away,
' ' 2 '
, , , , i u i u i
x x
are simply observed in
n
D , and
we are supposing that u create
n
D . So we hold:
a
u i =
'

x
s
x
u u =
B
x
u i =
2

j
u u =
'

g
u i =
'

For a number of integers g j B v , , , furthermore,
m j s c z
x
mod *
m g B c v z mod * *
Multiplying the first formula by v and subtracting the
second, we obtain: m g j v v s B c
x
mod * ) * (
(7)
So, the share is accurate when and only when
m v s B
j
mod * (8)
When (8) fails to retain, therefore it should be unsuccessful
to have
'
mod p
or
'
mod q
, and thus (7) uniquely finds out
mod c
one of these primes. Although in the random oracle
model, the distribution of
c
is consistent and separate from the
data to a hash value, and thus this still occurs with insignificant
probability.
Now, we will study zero-knowledge. We can build a
simulator that simulates the opponent vision without knowing
the result of
x
s
. This observation contains the results of the
random oracle at those situations where the opponent has
queried the oracle, thus the simulator is in entire charge of the
random oracle. When the opponent constructs a query to the
random oracle, if an oracle has not been determine before at the
provided point, the simulator describes it an arbitrary value,
and in all cases return the result to the opponent. If an honest
player is supposed to create a proof of correctness for a
provided
x
i i,
, the simulator selects
) 1 2 ,..., 0 (
1
e
L
c
and
) 1 2 ,..., 0 (
1
* 2 ) (
e
+ L n L
z
randomly, and for provided integers
x
i i,
determines the number of the random oracle
) * , * , , , , (
* 2 ' 2 ' c
x
z c
x
z
x x
i i u u x u i u

to be
c
. With insignificant
probability, the simulator has not described the random oracle
at this point, and thus it is limitless to do so. The proof is
) , ( c z
.
It is easy to check a distribution created by this simulator is
statistically near to perfect.
From soundness, we obtain the strength of the threshold
signature protocol. From zero-knowledge, we obtain the non-
forgeability of the threshold signature protocol, supposing the
common RSA signature scheme is secure, that is existentially
non-forgeable anti-adaptive chosen message attack. Such
approach is more correct in the random oracle model for
h
, this
typed follows from the RSA-based provided random
*
n
Z i e
, it
is difficult to find
j
where
i j
e
=
.
VI. CONCLUSION
In this paper, we illustrated the threshold signature scheme.
We introduced a strong threshold signature scheme relied on a
secret sharing scheme. The suggested signature scheme
simplifies threshold RSA signature in which relied on Shamir
secret sharing, and is an efficient. In addition, the method can
be extended to further public key cryptography as the secret
key is utilized in the exponent.
REFERENCES
[1] Cachin C, Kursawe K, and Shoup V, Random oracles in
Constantinople: practical asynchronous Byzantine agreement using
cryptography, Manuscript, 2000.
[2] Micali S and Sidney R, A simple method for generating and sharing
pseudo-random functions, with applications to Clipper-like key escrow
systems, Advances in Cryptology, Crypto'95, pages 185-196, 1995.
[3] Desmedt Y and Frankel Y, Threshold cryptosystems, Advances in
Cryptology Crypto'89, pp, 307-315, 1989.
[4] Desmedt Y and Frankel Y, Shared generation of authenticators and
signatures, Advances in Cryptology, Crypto'91, pp. 457-569, 1991.
[5] Frankel Y and Desmedt Y, Parallel reliable threshold multi-signature,
Technical Report TR-92-04-02, University of Wisconsin, Milwaukee,
1992.
[6] Harn L, Group-oriented (t; n) threshold digital signature scheme and
digital multi-signature, IEE Proceeding Computer Digital, Tech.,
141(5):307-313, 1994.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
116 | P a g e
www.ijacsa.thesai.org
[7] Gennaro R, Jarecki S, Krawczyk H, and Rabin T, Robust threshold
DSS, Advances in Cryptology, Eurocrypt'96, pp. 354-371, 1996.
[8] De Santis A, Desmedt Y, Frankel Y, and Yung M, How to share a
function securely, 26th Annual ACM Symposium on Theory of
Computing, pp. 522-533, 1994.
[9] Rabin T, A simplified approach to threshold and proactive RSA,
Advances in Cryptology Crypto'98, 1998.
[10] Jun Ao and Guisheng Liao, A Novel Non-interactive Verifiable Secret
Sharing Scheme, Chunbo Ma, Communication Technology, ICCT'06,
International Conference on 27-30 November 2006, pp. 1 4.
[11] Jin Li, Tsz Hon Yuen and Kwangjo Kim, Practical Threshold
Signatures without Random Oracles, Lecture Notes in Computer
Science, 2007, Volume 4784, 2007, 198-207.
[12] Feng Wang, Yousheng Zhou, Yixian Yang and Yajian Zhou, Comment
on a Novel Non-interactive Verifiable Secret Sharing Scheme,
Communication Software and Networks, ICCSN'10, Second
International Conference on, 26-28 Feb. 2010, pp. 157-159
[13] Chaum D and Pedersen T, Wallet databases with observers, Advances
in Cryptology, Crypto'92, pp. 89-105, 1992.
[14] Feldman P, A practical scheme for non-interactive verifiable secret
sharing, 28th Annual Symposium on Foundations of Computer Science,
pp. 427-437, 1987.
AUTHORS PROFILE
Sattar J Aboud is a Professor and advisor for Science and Technology at
Iraqi Council of Representatives. He received his education from United
Kingdom. Dr. Aboud has served his profession in many universities and he
awarded the Quality Assurance Certificate of Philadelphia University, Faculty
of Information Technology in 2002. Also, he awarded the Medal of Iraqi
Council of Representatives for his conducting the first international
conference of Iraqi Experts in 2008. His research interests include the areas of
both symmetric and asymmetric cryptography, area of verification and
validation, performance evaluation and e-payment schemes.
Mohammad Al-Fayoumi is a Professor at Umm Al-Qura University in Saudi
Arabia. He received his education from Romania. Dr. Fayoumi has served his
profession in many universities. His research interests include the areas of
software engineering, verification and validation, information security and
simulation.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
117 | P a g e
www.ijacsa.thesai.org
Fault Tolerant Platform for Application Mobility
across devices
T. N. Anitha
1

Associate Professor, Dept of Computer Science &
Engineering , SJCIT, Chickballapur-
562101,Karnataka,India
Jayanth. A
1

Project Engineer,Oracle India Private Limitted
Bhannerughatta Road, Near Diary circle, Koramangala,
Bangalore, Karnataka,India


AbstractIn the mobile era, users started using Smartphones,
tablets and other handheld devices, The advances in telecom
technologies like 3G accelerates the migration towards smart
phones. But still battery power and frequent change of handsets
is still a constraint. They burden on user had to manually
synchronize their contacts, applications they use to the new
phones. Also they loss whatever they are doing when the mobile
get power down. In this paper, we propose a solution to the
problem discussed with a new fault tolerant platform which can
provide application mobility across the devices.
Keywords- Fault tolerance; Middle ware; Midstore Manager.
I. INTRODUCTION
Application mobility refers to the idea application can run
across multiple devices seamlessly even if it fails on one device
say to battery down or users move to some other device. The
application does not stop and it continues doing its work across
users movement to devices. We choose smart phones as the
device and laptops as the devices to consider for application
mobility. Application can be anything from word editing,
preparing the power point presentations etc.
We explain the case for application mobility with some
scenarios. Say a user is a browsing a web page in his
smartphone, his battery power is low, can he work with he
work with the same page in browser in laptop automatically, or
he has another phone and want to work with same page in the
handheld. Say the user is watching a you tube video, he is at 2
min another 3 min video is pending, he wants to watch this
video on his tablet from the 2min or his phone switches off and
he powers again, can he still watch the video from 2min
automatically. Currently there is no solution available in
market for these problems which motivated our work. In this
paper, we address these challenges and design a software
solution these problems.
II. RELATED WORK
Fault Tolerance solutions usually are designed for
production servers like air traffic control, distributed disaster
system, railways reservation system, internet banking where a
single fault may lead to huge loss of money and even human
lives. Replication based technique is one of the popular fault
tolerance techniques [1]. A replica means multiple copies.
Replication is a process of maintaining different copies of a
data item or object. In replication techniques, request from
client is forwarded to one of replica among a set of replicas.
This technique is used for request that do not modify state of
service. Replication adds redundancy in system. In this way
failure of some nodes will not result in failure in system and
thus fault tolerance is achieved.
Checkpoint with rollback-recovery is a well-known
technique. Checkpoint is an operation which stores the current
state of computation in stable storage. Checkpoints are
established during the normal execution of a program
periodically. This information is saved on a stable storage so
that it can be used in case of node failures. The information
includes the process state, its environment, the value of
registers, etc. When an error is detected, the process is roll
backed to the last saved state [2].
Although replication method is widely used as a fault
tolerance technique but number of backups is a main drawback.
Number of backups increases drastically as coverage against
number of faults increases. As the number of backup increases
management of these backups is very costly. Fusion based
techniques overcome this problem. It is emerging as a popular
technique to handle multiple faults. Basically it is an alternate
idea for fault tolerance that requires fewer backup machines
than replication based approaches. In fusion based fault
tolerance a technique, back up machines is used which cross
product of original as fusions is corresponding to the given set
of machines [3]. Overhead in fusion based techniques is very
high during recovery from faults. Hence this technique is
acceptable if probability of fault is low.
All these fault tolerance solutions address only for the
server systems. We cannot use these solutions directly for our
problem; all these solutions are designed for applications
running in homogenous platform and the platform on which
application run are same across devices.
But in our problem we need to work across devices with
different platform. We need to provide application mobility for
application running in Android platform to the Application
running in Symbian platform or to windows on a laptop. So the
fault tolerance solution for application mobility becomes even
more complex.
III. PROPOSED SOLUTION
The proposed solution consists of designing a middleware.
The middle ware will provide fault tolerant application
mobility. The middleware provides APIs for application to
synchronize essential information.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
118 | P a g e
www.ijacsa.thesai.org
The middleware can write the synchronization information
into a memory card and it can also read this synchronization
information and start the applications. We could have used
cloud storage for synchronization of information, but relaying
on cloud will sometimes become a problem like network not
available. Using memory card has its advantages. The read and
write operations for synchronization becomes very much faster.
The memory card can be used to stat the application seamlessly
on another laptop or other handheld easily, just insert the
memory card and start the middleware.









So proposed system now standardizes the protocols and the
standards used for Application Mobility.
A. Middleware
There are two approaches for the design of middleware. In
first case, the application has to manually call the APIs to set
the synchronization information. In the next case, middle ware
is dynamic and abstract the device APIs.
In this work , we limit our scope to application has to
manually call the middles API.
Middleware APIs should take a minimal amount of time
for providing the fault tolerance without affecting the
application performance. Poor design of middleware will
result in lower application performance and will seriously
impact user experience. Middle ware records the session
activity in a XML format in the Memory Card.
Since the Memory Card is also used by user for his
contents, we should provide secure and incorruptible way for
our XML. This can be provided by creating a folder in the
Memory card for the purpose of middleware alone and using a
virtual file system. These way middleware operations are
assured a separate workspace in the Card. This virtual file
system is referred as MidStore from here on.
Applications uses the middleware APIs to record the
current application session activity into the MidStore. At any
point of time, the information in the middleware provides the
snapshot of mobile users current activity. The information is
MidStore has this snapshot. Our job of Application mobility in
case of fault tolerance is very easy, if we can reconstruct this
snap on any other device or platform. To make this easy, first
step is standardizing the format of recording the snapshot. We
use XML standard for this purpose. XML is well standardized
way to share data across cross platforms.
Middleware must also provide a reconstruct job as service
which can invoked on the other platform to reconstruct the
snapshot. Based on the discussion above, so far we formulate
the architecture of middleware as
















The core of our proposed solution is in the Middleware App
Layer, Mid Store Manager, MidStore and Snap Re constructor.
B. Middleware App Layer
This layer provides the APIs for the application to store the
session activity in the MidStore. Say the Application is browser
on the mobile handset. User has entered a URL and got the
webpage. This activity must be recorded into the midstore.
Say the Middleware provides the API like below.
RecordActivity (Application, DescritorParam);
The Browser Application will call RecordActivity with the
parameters filled as
Application: IE
DescriptorParam: www.yahoo.com
Once this API call is made, middle API Layer delegates the
API to the MidStore manager.
We care not addressing any specific implementation of
APIs in this paper. But as a general guideline, the API should
give the following information about user activity to the
Middleware platform. It should provide the activity is related to
which application and the additional descriptor of the activity.
If the user is browsing a webpage with internet explorer, the
application is IE and the descriptor is URL for browsing. If the
user is playing a song with media player, the application is
WMP and the descriptor is the song name.
C. MidStore Manager
As discussed previously the memory card have a virtual file
system called as MidStore. In order to manage the read and
write to the MidStore we need the MidStore manager.
At Handset At Laptop
Application
Middleware
Application
Middleware
Memory
Card
Memory Card

Applications
Applications

Middleware
API Layer

Mid Store
Manger
MidSt
ore
Mid Store
Manger

Snap
Re-constructor
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
119 | P a g e
www.ijacsa.thesai.org
MidStore Manger handles the api call from the APP Layer
and records this activity in the MidStore. It is upto the specific
implementation of Mid Store Manager , to record all the
activity in a single XML file or make separate XML file for
each application. The case of making separate XML for each
application has a particular advantage during reconstruct
process. This will be addressed in the Snap re constructor.
We only suggest a way for the XML format, but it is upto
the specific implementers to have their own way
<midstore>
<application>
< instance = IE/>
< descriptor url =www.yahoo.com />
</application>
</midstore>
Suppose the User have closed the browsing application , the
browser should call an API
RemoveInstance(Application , Descriptor )
With Application value as IE.
When this api is called, the midstore manager must remove
the application entries for IE.
D. SnapReConstructor
SnapReConstructor is an important module in our design. It
reads the midstore and replicates the user activity on any target
platform.
Suppose the mid store has the following
<midstore>
<application>
< instance = IE/>
< descriptor url =www.yahoo.com />
</application>
</midstore>
And reconstuctor is running on desktop system with
windows. Then by reading this snap the reconstuctor should be
able to open the IE browser with url as www.yahoo.com.
SnapReConstructor can also have additional logic called as
ApplicationAliter.
ApplicationAliter can help partial replication in target
platform. Say IE is not in the target platform, but its found out
that Mozilla is installed in the target. Since both IE and Mozilla
are web browsers, SnapReConstructor can start the Mozilla
with url as www.yahoo.com. So this guarantees partial
restoration.
SnapReConstructor must always try to achieve maximum
restoration without user intervention. Instead of
SnapReConstructor to be provided as an application and user
always click to start , it can be made auto start whenever the
device is introduced into target platform. But we leave this to
implementers choice.
For any new application to be supported by the middleware,
we need to extend the middleware API layer for providing
parameters and SnapReConstructor must be extended to start
the application or must ApplicationAliter must be extended for
doing partial reconstruction.
Instead of application reading each application tag in the
MidStore , if it can get all instances for that specific
application, it can start the application at one start with all
instances. We find this very useful in certain targets for some
applications. In windows, stating a IE can be done with
multiple url , so one ie instance comes up with different tabs
for each url.
Not all application available in mobile platform is available
for desktop platform. Also not all applications are available
across different mobile platform. But popular applications for
consumer user and the enterprise application are moving
towards availability in all platforms. Also with the use of
ApplicationAliter we can always find alternative for the
applications.
E. Application of proposed Platform
In this section we detail the possible usage scenarios for our
project.
A user adds contacts to his phone. He adds as many
contacts to his phone. Since mobile phones are cheaper and
new model comes to market every day, and in rapid mobile use
in countries like India , people often change phone , so every
time they change they have sync with contacts , important
messages , calendar events etc , but with our platform in place
, the user can just change the handset and connect to his
computer, the platform will immediately get the application
information like messages , calendar events etc applications
continues in the new handset.
F. Platform against Battery Backup
Modern smart phones even claim a battery backup of 10
days or so, but for busy business users, they see that battery
does not last for even 2 days. Many times business user has to
carry another phone. He has to continue his operation using the
new handset, but from the same point of continuity.
Currently there is no solution for this problem and the
transitory phase for business user from handset to handset is a
very bad experience, but with our platform the transitory is
smooth for the user, with the application snapshot same as he
worked previously.
Since our platform is not so complex and uses less
resources it can even run with low end platforms.
IV. CONCLUSION
In this paper work, we have provided a solution for
application fault tolerance and mobility across different
platforms. Any Implementation can use the solution to realize it
on different platforms. The Application mobility will be of
great advantage to the enterprise and consumer applications.
The application has to change a bit to save the user activity into
the MidStore. This will be a disadvantage for the existing
applications. In Android platform , with the concept of
Broadcast receiver, any application can watch for certain
events , so browsing a url , playing a song etc are all events , so
a watcher application cab be written to watch for the events and
record user activity to the MidStore. We are looking for the
same kind of solutions on other platforms too. Once we are
able to find the generic watcher solution, the Watcher
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
120 | P a g e
www.ijacsa.thesai.org
component can also be added to the middleware. This will be a
big advantage for providing mobility for existing applications.
REFERENCES
[1] Bhargava B. and Lian S. R., Independent Checkpointing and
Concurrent Rollback for Recovery in Distributed Systems-An Optimistic
Approach, Proceedings of 17th IEEE Symposium on Reliable
Distributed Systems, pp. 3-12, 1988.
[2] Cao G. and Singhal M., On coordinated checkpointing in Distributed
Systems, IEEE Transactions on Parallel and Distributed Systems, vol.
9, no.12, pp. 1213-1225, Dec 1998.
[3] V. Agarwal, Fault Tolerance in Distributed Systems, I. Institute of
Technology Kanpur, www.cse.iitk.ac.in/report-repository, 2004. ,
[4] "XML Media Types, RFC 3023". IETF. 2001-01. pp. 911.
http://tools.ietf.org/html/rfc3023#section-3.2. Retrieved 2010-01-04.^
"XML Media Types, RFC 3023".
[5] Adnan Agbaria, William H. Sanders, Distributed Snapshots for Mobile
Computing Systems, Proceedings of the Second IEEE Annual
Conference on Pervasive Computing and Communications](Percom04),
pp. 1-10, 2004.
[6] Parveen Kumar, Lalit Kumar, R K Chauhan, A low overhead Non-
intrusive Hybrid Synchronous checkpointing protocol for mobile
systems, Journal of Multidisciplinary Engineering Technologies,Vol.1,
No. 1, pp 40-50, 2005.
[7] Parveen Kumar, Lalit Kumar, R K Chauhan, Synchronous
Checkpointing Protocols for Mobile Distributed Systems: A
Comparative Study, International Journal of information and computing
science, Volume 8, No.2, 2005, pp 14-21
[8] "XML Serialization in the .NET Framework". Msdn.microsoft.com.
http://msdn.microsoft.com/en-us/library/ms950721.aspx. Retrieved
2009-07-31
[9] V.K Garg,. Implementing fault-tolerant services using fused state
machines, Tech-nical Report ECE-PDS-2010-001, Parallel and
Distributed Systems Laboratory,ECE Dept. University of Texas at
Austin (2010).
[10] Extensible Markup Language (XML) 1.1 (Second Edition)".W3.org.
http://www.w3.org/TR/xml11/#charsets. Retrieved 2010-08-22.IETF.
2001-01. pp. 79. http://tools.ietf.org/html/rfc3023#section-3.1.
Retrieved 2010-01-04.
[11] M. Murata, D. Kohn, and C. Lilley (2009-09-24). "Internet Drafts: XML
Media Types". IETF. http://tools.ietf.org/html/draft-murata-kohn-lilley-
xml-03. Retrieved 2010-06-10.
[12] "XML 1.0 Specification". W3.org. http://www.w3.org/TR/REC-xml.
Retrieved 2010-08-22.
[13] M. Wiesmann, F. Pedone, A. Schiper, B. Kemme, G. Alonso,
Understanding Replication in Databases and Distributed Systems,
Research supported by EPFLETHZ DRAGON project and OFES).
[14] Checkpoint-based Fault-tolerant Infrastructure for Virtualized Service
Providers.
[15] A Review of Checkpointing Fault Tolerance Techniques in Distributed
Mobile Systems.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
121 | P a g e
www.ijacsa.thesai.org
Viable Modifications to Improve Handover Latency
in MIPv6

Mr.Purnendu Shekhar Pandey
Department of ICT,
Gautam Buddha University,
Gr.Noida, Gautam Buddha Nagar, Uttar Pradesh India
Dr.Neelendra Badal
Department of Electrical Engineering,
Kamla Nehru Institute of Technology, Sultanpur,
Uttar Pradesh, India


Abstract Various Handover techniques and modifications for
Handover in MIPv6 come into light during past few years. Still
the problem remains, such as quality of services, better resource
utilization during Handover and Handover latency. This paper
focuses on such problems within various Handover techniques
and proposes some modifications to reduce the Handover latency.
This also improves the quality of services related with Handover
in MIPv6. Experimental results presented in this paper shows
that the Handover latency in MIPv6 will be reduced by applying
these proposed modifications. This paper is organized in
following sections: section I gives the introduction, Section II
presents the Basic operations for MIPv6 Handover, Section III
focuses on the related work with background, Section IV
proposes modifications and Section V concludes the paper while
underlining future prospects in this domain.
Keywords- Handover; CoA; Home Agent; Foreign Agent; MIPv4;
MIPv6.
I. INTRODUCTION
Various Handover techniques for MIPv6 is applied so that
the user enjoys continuous internet connectivity and avoids
rebooting their application as they move from one subnet to
another subnet.
Internet connectivity is based on certain protocols such as
Transmission Control Protocol (TCP) and Internet Protocol
(IP). Such protocols require a unique IP Address for identifying
the physical location of the Mobile Node (MN) for setting up
the connection but the mobility of the MN is only achieved if
its IP address keeps on changing from one subnet to another.
To overcome this problem, Mobile IP was developed which
removes the problem by providing MN with two types of
addresses, i.e. first, a Home Address (HA) that does not change
as the node moves and second, a Care-of-Address (CoA) that
keeps on changing as the MN moves from one subnet to other
subnet [4].
Mobility support was first incorporated using Internet
Protocol Version 4(IPv4) bringing forth MIPv4,later on
Internet Engineering Task Force (IETF) developed another
Mobile IP that is MIPv6.The major factor that led to switching
from MIPv4 to MIPv6 are Route Optimization in MIPv6 and
Triangular Routing problem that existed in MIPv4.
Until now, there are approaches to solve the problem of
Handover such as L3- Driven Fast Handover which not only
uses network layer (L3-layer) information but also uses the
link-layer (L2-Layer) information for better Handover process
and Resource Efficient CoA Provisioning which makes use of
the various caches such as active proxy cache and active
garbage cache for better performance in terms of Handover and
quality of services.
The basic operations related to Handover are presented in
the next section.
II. MIPV6 BASIC OPERATIONS: A QUICK LOOK
MIPv6 protocol was developed which allows MN to be
communicated and reachable while moving around in IPv6
internet. MIPv6 performs mobility with the help of three
addresses such as HA (static address), CoA and Link Layer
Address (Prefix).
Packets are to be transmitted to the MN using HA
without actually bothering about MN current point of
attachment to the internet. Only HA is not sufficient at all, the
CoA generated by the Foreign Agent (FA) is also required.
After knowing CoA, a correspondent node or HA can send the
desired packets to the MN. So, these two addresses are
sufficient enough for the proper flow of the packets. HA is a
router on a MN's home network which maintains current
location information for the mobile. Now, the question arises
as to what for the Link Layer address is used. In brief, it is used
for the better Handover. Handover or Handoff refers to the
process of transferring an ongoing call or data session from one
subnet connected to the next subnet where the mobile node is
going to get attached. While, Correspondent Node (CN) is a
peer node with which a MN is communicating. It may be either
mobile or stationary.
When a MN moves from HA to FA, the actual
communication starts between MN and CN after performing
the following processes as mentioned also in figure 1.
A. Agent Solicitation
It refers to messages that are sent by the MNs which is
looking for a Router to carry out its previous communication
activity with its HA or Correspondent node [2].


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
122 | P a g e
www.ijacsa.thesai.org

Figure 1: Qualnet Simulated example of Process Involved in Handover
B. Router Discovery
In this process, the Access Router (AR) multicasts Router
Advertisement message on a loosely periodic basis and MN
senses these messages to determine whether it is in the same
AR or it has switched to other Access Routers domain.
C. Address Configuration.
As shown in Fig.1 MN configures a new global IP
address called Care-of-Address by the help of prefix received
which is present in the Router Advertisement message.
D. Movement Detection.
As shown in Fig.1 the Router Advertisement message
contains a prefix that is generated by all the access routers but
these prefixes are different for different routers respectively.
This allows the MN to check for the change in prefix and, as
soon as, it detects the change in prefix, it is able to decide that
it has now changed its domain [2].
E. Mobile IPv6 Registration
As shown in Fig.1 After following all these previous
processes, the MN comes to know that it has left its previous
domain and has entered into domain of new access router.
Now, MN has got the responsibility to get its New IP Address
i.e. CoA registered with its HA and the Correspondent Node
[5].
To make the registration more authentic and appropriate
following processes are followed:
a) a) Home Registration
The MN sends a Binding Update Message to the HA after
notifying it about the New Care-of-Address (CoA) and, in
response, the HA sends a Binding Acknowledgement message.
b) b) Correspondent Registration
As shown in Fig.1 In Correspondent Registration, the
Binding Update message are send to the Correspondent Node
and, in response, Correspondent Node sends the Binding
Acknowledgement message.
III. RELATED WORK AND BACKGROUND
To improve the Handover process various techniques are
described in this section to improve Handover process and
quality of services. The various techniques are as follows:-
A. L3 (Network Layer) Driven Fast Handover Approach
The Handover is the main concern in MIPv6.The
bottleneck of Handover is due to the fact that is that network
layer use only the network layer information to detect whether
the Handover had taken place or not. To solve this problem
several fast Handover approaches came in to lime light which
started using link layer information (L2) to speed the process of
Handover [12].
There are two sub-processes related to the main Handover
approach:
Link-Layer Handover (L2):-In this Handover, the MN
comes to know about change of AR as MN detects link layer
address change to which it is linked with.
Network layer Handover (L3)
L3 Handover consists of two phases:
Preparation Phase:-In this phase a CoA for mobile is
generated as well as Duplicate Address Detection (DAD)
protocol is executed.
Signaling Phase:-In this phase the CoA is registered with its
HA.
The evolution of L3- Driven Handover was not at all
abrupt. First of all Normal Handover Sequence was followed
then after some time. Later on Handover Sequence was
developed Using L2 information. It was followed for some
time and then L3- Driven Handover was developed which was,
indeed, more efficient than the other two previous approaches
[7].
a) Normal Handover Sequence
First of all L2 Handover will take place, whose functioning
is not at all known to network layer. After a while the MN
would receive a Router Advertisement and only then the L3
Handover would start its sub phases: preparation phase and
signaling phase will occur. However the figure clearly
establishes that there is delay of few seconds before L3
Handover starts [6].
b) Handover sequence Using L2 Information
As soon as the L2 Handover is finished, the link layer
notifies the network layer of the end of L2 Handover. After
getting this notification, the network layer sends the Router
Solicitation message to the AR and, consequently, AR starts
sending router advertisement message [7].
c) L3- Driven Fast Handover using L2 Information
As shown in the Ref [7] Fig.2 the link-layer senses the link
quality and it notifies the network-layer. As soon as the link is
down below the threshold the network- layer correspondingly
executes the preparation phase immediately.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
123 | P a g e
www.ijacsa.thesai.org

Figure 2: L3-Driven Handover
When the link layer finishes the L2 Handover, it informs
the network layer for executing its signaling phase. It is
important to know that the bottleneck of L3-Driven approach
lies in being subservient to device dependent information. It is
manifested through the radio wave strength by which to link
layer can judge exact signal strength and notify the network
layer that the link is down.
B. Resource Efficient Care-of-Address Provisioning
Duplicate Address Detection (DAD) protocol is followed so
that each MN is using a unique address CoA i.e. no other node
existed within the same domain subnet using the same global
IP address (CoA) but it was found that DAD was a time
consuming process ,which might interrupt the seamless
Handover. in e.g. RFC 2462 DAD algorithm took more than
1000 millisecond to complete the DAD process which was not
at all viable[3].
After that a protocol called aDAD (advance DAD) was
developed. The major advantage of aDAD is that it reserves
unique New Careof-Address in advance, as a result, it almost
eliminated the latency needed for the Address Configuration
and Confirmation .Another major advantage of aDAD is that it
follows a concept called Piggybacking in accordance with
which the MN sends a message called Router Solicitation to
the AR and then the AR reverts the same message giving reply
in the form of Router Advertisement [4].The major drawback
of this protocol is that it uses excessively a lot of network
resources such as bandwidth to generate new CoA and verify
its uniqueness as aDAD generates New CoA and checks their
uniqueness one by one [11].
Third protocol is called Agent-based-DAD (XDAD).Its the
most efficient approach for providing New Care-of Address
,the advantage of this protocol is that it readily decreases the
latency during Handover. Another advantage is that it also
reserves New CoA in an optimized and effective way. It
generates New CoA and stores them in a cache [10].
XDAD uses two types of cache:-Active Proxy Cache and
Active Garbage Cache. Active Proxy Cache contains the newly
generated CoA and the Active Proxy Garbage contains the
CoA which MN relinquishes as soon as it leaves the subnet.
Actual Processes Involved In XDAD are:-
After receiving a solicitation message AR tries to reuse the
CoA from the Active Proxy Garbage. If it doesnt get it from
there, it uses store Active Proxy Cache address subsequently
[6]. Access routers then check the uniqueness of the generated
CoA. If it is unique, it sends back the CoA within the same
message as a Router Advertisement message through
Piggybacking process. Otherwise AR drops the CoA and
regenerates it [8].
As soon as the CoA is assigned to a MN, CoA is deleted
from the cache.
Therefore it is observed while reviewing various paper that
there exist following challenges such as for L3 driven approach
use of devices to gauge whether the link is down or up and for
Resource efficient approach the problem is use of various
caches that directly affect the quality of service. Various
modifications for challenges are proposed in subsequent
sections.
IV. PROPOSED MODIFICATION
There are two modifications proposed in this sections these
are described as follows:-
A. Modification 1: In Domain Of L3-Driven Fast Handover
Approach
In the previously defined L3-Driven Fast Handover
approach, there is need to frequently check whether link is
down or not, and if the link is down that means that the present
value of signal strength is below than threshold value. It is
important to note here that L3Driven Handover requires some
of the devices dependent information. So this dependence on
the device for constant measuring and monitoring the signal
strength is not at all required as it will frequently create
problems such as wrong measurement of signal strength in a
real dynamic environment.
To remove this device dependent problem, the internal
entities such as MN and AR should decide about the weak
signal strength (Link). To know whether the link is down or not
MN sends a packet to the AR and starts a timer, the router will
revert the same packet back to the MN. This process is done
when MN is very much connected to the access router. The
running timer will calculate the average time taken by the
packet to traverse to and fro i.e. from MN to AR and from AR
to the MN. Standard measurement of average time will be
calculated by sending the packets at least 5 times. Then the
average of separate to and fro dispatching will be calculated
and compared if these values remain same in all 5 remittances.
Only then it will be set as standard time within the subnet.
Now the MN will keep on moving and sending this
packet. After some time, it will happen that packet would start
taking more time to reach MN, because when MN moves away
beyond the reach of present Access Router, packets will not be
able to reach the MN within set standard limit of the time.
Now to enhance the efficiency of proposed modified
version of MIPv6, it is also seen that if the to and fro dispatch
time is more than standard time and shows increasing trend of
delay for 3 subsequent packets, then the device is so
customized that it understands the feeble signal strength
immediately and starts multicasting solicitation message to the
adjacent router. for e.g. in Fig.3.the to and fro dispatch of three
consecutive dispatch are 2ms,7ms,10ms as a result it will again
watch the another set of 3 to and fro dispatch and if it shows
the increasing trend in these values the Handover will take
place.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
124 | P a g e
www.ijacsa.thesai.org
If the to and fro dispatch time of 3 subsequent packet for
this router is less than the standard time, the MN is not going to
immediately perform the L2 Handover. Whereas in Fig.4 the to
and fro dispatch of three consecutive dispatch are 2ms, 10ms,
3ms. So Handover will not take place. This process finally
eliminates the need of device-based-signal-based-monitoring
system and also overcome the delay on account of it.

Figure.3 Handover is required as To and Fro Dispatch time keeps on
increasing with the time as compared to Standard Measurement


Figure.4 No Handover is required as To and Fro Dispatch time keeps on
fluctuating with the increase in time as compared to Standard
Measurement
B. Modification 2: In Domain of Resource Efficient CoA
Provisioning
In the previously defined Resource-Efficient CoA
Provisioning two caches are used:
a) Active Proxy cache.
b) Active Proxy Garbage.
These two caches are being managed by the Access Routers
which will definitely increase to its overhead. To subside the
overhead, it is recommended to reduce the no. of caches to one.
Now, instead of using two caches, scope is reduced to one
cache and named this cache as Working Cache. This Working
Cache will store and allocate new unique CoA to the MN.
The allocated CoA will be removed from the Working
Cache and as soon as the MN leaves the subnet, again that CoA
is reallocated to the Working Cache for reuse purpose. The
CoA is generated by the help of an algorithm and stored in the
Working Cache.
The algorithm will always generate an unique CoA that will
remove the need of running DAD protocol and multicasting
Neighbor Solicitation messages as every CoA is unique. If the
need of running DAD protocol and multicasting Neighbor
Solicitation is removed, it will definitely reduce the latency.
Now the question arises that if the same algorithm is used
by all the Access Routers to create unique CoA then what will
differentiate the CoA if MN moves from one network to
another and it will again get the same CoA in another Access
Router. To resolve this problem use of the Link-Layer Address
Prefix is suggested as it is very much unique for a subnet. The
Prefix will be attached to the CoA and thats what will make it
Unique within the subnet and outside the subnet.
V. CONCLUSION AND FUTURE SCOPE
This paper focuses on L3 Driven fast handover technique
and Resource efficient CoA provisioning techniques and
proposes some viable comprehensive modification to enhance
the quality of services and resolve the problems of Handover in
MIPv6. Experimental results show that the challenges during
Handover can be drastically improved by applying the
proposed comprehensive modification. Hopefully, if the
proposed radical changes are incorporated in the existing
MIPv6, it will not only bring about large scale viability in the
system but also make it cost-effective, robust, and dynamic all
the more.
The work may be further extended in the domain of care-
of-address provisioning and fast Handover using the techniques
as proposed in this paper.
REFERENCES
[1] Bi-Lynn Ong, Suhaidi Hassan, Interworking of Protocols in IPv6
Mobility Management IEEE International Conference on
Telecommunications and Malaysia International Conference on
Communications, May 2007.
[2] Christian Vogt A Comprehensive and Efficient Handoff Procedure for
IPv6 Mobility Support, Proceedings of the 2006 International
Symposium on a World of Wireless, Mobile and Multimedia Network
(WoWMoMo0),IEEE, September 2009.
[3] D. Johnson, C. Perkins, and J. Arkko, "Mobility support in IPv6", RFC
6275, July 2011.
[4] Deng Ya-ping, Wu Ying-qiu, Research on HMIPv6 Handover Latency
of Improved DAD Policy, IEEE, February 2010.
[5] Gaogang XIE, Ji CHEN, Hongxia ZHENG, Jianhua YANG, Handover
Latency of MIPv6 Implementation in Linux ,IEEE GLOBECOM 2007
proceedings, July 2007.
[6] Longjiang Li, Yuming Mao, Yonggang Li, Resource-efficient Care-of
Address Provisioning for Seamless IPv6 Mobility Support., IEEE, July
2008.
[7] Kazutaka GOGO, Rie SHIBUI, Fumio TERAOKA An L3-Driven
Fast Handover Mechanism in IPv6 Mobility Proceedings of the
International Symposium on Applications and the Internet Workshops
(SAINTW 09), IEEE, January 2009.
[8] P. Sangheon and C. Yanghee, Performance analysis of fast handover in
Mobile IPv6 networks, Lecture Notes in Computer Science (LNCS),
vol.2775, pp.679-691, Springer-Verlag, 2003
[9] Reza Malekian , The Study of Handover in Mobile IPv6 Networks ,
IEEE International Conference on Telecommunications and Malaysia
International Conference on Communication, 2008
[10] R. Koodli , "Fast Handovers for Mobile IPv6," RFC 4068,2005

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
125 | P a g e
www.ijacsa.thesai.org
[11] S. Menezes, An Efficient Handover Scheme Based on Fast Mobile
IPv6, IEEE 802.21Media Independent Handoff Working Group , 2007
[12] Seung Wook Moon , Reducing Handover Delay in Mobile IPv6 by
cooperating with Layer 2 and Layer 3 Handovers, IEEE 802.21 Media
Independent Handoff Working Group, 2007
AUTHORS PROFILE
Mr.Purnendu Shekhar Pandey is a M.Tech Student at
Gautam Buddha University. The author has done his
B.Tech from Noida Institute of Engineering and
Technology, Uttar Pradesh Technical University, Uttar
Pradesh. The author has published and presented various
papers in national and international conferences.


Dr.Neelendra Badal is an Asst. Prof. in the Department
of Computer Science & Engineering at KNIT,Sultanpur
(U.P), INDIA. He received B.E. (1997) from
Bundelkhand Institute of Technology (BIET), Jhansi in
Computer Science & Engineering, M.E. (2001) in
Communication, Control and Networking from Madhav
Institute of Technology and Science (MITS), Gwalior
and PhD (2009) in Computer Science & Engineering from Motilal Nehru
National Institute of Technology (MNNIT), Allahabad. He is Chartered
Engineer (CE) from Institution of Engineers (IE), India. He is a Life Member
of IE, IETE, ISTE and CSI-India. He has published about 30 papers in
International/National Journals, conferences and seminars. His research
interests are evinced at Distributed System, Parallel Processing, GIS, Data
Warehouse & Data mining, Software engineering and Networking.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
126 | P a g e
www.ijacsa.thesai.org
Different Protocols for High Speed Networks

Dr.Srinivasa Rao Angajala
Professor, Mekapati Rajamohan Reddy Institute of Technology & Science,
Udayagiri 524 226, A.P., India.

AbstractNew challenges arise with the presence of various
types of physical links, such as wireless networks, high speed and
satellite in todays ever-changing network. It is clear that the
TCP throughput deteriorates in high-speed networks with large
bandwidth-delay product, and new congestion control
algorithms have been proposed to address such deterioration.
Traditional TCP protocols treat all packet loss as a sign of
congestion. Their inability to recognize non-congestion related
packet loss has significant effects on the communication
efficiency. The proposed protocols such as TCP Adaptive
Westwood, Scalable TCP, HS-TCP, BIC-TCP, FAST-TCP and
H-TCP all have some improvement in functionality over the
traditional TCP protocols. This survey gives a summarization of
all the protocols for high speed networks.
Keywords- TCP protocols; high speed networks; TCP congestion.
I. INTRODUCTION
The Transmission Control Protocol (TCP) congestion
control algorithm is successful in making the functioning of
internet efficiently. However, it shows very poor performance
over the networks with high bandwidth - delay product paths.
The problem starts from the fact that the standard AIMD
(Additive increase and multiplicative decrease) congestion
control algorithm increases the congestion window too slowly.
This would lead to probably long file transfer times. With the
present algorithms, latencies of only a few tens of milliseconds
are quire sufficient to create bandwidth-delay products that
yield poor throughput performance [1]. Therefore innovative
schemes were designed to provide a solution to certain issues.
TCP designed mainly for wired networks is a connection
oriented transport protocol that provides reliable data
communications. The reason for the performance degradation
is that the congestion control mechanism in TCP cannot
distinguish between the packet loss caused by wireless link
error and that caused by network congestion, thus, reacting to
the loss by reducing its congestion window (cwnd). Therefore,
these in appropriate reductions of the cwnd lead to
unnecessary throughput degradation for TCP applications [2].
A solution to the problem has been given by many authors.
The main aim is to increase the rate at which cwnd is
increased and thereby shortens the congestion epoch duration
and also it should be suitable to standard TCP paths with low
bandwidth delay product (BDP). The previous research on
these lines includes the HS-TCP proposal [8], the scalable
TCP [7] and the FAST TCP [11] and many more recent
proposals include BIC-TCP [8] and H-TCP [10]. The rest of
the document is organized as follows. Section 2 offers an
overview of the basic design of TCP protocol and its
mechanisms. Section 3 discusses various protocols for high
speed networks provided in the previous research papers.
Section 4 describes the experimental results from the
related work and Section 5 concludes the review of the study
of various TCP protocols for high speed networks.
II. BASIC DESING OF TCP PROTOCOLS
TCP is a transport layer protocol that provides a reliable
and in-order delivery of data between two hosts. It is a
defensive protocol highly sensitive to network congestion.TCP
issues an acknowledgement packet (ACK) as a response to a
successfully delivered packet to ensure a reliable
communication. The standard TCP congestion control
mechanism is based upon a sliding window, which defines the
number of packets injected in the network. The congestion
window value is updated after the reception of an ACK and
upon the detection of a packet loss. The reception of an ACK
is interpreted by TCP as a signal of available bandwidth, so
that the congestion window value, W, may be increased. The
increase is fast at the beginning of a new connection (the so-
called slow-start phase), where the TCP sender enlarges W by
one segment for each received ACK. In this way, the
congestion window grows exponentially on a Round Trip
Time (RTT) basis. Once a given threshold value is crossed,
the connection enters the congestion avoidance phase, where
the TCP sender gently increases its transmission rate in a
linear fashion over RTT (the congestion window is increased
by 1 W upon an ACK reception).
The decrease phase is triggered by the detection of a packet
loss. A packet loss can be detected either by a timeout
expiration or by the reception of 3 duplicate ACKs
(correspondently, the congestion window is halved). Since the
timeout granularity is fairly large, the first case is interpreted
by TCP as a signal of severe network congestion so that
it reduces its transmission rate to the minimum. In the other
case, since some packets have been correctly received after
the lost one(s), the congestion is assumed to be a transient
phenomenon It is easy to verify that this mechanism is
pretty inefficient in the presence of large W values (or,
equivalently, large BDP values). Indeed, let us consider an
error occurring in the presence of a congestion window value
of W. Since W grows approximately by 1 for RTT, the time it
takes to go back from W/2 to W equals Trec = (W RTT) / 2 .
This will be called as the time to recover from a loss. The fact
that Trec depends linearly on W suggests that the TCPs
AIMD mechanism does not scale well with BDP. This is one
of the main reasons that have motivated the study of
congestion control mechanisms able to perform well in
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
127 | P a g e
www.ijacsa.thesai.org
a large BDP scenario. The packet loss is detected by (1) a
timeout and (2) duplicate acknowledgement (ACK). A timeout
occurs when the TCP sender does not receive any
acknowledgement from the receiver even after a prescribed
time. When timeout occurs, TCP treats the situation as
network congestion and performs slow start. In the second
case, when the TCP sender receives duplicate ACK, it
identifies receiver received out-of order packets. TCP enters
fast retransmit and fast recovery algorithm.
III. VARIANTS OF TCP
A. TCP Adaptive Westwood
TCP Adaptive Westwood (TCP-AW) is a combination of
the best features from both TCP Westwood ABSE (TCP-
ABSE) [4] and TCP Adaptive Reno (TCP-AReno) [5]. TCP-
ABSE's best feature is its eligible rate estimation (ERE)
mechanism, which helps predict imminent congestion.
Essentially, the congestion window is adjusted according to
the network congestion rate, and is not based on the traditional
delay model. TCP-AReno's best feature is its use of packet loss
interval time to adjust the congestion window, instead of using
bandwidth estimation.

This approach improves RTT-fairness even when accurate
bandwidth estimation is not available. For the high-speed
scenario, TCP-AW[6] remarkably improved the aggregate
throughput of the control group. This is due to the delay-based
nature of the protocol. The protocol feels the presence of
incoming TCP standard Reno flows, and then reacts
accordingly to grab the extra bandwidth whenever the
bottleneck is less loaded. TCP Adaptive Westwood shows
good throughput in High Speed networks and performs safely.
TCP-AW obtains a substantial improvement in coexistence
due to its embedded loss discriminator component.
B. HS-TCP
HS-TCP was introduced to achieve high throughput in high
bandwidth-delay product links without requiring
unrealistically low packet loss rates. The HS-TCP modifies the
standard TCPs Additive increase and multiplicative
decrease (AIMD) algorithm to improve its loss discovery time.
Such alteration would only be effective when higher
congestion windows are encountered. This implies that if the
congestion window is smaller than a given threshold, it makes
use of the Standard (TCP Tahoe) AIMD algorithm, otherwise
the High Speed algorithm is used [3]. The Standard AIMD
algorithm upon receipt of ACK and in the event of congestion,
the window is respectively given as:
w = w + 1/w
w = 0.5*w
Parameters a and b for the increase and the decrease of the
AIMD algorithm are fixed at 1 and 0.5 respectively. The
modified HS-TCP algorithm is as follows: upon the receipt of
acknowledgement,
w = w + a(w)/ w
And when congestion is encountered, the window (w) is,
w = w b(w)w
The increase and decrease parameters thus vary depending
on the current value of the congestion window.
C. Scalable TCP
STCP seeks to improve the loss recovery time of Standard
TCP; this idea mirrors that of the HS-TCP [3]. For standard
TCP and HS-TCP connection, the packet loss recovery times
increase (or reduce) in a proportion as the connections
window size and round trip time (RTT) does. In a Scalable
TCP connection, packet loss recovery times are proportional to
connections RTT only. For the STCP [7], the slow start phase
of the standard TCP algorithm is not modified, but its
congestion avoidance phase is modifies thus: for every
acknowledgement received in a RTT, the congestion window
(cwnd) is
cwnd = cwnd + 1/cwnd (for Std TCP)
cwnd = cwnd + 0.01 (for STCP) -------(1)
and when congestion is encountered in a given RTT,
cwnd = cwnd - cwnd * 0.5 (for Std TCP)
cwnd = cwnd cwnd * 0.125 (for STCP) ----(2)

Similarly the evolution of STCPs cwnd is similar to that
of the HS-TCP; its threshold window size and the modified
algorithm in equation (1) and (2) are used only when the size
of the congestion window exceeds threshold window size. The
values of 0.01and 0.125 are suggested for the increase and the
decrease parameters. STCP default value for the threshold
window size is given as 16 segments [7].
D. BIC TCP
BIC-TCP [8] employs a binary search algorithm to
update its congestion window.Briefly a variable w1 is
maintained which holds a value halfway between the values of
cwnd just before and Just after the last loss event. The
window update rule seeks to rapidly increase its window
beyond a specified distance Smax from w1, and update cwnd
more slowly when its value is close tow1.
Multiplicative backoff of cwnd is used on detecting a
packet loss with a backoff factor of 0.8. It also implements
an algorithm whereby upon low utilization detection, it
increases its window more aggressively. This is controlled
with two factors namely, low utility and utility check. In order
to maintain backwards compatibility, it uses the standard TCP
update parameters when cwnd is below the threshold.
E. FAST TCP
FAST TCP [9] is a delay based algorithm. It also
includes rate packing. Rate pacing is a functional change and
is thus it can be viewed as a part of congestion control
algorithm. The FAST TCP flows typically converge quickly
initially, flows may later diverge again to create significant and
sustained unfairness. The main drawback of this is where the
threshold is somewhat higher, owing to the standing queue
created by the delay-based congestion control action used here.
F. H TCP
H-TCP [10] uses the elapsed time since the last
congestion event, rather than cwnd, to indicate path
bandwidth-delay product and the AIMD increase parameter is
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
128 | P a g e
www.ijacsa.thesai.org
varied as a function of . The AIMD increase parameter is
also scaled with path round trip time to mitigate unfairness
between competing flows with different round-trip times. The
AIMD decrease factor is adjusted to improve link utilization
based on an estimate of the queue provisioning on a path.
IV. EXPERIMENTAL RESULTS FROM THE RELATED WORK
The performance analysis relevant to the present study is
derived from many papers. In [7], Kelly presents an
experimental comparison of the aggregate throughput
performance of scalable-TCP and standard TCP. In [11],
aggregate throughput measurements are presented for FAST-
TCP and TCP-Reno. In all of these studies, measurements
focus on aggregate throughput i.e., link utilization. Here
efficiency as a function of queue size is not considered, nor
fairness, friendliness, responsiveness and convergence times.
In [9], throughput and cwnd time histories of FAST-TCP,
HS-TCP, Scalable-TCP, and TCP-Reno are presented for a lab
scale experiment test bed. Aggregate throughput, throughput
Fairness and a number of other measures are presented.
However, results are confined solely to an 800-Mb/s
bottleneck link with a 2000-packet buffer. The impact of link
rate, RTT, queue size, and level of Web traffic on fairness and
responsiveness are not considered nor is the impact of queue
size on efficiency. In [8], NS simulation results are presented
comparing the performance of HS-TCP, Scalable-TCP, BIC-
TCP, and standard TCP.
V. CONCLUSION
This article concludes by presenting a comprehensive
survey of current research on running TCP in high speed
networks from various experimental results evaluating the
performance of Scalable-TCP, HS-TCP, BIC-TCP, FAST
TCP and H-TCP. All the protocols studied are all successful in
improving the link utilization in a relatively static environment
with long-lived flows. And many of them showed poor
responsiveness to changing network conditions. Though there
are various schemas and mechanisms proposed, there is no
single mechanism that can overcome the unreliable nature of
network in a reliable way.
Each and every mechanism has its own advantages and
disadvantages. In short, any mechanism will be effective based
on the factors that are to be taken into consideration. To
conclude this area is not completely explored to its maximum
and still lot more research can be done towards establishing a
basis for the development of new protocols.
REFERENCES
[1] M.Allman, TCP congestion control with appropriate byte counting,
IETF RFC 3465, Feb 2003.
[2] Lakshman, T. V. and Madhow, U.: The performance of TCP/IP for
networks with high bandwidth- delay products and random
loss. IEEE/ACM Transactions on Networking, Vol. 5, 336350, 1997.
[3] Sally Floyd, RFC 3649, Highspeed TCP for large congestion
window, 2003.
[4] R. Wang, M. Valla, M. Sanadidi, and M.Gerla, Using adaptive rate
estimation to provide enhanced and robust transport over
heterogeneous networks, IEEE ICNP, Nov 2002.
[5] Cesar Marcondes, Jerrid Matthews, Robert Chen, A cross comparison
of advanced TCP protocols in High speed and satellite
environments,IEEE, pp 179-185, 2008.
[6] T. Kelly, Scalable TCP: Improving Performances in High speed
Wide Area networks, Computer Commn., Rev., vol.33, no.2, pp. 83-
91, April 2003.
[7] C.Jin,D.X.Wei,S.H.Low,G.Buhrmaster, J. Bunn, D. H. Choe, R. L. A.
Cottrell, J. C. Doyle, W.Feng, O. Martin, H. Newmann, F. Paganini,
S. Ravot and S. Singh, FAST TCP : From theory to experiments,
IEEE Network, vol. 19, no. 1, pp.4-11, 2005.
[8] L.Xu, K.Harfoush and I.Rhee, Binary increase obstacle control for
fast long-distance networks, in Proc. IEEE INFOCOM, HongKong,
2004.
[9] D.J.Leith and R.N.Shorten, H-TCP protocol for high speed long
distance networks,presented at thend Workshop Protocols Fast
Long Distance Networks, Argonne, Canada, 2004.
AUTHORS PROFILE
Dr. ANGAJALA SRINIVASA RAO was born on 1964 and received
Ph.D from University of Allahabad in 2008 and M.S(CSE) from Donetsk
State Technical University, Ukraine in 1992. At present he is working as a
Principal / Professor in CSE Dept. in Mekapati Rajamohan Reddy Institute of
Technology&Science, Udayagiri- 524226, Sri Potti Sreramulu Nellore Dist.,
Andhra Pradesh. He wrote 5 books and published by Vikas Publications,
New Delhi for various Universities in India. He is a member of IEEE, CSI,
ISTE, IACSIT, CSTA, IWA AND IAENG and also attended no. of National
and International Conferences.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
129 | P a g e
www.ijacsa.thesai.org
Wideband Wireless Access Systems Interference
Robustness: Its Effect on Quality of Video Streaming
1
Aderemi A. Atayero,
2
Oleg I. Sheluhin
1,4
Department of Electrical and Information Engineering
Covenant University
Ota, Nigeria
3
Yuri A. Ivanov,
4
Julet O. Iruemi
2,3
Department of Information Security
Moscow Tech. Univ. of Communication and Informatics
Moscow, Russia


AbstractA necessary requirement incumbent on any
information communication system and/or network is the
capacity to transmit information with a predefined degree of
accuracy in the presence of inevitable interference. The
transmission of audio and video streaming services over different
conduits (wireless access systems, Internet, etc.) is becoming ever
more popular. As should be expected, this widespread increase is
accompanied by the attendant new and difficult task of
maintaining the quality of service of streaming video. The use of
very accurate coding techniques for transmissions over wireless
networks alone cannot guarantee a complete eradication of
distortions characteristic of the video signal. A software-
hardware composite system has been developed for investigating
the effect of single bit error and bit packet errors in wideband
wireless access systems on the quality of H.264/AVC standard
bursty video streams. Numerical results of the modeling and
analysis of the effect of interference robustness on quality of
video streaming are presented and discussed.
Keywords-codec; H.264/AVC; polynomial approximation coding;
signal-to-noise ratio; video streaming.
I. INTRODUCTION
One of the most important Quality of Service (QoS)
parameters for wireless networks is the probability of
occurrence of bit and packet errors measured by the Bit Error
Rate (BER) and Packet Error Rates (PER) respectively. Neither
single packet losses nor single bit errors can provide a
comprehensive imitation modeling of fading channels. In
digital systems, errors often occur in packets as a result of
transmission conditions. Specifically, signals are attenuated
during transmission and this consequently leads to grouping
(packetization) of errors. A group of erroneous packets is
essentially a sequence of packets that are either lost in transit or
received with error after transmission over a communication
channel within a given period of time. Burst Error Length
(BEL) is defined as the number of erroneous packets included
in a given group of errors.
II. METHOD
The H.264/AVC standard is a compendium of innovations
and improvements on prior video coding technologies vis--vis
enhancement of coding efficiency and effective usage over a
wide gamut of networks and applications [1]. For a complete
analysis of the impact of errors on resultant signal quality, we
investigate the influence of the conduits (i.e. wireless
transmission mediums) robustness on the perceivable quality
of streaming video standard H.264/AVC using the developed
hardware and software complex (HSC) [2], [3].
Objective and subjective indicators of video quality were
obtained using methods described in [3]. For qualitative
assessment, it is imperative to have the video file data before
transmission over the network (on the transmitting end), and
after reception from the network (at the receiving end).
Data required for the qualitative assessment at the
transmitting end are:
a)the original unencoded video in YUV format,
b)the encoded video in MPEG-4,
c)transmission start time and
d)type of each packet sent to the network.
The following data are required for qualitative assessment
at the receive end:
a)time of reception of each packet from the network,
b)type of each packet received from the network,
c)the encoded video (possibly distorted) in MPEG-4
format, and
d)the decoded video in YUV format for display.
We performed data evaluation by comparing the
transmitted and received files.
Structure of the Hardware-Software Complex (HSC)
Data processing is carried out in the three phases described
below:
First phase: the time taken in sending and receiving each
packet on both sides as well as the packet type are analyzed.
This results in a record of the type of frame and the time
elapsed between transmitted and received packets. The
distorted video file at the receive end is restored using the
originally encoded video file and information about lost
packets. Subsequently, the video is decoded for playback to the
viewer.
Assessment of video quality is done at this stage. Video
quality indicators always require a comparison of the received
(possibly distorted) video frame and the corresponding source
frame. In the case of a total loss of frame in transit, the
necessary frame synchronization before and after transmission
over the network becomes impossible.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
130 | P a g e
www.ijacsa.thesai.org
Second phase: In this phase of data processing, the problem
of quality assessment is resolved based on the analysis of
information about frame loses. Substituting the last relayed
frame for the lost frame restores frame synchronization. This
methodology allows for subsequent frame-wise assessment of
video quality.
Third phase: At this stage, the assessment of the quality of
decoded video is achieved by means of both the restored and
source video files.
Fig. 1 shows a block diagram of the HSC for assessing the
quality of streaming video. The schematic diagram reflects the
interaction between modules in the transmission of the digital
video from a source through the network connection to the
viewer. The HSC modules interact with the network by using
traces containing all the necessary data listed above. Thus, for
proper functioning, the HSC requires two traces, the source
video and the decoder. The data network can be considered
simply as a two-port black-box that introduces delay, packet
loss, and possibly packet rearrangement. The network was
simulated based on the aforementioned assumptions [4] in the
NS2 environment. A detailed description of the functional
modules of the HSC is given in [3].
III. PRESENTATION OF DATA AND SIMULATION
PREPARATION
Video encoding begins with color space conversion from
RGB to YUV also known as Y, Cr, Cb i.e. one luma and two
chroma components [5]. It is common knowledge that there is a
significant correlation of color components in any typical
image of the RGB format, which makes it an obviously
redundant format in terms of compression. The standard
television uses a different representation of images, which also
employs three components of the signal, but these components
are uncorrelated (i.e. void of inter-componential redundancy).
R, G and B components are converted to luminance Y
component and two color difference components U and V of
the YUV format. Since most information is stored in the
luminance component, little information is lost if a thinning of
the U and V components is done.
Standard test videos in the YUV format may be used as
initial test video sequences. However, these videos have limited
playback time and hence do not allow for the assessment of
change in quality under prolonged video broadcast. Similarly, a
vast amount of experimental data cannot be obtained from
them. It is for these reasons that we recorded our own 30-
minute video in YUV format (send.YUV) with a resolution of
640x480 pixels and frame rate of 25fps using special software.
The first step is to encode the source video to H.264 format
(video stream file). This is done by the video codec (a device
used for encoding and decoding video signals). Video codecs
are usually characterized by a) channel throughput, b) decoded
video distortion rate, c) startup latency, d) end-to-end delay,
computational complexity, and e) memory capacity. A good
codec is one capable of providing the necessary trade-off vis--
vis these characteristics [5].
In the next step, the encoded video stream is packaged in an
MP4-container for onward transport over the network using the
User Datagram Protocol (UDP). The result of encoding the
original video is an MP4-file. Since it is necessary to evaluate
the quality of video transmitted over the network, the need
arises to create a spare decoded YUV file from the newly
created MPEG layer-4 file, which serves as the control in
evaluating the quality of video transmitted over the network,
excluding the impact of the codec. It is thus possible to
estimate the influence of a wireless network on the received
visual video quality, while excluding encoding and decoding
losses.
For simulation purposes, it is necessary to create a video
trace file that contains the following information: frame
number, frame type, frame size, and the number of segments in
which the frame is divided into packets. This video trace serves
as the input to the simulator network, where the sending and
reception of video data occurs. As a result of video
transmission over the network, it is necessary to obtain
transmission trace files and reception trace files, which contain
the following packet data: the transmission/reception time, a
unique identifier and trace file size. These two traces are used
to determine lost packets in the network. In the end, we obtain
files of the sent and received packets containing detailed
information about the time of sending from the transmitter and
the time of reception by the receiver.
IV. MODELING AND SIMULATION OF TRANSMISSION OVER A
WIRELESS NETWORK
The HSC allows for the simulation of the main types of
errors encountered when transmitting video data over wireless
networks. The two types of simulation required are as listed
below:
A. Bit error simulation
Simulated transmission over a wireless channel model with
Additive White Gaussian Noise (AWGN) is conducted. In the
process of simulation, certain bits in the sequence are distorted
(i.e. inverted) with a given probability. The probability value
used is defined by the Bit Error Rate (BER).
B. Packet error simulation.
The UDP packets can be manually deleted from the
received trace file. This allows for the observation of codec
functionality and analysis of change in visual quality in cases
of packet loss. At the same time, both the received and
undistorted files can be obtained during transmission over an
"ideal" channel with unlimited bandwidth and no delay, with
subsequent removal of some packets.


Figure 1. Block diagram of HSC.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
131 | P a g e
www.ijacsa.thesai.org
V. CALCULATION OF LOSSES AND ESTIMATION OF
OBTAINED VIDEO QUALITY
Calculation of losses given the availability of unique id of
the package is quite easily achieved. With the aid of the video
trace, each packet is assigned a type. Each package of the
assigned type that is not included in the received trace is
deemed lost. Loss of frame is calculated for any (and all)
frame(s) with a lost packet. If the first packet in a frame is lost,
then the whole frame is considered lost since the video decoder
cannot decode a frame, which is missing the first part.
Assessment of received traces is done by the module for trace
assessment (see Fig. 1). The recovered file must be decoded in
YUV format.
There are two major methods of estimating the quality of
digital video, namely, the subjective and objective methods:
Subjective quality assessment is always based on viewer
impression. It is extremely costly, very time consuming and
requires specialized equipment. Traditionally, subjective video
quality is determined by expert assessment and calculation of
the average Mean opinion Score (MOS), which is assigned a
value from 1 to 5 (ITU scale) [6], [7], where 1 and 5 represent
worst and best received video quality respectively.
Objective video quality assessment is usually done by
measuring the average luminance peak Signal-to-Noise Ratio
(PSNR). The PSNR is a traditional metric, which allows for the
comparison of any two images [8]. The PSNR module
evaluates the objective quality of received video stream in
polynomial approximation coding (PAC). The end result is the
values of PSNR calculated for the original and distorted image
(as shown in Fig. 2). MOS values are calculated from the
PSNR indicator.
VI. RESULTS, ANLYSIS AND DISCUSSION
In order to study the effect of transmission errors on the
resulting video quality, the transmission of a 30-minute video
over a wireless network with random packet errors in the
channel was simulated. Characteristics of sequences used are
listed in Table I.
For modeling purposes, the encoded video stream was
split into RTP/UDP-packets using the hardware-software
tool reported in [3]. Bit error simulation for transmission
over a wireless channel was done using an AWGN error
generator contained within the PAC structure. Simulation of
packet errors during transmission over a wireless channel
was done by deleting packets from the received trace file [3].
This allowed us to explore and analyze the change in visual
quality during loss of packets. The received and undistorted
trace files were obtained for transmission over an "ideal"
channel with unlimited bandwidth and no delay in using the
NS2 software environment [4], followed by a random removal
of packets, according to PER and BER parameters. Quality
assessment was carried out using PSNR and MOS indicators,
calculated by using hardware and software tools [4]. The
standard deviation of the quality of the average PSNR values
was calculated using equation (1) [9].

()
C. Effect of bit error
Fig. 2 shows the effect of BER on the quality of video
streaming.
Analysis of the results of streaming video over the
simulated wireless network with different values of BER
revealed the following:
i) Simulating a wireless channel using AWGN model,
and additive, bit errors with a value of BER 3x10
-5

does not affect the quality of the video. However, when
BER 4x10
-3
packet loss in the network reaches its
maximum value of 99,9%.
ii) Objectively, excellent quality of video transmission
over a channel can be guaranteed for all bit error
probabilities less than 1x10
-4
, good quality is in the
range of 1x10
-4
to 4x10
-4
, satisfactory quality is in the
range of 4x10
-4
to 8x10
-4
, poor quality is in the range of
8x10
-4
to 1x10
-3
, while very bad quality is for any BER
> 1x10
-3
.
iii) The histograms of the distribution of PSNR values
during simulation and broadcast over a real network in
general are of a twin-peak form. One of the peaks
characterizes the PSNR value of error-free video
stream (the decoder is able to correct bit errors when
they are relatively few in the frame). The second peak
characterizes PSNR degradation due to the large
number of corrupted video frames in fading moments
(the decoder is unable to fix large numbers of bit
errors). As the number of errors increases, this
maximum increases commensurately with a decrease in
the second. During transmission, depending on error
TABLE 1. CHARACTERISTICS OF ENCODED VIDEO
Format MPEG-4 Part14 (MP4)
Codec H.264
Bit rate Constant @ 1150 kbps
Frame frequency 25 fps
Resolution 640 x 480 pixels
GOP type IBBPBB

TABLE 2. RELATIONSHIP BETWEEN QUALITY INDICATORS AND BER
PSNR
[dB]
MOS
[%]
BER ITU
Quality Scale
Picture
Degradation
> 37 81100 < 1X10
-4

5
EXCELLENT
NOTICEABLE
3137 6180 1X10
-4
4X10
-4

4
GOOD
NOTICEABLE,
BUT NOT
IRRITATING
2531 4160 4X10
-4
8X10
-4

3
SATISFACTORY
SLIGHTLY
IRRITATING
2035 2140

8X10
-4
1X10
-3

2
POOR
IRRITATING
< 20 020 > 1X10
-3
1-VERY POOR VERY IRRITATING
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
132 | P a g e
www.ijacsa.thesai.org
iv) level, values of either of the maxima increases. In cases

a)

b)

c)

d)
Figure 3. Values of video sequence quality indicators for different values of
wireless channel BER: ) PSNR value distribution histogram; b) PSNR value
distribution histogram for certain values of BER; c) Quality deviation from
average PSNR value; d) quality gradation for MOS values.

a)
b)

c)
d)
Figure 2. Values of video sequence quality indicators for different values of
wireless channel PER: ) PSNR value distribution histogram; b) PSNR value
distribution histogram for certain values of PER; c) Quality deviation from
average PSNR value; d) quality gradation for MOS values.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
133 | P a g e
www.ijacsa.thesai.org
when errors in the communication channel are
negligible, the PSNR distribution has only one
maximum.
Empirical values of BER transitions from an acceptable
quality to the poor, according to the relationship between
PSNR and MOS [6], are presented in Table II.
However, the AWGN model does not allow for adequately
simulation of a fading channel. Typically, errors are often long
term, since high probability of bit loss occurs in specific
periods of transmission, e.g. during poor propagation.
Attenuation of the transmitted signal results in packetizing
(grouping) of errors. Another cause of error grouping can be
physical defects of, and failures inherent in the information
storage system. When using VLC, bit error occurrence results
in group errors or packetization of errors.
B. Effect of packet error
Fig. 3 shows the effect of PER indicator on streaming video
quality.
The range of values of PER, within which the resulting
quality is maximal (i.e. almost equal to the original) and
minimal are indicated. It is shown that with PER 1x10-4 error
does not affect the resultant video quality and can be easily
eliminated with decoders and existing methods of error
correction. A further change in the quality has a stepwise
nature and decreases with increasing PER.
Empirical values of PER transitions from an acceptable
quality to the poor, according to the relationship between
PSNR and MOS, are presented in Table III.
Analyzing the results of streaming video over a simulated
wireless network with a given probability of packet loss, we
safely conclude that:
a) A PER value of 1x10-4 in simulation of a wireless
network does not affect the video quality. When PER
1x10-3 impact of errors on video quality is not
noticeable and does not irritate during viewing
experience. When PER 0.1, packet loss in the network
has the worst effect on visual quality.
b) Objectively, excellent quality of video transmission
over a channel can be guaranteed for all packet error
probabilities less than 1x10-3, good quality is in the
TABLE 3. RELATIONSHIP BETWEEN QUALITY INDICATORS AND PER
PSNR
[dB]
MOS
[%]
PER ITU
Quality Scale
Picture
Degradation
> 37 81100 < 1X10
-4

5
EXCELLENT
NOTICEABLE
3137 6180 1X10
-3
3X10
-3

4
GOOD
NOTICEABLE,
BUT NOT
IRRITATING
2531 4160 3X10
-3
1X10
-2

3
SATISFACTORY
SLIGHTLY
IRRITATING
2035 2140

1X10
-2
5X10
-2

2
POOR
IRRITATING
< 20 020 > 5X10
-2
1VERY POOR VERY IRRITATING


a)

b)

c)

d)
Figure 4. Values of video sequence quality indicators for PER=1x10-3 and
varying values of wireless channel BEL: ) PSNR value distribution histogram;
b) PSNR value and RTP/UDP packet distribution (black spaces correspond to
lost packets) for certain values of BEL; c) Quality deviation from average
PSNR value; d) quality gradation for MOS values.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
134 | P a g e
www.ijacsa.thesai.org
range of 1x10-3 to 3x10-3, satisfactory quality is in the
range of 3x10-3 to 1x10-2, poor quality is in the range
of 1x10-2 to 5x10-2, while very bad quality is for any
PER > 5x10-2.
Histograms of the distribution of values of PSNR when
PER 6x10-4, in general, have a bimodal shape. One of the
peaks characterizes the value of PSNR of video stream
distorted due to packet loss. The second maximum
characterizes deterioration in the PSNR of dependent frames.
As the number of errors increases, one of the peaks increases
due to a decrease in the other.
C. Effect of length of error groups
To study the effect of the length of error groups on resultant
quality, the simulation of a 30-minute video transfer over a
wireless network for the values of PER of 1x10
-3
to 5x10
-2
is
repeated, since a visual change in video quality is observed at
this range. The simulation of groups of error packets during
transmission over a wireless channel was done by means of
random deletion of packet groups from the receive trace file
with a given BEL. For this particular example BEL=100
implies that the total random number of consecutively deleted
packets does not exceed 100. The total sum of erroneous
(deleted) packets in the video sequence for the whole
experiment given PER=const. remained the same, irrespective
of the value of BEL. Fig. 4 shows the effect of BEL on the
quality of streaming video for PER=1x10
-3
.

Analyzing the results of streaming video over the simulated
wireless network with a given grouping of erroneous packets,
we can draw the following conclusions:
v) For PER 1x10
-3
the effect of single packet errors on
quality is insignificant and does not irritate the viewing
experience.
vi) Histograms of the distribution of values of PSNR have
two maxima. One of the peaks characterizes the value
of PSNR of video frames distorted due to the loss of
packets. The second maximum characterizes the
deterioration of PSNR of dependent frames. With
increasing quantities BEL is one of the peaks decreases
as the number of dependent frames are also reduced,
whereas the second peak remains unchanged. This is
explained by the fact that the single scattered
throughout the video sequence error number of
distorted frames is large due to error propagation to
dependent frames. An increase in the BEL value leads
to a decrease in one of the maxima, since the number of
dependent frames also decreases, while the second
maximum remains the same. This is due to the fact that
under singular errors spread across the whole video
sequence, the number of distorted frames is large
because of the distribution of errors on dependent
frames.
vii) Increasing the length of the error groups leads to an
increase in the average quality of the video sequence.
viii) Effect of error groups on the quality is more powerful
because of the local concentration of errors. However,
the average quality of the video sequence increases
with increase in the length of the grouping for a given
value of probability of occurrence of packet errors.
For BEL 60 the average quality is almost identical to the
original video.
D. Relationship between PER and BEL
The average quality of the experimental video sequence for
different values of PER and BEL is shown in Fig. 5.
In assessing the impact of erroneous packets received on
quality, it is necessary to analyze not only the likelihood of
occurrence of errors, but also their structure and length of their
grouping. Additionally, the following conclusions can be
drawn:
i) Increasing the length of error groupings leads to an
increase in the average quality of the video sequence.
This is due to the deterioration of a small section of
video, where error groups are concentrated, whereas in
the case of single bit errors deterioration in the quality
of video may be observed across the whole sequence;
ii) When the length of erroneous packets is BEL 6 the
change in quality is minor and identical to the influence
of single packet errors (BEL=1);
iii) When BEL 60 the average quality is almost identical
to the original (PSNR <90 dB). It is logical to assume
that the value of BEL in the longer video sequences,
with the same average quality may have a higher value;
iv) The highest dynamics of change in PSNR=60 dB is
observed in two cases: a) for a fixed PER = 1x10
-3
and
the variable values of BEL; and b) at BEL 80 and the
varying values of the PER. In other cases, the dynamics
is not essential and minimal in the absence of clustering
of errors (BEL = 1)
v) With increasing PER, the effect of BEL on quality
decreases due to increase in denseness of single errors.
vi) Analysis of the results of PER and BEL shows that for
effective assessment of the impact of transmission
errors on resultant quality it is necessary to analyze not
only the likelihood of errors, but also their structure and
length of their grouping. The most realistic and


Figure 5. Estimate of video sequence quality indicator for different values
of wireless channel PER and BEL.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
135 | P a g e
www.ijacsa.thesai.org
accurate method of modeling statistical errors in
communication channels is the use of probability data
obtained from real networks.
At BER values 3x10
-5
bit errors do not affect the quality
of the received video and are easily eliminated by well-known
methods of error correction implemented in WiMAX. When
BER 4x10
-3
packet loss in the network reaches its maximum
value and leads to an unacceptable quality of the received
video. Ensuring objectively excellent quality of video sequence
over a channel can be done for probabilities of bit error rate
less than 1x10
-4
; good quality in the range of 1x10
-4
4x10
-4
;
satisfactory quality in the range of 4x10
-4
8x10
-4
; poor quality
in the range of 8x10
-4
1x10
-3
and very bad at BER 1x10
-3
.
The use of H.264/AVC video in wireless access systems
with VLC codec of variable length leads to a disruption of the
synchronization of decoded video sequences and the
occurrence of additional grouping of errors, whose impact on
the quality for video decoding is much stronger than that of the
bit error, since it leads to loss of large segments of the
information. It is shown that the quality of the video affects not
only the probability of error, but also the structure and length of
errors. Analysis of individual errors showed that at PER
1x10
-4
packet errors do not affect the quality of the received
video and are easily eliminated by well-known methods of
error correction deployed in wireless networks. When PER
1x10
-3
, the effect of errors on quality is not noticeable and does
not irritate the viewing experience. For values of PER 0.1
packet loss in the network leads to an unacceptable quality of
the received video. Ensuring objectively excellent quality of
video sequence over a channel can be done for probabilities of
bit error rate less than 1x10
-3
; good quality in the range of
1x10
-3
3x10
-3
; satisfactory quality in the range of 3x10
-3

1x10
-2
; poor quality in the range of 1x10
-2
5x10
-2
and very
bad at BER 5x10
-2
.
To assess the impact on quality of video playback under
error grouping conditions of error groups BEL, the use of a
regular (deterministic) model is proposed. It is shown that the
effect of errors on the average quality is stronger due to local
concentration of errors. The average quality of the video
sequence at the same time increases with increase in the length
of the grouping for a given value of probability of occurrence
of packet errors. For groupings of length BEL 60, average
quality is almost identical to that of the source video. With
increasing PER, the effect of BEL on quality decreases due to
increase in the denseness of single errors. Increase in the BEL
leads to an increase in the average quality of the video
sequence irrespective of the PER value. The highest dynamics
of change in PSNR is observed for fixed PER =1x10
-3
and the
variable values of BEL; at BEL 80 as well as for the
changing values of the PER. In other cases, the dynamics is not
significant and is minimal in the absence of clustering of errors.
To assess the quality of video under packetization of errors
under real conditions, it is necessary to investigate the actual
distribution of packetization of errors in the communication
channel.
VII. CONCLUSION
We have presented in this paper the results of investigating
the effect of single bit error and bit packet errors on the quality
of H.264/AVC standard bursty video streams in wideband
wireless access systems, using a software-hardware composite
system that was developed specifically for this purpose. From
analyses of simulation results, we safely conclude as discussed
in sections VI A through D for BER, PER, BEL, and
relationship between PER and BEL respectively.
REFERENCES
[1] T. Wiegand, G. J. Sullivan, G. Bjntegaard, and A. Luthra, Overview
of the H.264/AVC Video Coding Standard, IEEE Transactions on
Circuits and Systems for Video Technology, Vol. 13, No. 7, pp. 560-
576, July 2003.
[2] O. I. Sheluhin, Y. A. Ivanov, Assessment of the quality of streaming
video in telecommunication networks using software-hardware
methods, Electro-technical and Information Complexes and Systems,
vol. 5, No.4, pp. 48-56, 2009.
[3] Y.A. Ivanov, V.S. Pryanikov, Imitation Modeling of Wireless
Networks using Hardware-Software Complex for the Assessment of
Streaming Video Quality, Chuvash University Digest, vol.1, No.1,
pp.35-48, 2010.
[4] NS-2 documentation, available at: http://www.isi.edu/nsnam/ns/ns-
documentation.html, accessed 29.06.2010.
[5] G. J. Sullivan, T. Wiegand, Video Compression-From Concepts to the
H.264/AVC Standard, Proceedings of the IEEE, Vol. 93, No. 1, pp.18-
31, Jan. 2005.
[6] ITU P.800: Methods for subjective determination of transmission
quality, available at: http://www.itu.int/rec/T-REC-P.800-199608-I/en.
[7] Atayero A.A., Estimation of the Quality of Digitally Transmitted
Analogue Signals over Corporate VSAT Networks, Ph.D Thesis
(unpublished), Moscow, Jan. 2000.
[8] J. Ostermann, et al., Video coding with H.264/AVC: Tools,
Performance, and Complexity, IEEE Circuits and Systems Magazine,
pp. 7 28, Q1. 2004.
[9] J. J. Lemmon, Wireless link statistical bit error model, NTIA Report.
02-394, U.S. Department of Commerce, June 2002.
AUTHORS PROFILE
Aderemi A. Atayero graduated from the Moscow Institute of Technology
(MIT) with a B.Sc. Degree in Radio Engineering and M.Sc. Degree in
Satellite Communication Systems in 1992 and 1994 respectively. He earned a
Ph.D in Communications/Signal Processing from Moscow State Technical
University of Civil Aviation, Russia in 2000. He is a two-time Head,
Department of electrical and Information Engineering, Covenant University,
Nigeria. He was the coordinator of the School of Engineering of the same
University.
Dr. Atayero is a member of a number of professional associations including:
the Institute of Electrical and Electronic Engineers, IEEE, the International
Association of Engineers, IAENG, among others. He is a registered engineer
with the Council for the Regulation of Engineering in Nigeria, COREN, as
well as a professional member of the International Whos Who Historical
Society (IWWHS). He is widely published in International peer-reviewed
scientific journals, proceedings, and edited books. He is on the editorial board
of a number of highly reputed technical and scientific publications. He is a
recipient of the 2009/10 Ford Foundation Teaching Innovation Award. His
current research interests are in Radio and Telecommunication Systems and
Devices; Signal Processing and Converged Multi-service Networks.
Oleg I. Sheluhin was born in Moscow, Russia in 1952. He obtained an M.Sc.
Degree in Radio Engineering1974 from the Moscow Institute of Transport
Engineers (MITE). He later enrolled at Lomonosov State University
(Moscow) and graduated in 1979 with a Second M.Sc. in Mathematics. He
received a PhD at MITE in 1979 in Radio Engineering and earned a D.Sc.
Degree in Telecommunication Systems and Devices from Kharkov Aviation
Institute in 1990. The title of his PhD thesis was Investigation of interfering
factors influence on the structure and activity of noise short-range radar.
He is currently Head, Department of Information Security, Moscow Technical
University of Communication and Informatics, Russia. He was the Head,
Radio Engineering and Radio Systems Department of Moscow State
Technical University of Service (MSTUS).
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
136 | P a g e
www.ijacsa.thesai.org
Prof. Sheluhin is a member of the International Academy of Sciences of
Higher Educational Institutions. He has published over 15 scientific books and
textbooks for universities and has more than 250 scientific papers. He is the
Chief Editor of the scientific journal Electrical and Informational Complexes
and Systems and a member of Editorial Boards of various scientific journals.
In 2004 the Russian President awarded him the honorary title Honored
Scientific Worker of the Russian Federation.
Yury A. Ivanov was born in Moscow, Russia in 1985. He obtained an M.Sc.
degree in Systems, network and devices in telecommunications from Chuvash
State University in 2007. He obtained a Ph.D in Telecommunication Networks
and Systems in 2011 from Moscow State University of Communication and
Informatics. His dissertation topic was "The impact of errors in channels of
broadband wireless access systems on the quality of streaming H.264/AVC
video". Dr. Ivanov has published over 35 scientific papers and his current
research interests include Radio and Telecommunications Systems and
Devices: transmission of multimedia data across telecommunication networks,
assessment of the quality of video sequences.
Juliet O. Iruemi was born in Kaduna, Nigeria in 1984. She obtained a B.Eng.
degree in Information and Communication Technology from Covenant
University in 2008. She is currently on her M.Eng. Programme in Information
and Communication Technology in Covenant University. Her thesis topic is
Hybrid WLAN Access Point (AP) based on Software Defined Radio (SDR).
Her current research interests include Radio and Telecommunication Systems:
Wireless access transmission over broadband network.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
137 | P a g e
www.ijacsa.thesai.org
Survey on Impact of Software Metrics on Software
Quality
Mrinal Singh Rawat
1

Department of Computer Science
MGMs COET,
Noida, India
Arpita Mittal
2

Department of Computer Science
IIMT,
Merrut, India
Sanjay Kumar Dubey
3

Department of Computer Science
Amity University,
Noida, India


AbstractSoftware metrics provide a quantitative basis for
planning and predicting software development processes.
Therefore the quality of software can be controlled and improved
easily. Quality in fact aids higher productivity, which has brought
software metrics to the forefront. This research paper focuses on
different views on software quality. Moreover, many metrics and
models have been developed; promoted and utilized resulting in
remarkable successes. This paper examines the realm of software
engineering to see why software metrics are needed and also
reviews their contribution to software quality and reliability.
Results can be improved further as we acquire additional
experience with variety of software metrics. These experiences
can yield tremendous benefits and betterment in quality and
reliability.
Keywords- Software metrics; Software quality; Software reliability;
Lines of code; Function points; object oriented metrics.
I. INTRODUCTION
Software metrics are valuable entity in the entire software
life cycle. They provide measurement for the software
development, including software requirement documents,
designs, programs and tests. Rapid developments of large
scaled software have evolved complexity that makes the
quality difficult to control. The successful execution of the
control over software quality requires software metrics. The
concepts of software metrics are coherent, understandable and
well established, and many metrics related to the product
quality have been developed and used.
It is essential to introduce definition of software metrics.
Software metrics provides measurement of the software
product and the process of software production. In this paper,
the software product should be seen as an abstract object that
begins from an initial statement of requirement to a finished
software product, including source and object code and the
several forms of documentation exhibited during the various
stages of its development.
Good metrics should enable the development of models that
are efficient of predicting process or product spectrum. Thus,
optimal metrics should be: [1]
Simple, precisely definableso that it is clear how
the metric can be evaluated;
Objective, to the greatest extent possible;
Easily obtainable (i.e., at reasonable cost);
Validthe metric should measure what it is
intended to measure; and
Robustrelatively insensitive to (intuitively)
insignificant changes in the process or product.
II. OVERVIEW OF SOFTWARE METRICS
A. Classification of Software Metrics
There are three types of software metrics: process metrics,
project metrics and product metrics. [3]
1) Process Metrics:
Process metrics highlights the process of software
development. It mainly aims at process duration, cost incurred
and type of methodology used. Process metrics can be used to
augment software development and maintenance. Examples
include the efficacy of defect removal during development, the
patterning of testing defect arrival, and the response time of the
fix process.
2) Project Metrics:
Project metrics are used to monitor project situation and
status. Project metrics preclude the problems or potential risks
by calibrating the project and help to optimize the software
development plan. Project metrics describe the project
characteristics and execution. Examples include the number of
software developers, the staffing pattern over the life cycle of
the software, cost, schedule, and productivity. [4]
3) Product Metrics:
Product metrics describe the attributes of the software
product at any phase of its development. Product metrics may
measure the size of the program, complexity of the software
design, performance, portability, maintainability, and product
scale. Product metrics are used to presume and invent the
quality of the product. Product metrics are used to measure the
medium or the final product.
We can find more efficient ways of improving software
project, product and process management.
B. Mathematical Analysis
A metric has a very explicit meaning in mathematical
analysis .It is a rule used to determine distance between two
points. More formally, a metric is a function d defined on
pairs of objects p and q such that d (p, q) expresses the distance
between p and q. Such metrics must satisfy certain properties:
[11]
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
138 | P a g e
www.ijacsa.thesai.org
d (p,p) = 0 for all p : that is, the distance from point p to
itself is zero;
d (p, q) = m (q, p) for all p and q: that is, the distance from
p to q is similar to the distance from q to p;
d (p, r) d (p, q)+d (q, r) for all p, q and r: that is, the
distance from p to r is no larger than the distance measured by
stopping through an intermediate point.
A prediction system comprise of a mathematical model
along with a set of prediction processes for determining
unknown parameters and depicting the results. The model
should not be complicated for use. Suppose we want to predict
the number of pages, P that will print out as a source code
program, so that we can bring sufficient paper or calculate the
time the program will take for printing. We can use a simple
model,
P = x/a (1)
Where x is a variable, acts as a measure i.e. length of source
code program in LOC (line of code), and a is a constant that
represents the average number of lines per page. There are
number of models to determine effort estimation; from analogy
based estimation to parametric models. A generic model can be
used to estimate effort predication.
E = aS
b
(2)
Where a and b are constants. E is effort in person-months. S
is the size of source code in Line of code.
III. IMPORTANCE OF SOFTWARE QUALITY
In recent times the importance of software quality has come
to light when random errors on a say a telephone bill, or on a
bank statement were randomly attributed to a bug in the
computer code or using the ignorant adage of the computer
does things without making an effort to undermine the cause
of the problem or even separating it by hardware or software.
The problem arises when computer errors creep into highly
critical aspects of our lives involving situations where a small
error can lead to a cataclysmic chain of events. Bearing all this
in mind, the importance of enforcing software quality in
computer practices has become highly important. Seeing the
penetration of computer code into everyday objects like
washing machines, automobiles, refrigerators, toys and even
things like the mars rover, any system be it a large one or a
small system running embedded IC technology, ensuring the
highest levels of software quality is paramount.
However, that brings us to the next logical question, how
do we assess the quality of something intangible like software
quality? This is a highly subjective question whose answer will
vary according to the situation. For example, a small word
processing error in a students assignment will not be a huge
issue. But a slight code error in a space shuttles guidance
computer might be mission critical and endanger human lives.
Hence in terms of software quality, it is imperative that we
understand that its impossible to have a boilerplate definition
or meaning of software quality. The definition will differ
according to factors like quality of products and business. Also
crucial is the proper setting of goals as well as proactive
monitoring of quality factors and goals making sure that that
the goals set are resolved and completed in the given timelines
and specifics. Views on Software Quality
Software quality, as stated earlier, depends on a number of
factors. Also as theorized by David & Garwin, quality is a
complex as well as multifaceted concept, which can be viewed
according to different points of view as follows
1) User View
The user viewpoint of software quality tends to be a lot
more concrete and can be highly subjective depending upon the
user. This view evaluates the software product against the
users needs. In certain types of software products like
reliability performance modelling and operational products, the
user is monitored according to how they use the product.
2) Manufacturing View
This viewpoint looks at the production aspect of the
software product. It basically stresses on enforcing building the
product without any defects and getting it right the first time
rather than subsequently making a defective product and
spending valuable project time and more importantly costs
ironing out the defects or bugs at a later stage. Being process
based, this viewpoint focuses on conformity to the process,
which will eventually lead to a better product.
Models such as ISO 9001 as well as the Capability Maturity
Model do encompass this viewpoint that stress on following the
process as opposed to going by specification. However, that
being said, the theory that following the best and high quality
manufacturing process will automatically lead to a better
product cannot be inferred. The critics viewpoint is that
following an optimized and high quality product manufacturing
method can also lead to the standardization of a product
making it more of a commodity rather than a standout product.
That being said, there have been a lot of industry example
where the philosophy of doing it right the first time been
profitable. Also both the models CMM as well as the ISO,
indirectly do imply by following the principle of
Documenting what you do and doing what you say helps in
improving the product quality.
3) Product View
The product viewpoint looks at the internal features as well
as the characteristics of the product. The idea behind this
viewpoint is that in case a product is sound in terms of the
features and functionality it offers, and then it will also be
favourable when viewed from a user viewpoint in terms of
software quality. The idea is that controlling the internal
product quality indicators will influence positively the external
product behaviour (user quality) There are models trying to
link both the views of software quality but more work is
needed is this area.
4) Value based view
The value-based view becomes important when there are
lots of contrasting views, which are held by different
departments in an organization. For example, the marketing
department generally take a user view and the technical
department will generally take a product-based view. Though
initially these contrasting viewpoints help to develop a 360-
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
139 | P a g e
www.ijacsa.thesai.org
degree product with the different viewpoints complementing
each other, the later stages of the software product development
might have issues
The issues arise when there might be a set of change
proposed to a certain view that can end up throwing a conflict
in the other view. For example, say the marketing department
(user view) want changes to the user interface that are not
technically feasible (product view).
This is where a value-based view comes into play helping
resolve such conflicts so that the software product is not
delayed indefinitely. The value-based viewpoint looks the
conflict with a cost to benefit angle. It help in resolving such
issues by looking at the issue in relation to terms like costing,
constraints, resources, time. Using this viewpoint, its possible
to resolve interview conflicts helping to keep the software
product on track and within initial cost and timeline estimates.
IV. CASE STUDY ON SOFTWARE QUALITY
The Boeing 777 project Boeing with its 777 airplane
project was a giant leap forward in the direction of Software
quality and is compelling case and point in the importance in
reinforcing strong software quality management. With almost
2.5 million lines of code written for the new jetliners state of
the art avionics and other on board software, it was super
critical to ensure best software quality practices and
implementation. Complications like an extensive network of
third party suppliers who would supply crucial components for
the 777 made it a large challenge to ensure that deadlines are
met without a compromise on software quality as a whole. [15]
Measures taken by Boeing interestingly, at the beginning
of the 777 project, since there was extensive vendor
fragmentation, each vendor was using different measures and
metrics to keep in track of software quality and measure the
status of the work. As a result, this soon snowballed into a
situation where due to non-standard practices being followed, it
was extremely hard to understand the progress of the project as
a whole. Therefore, around the 777 projects midway point,
Boeing implemented measures, which called for uniformity in
reporting as well as monitoring all variables related to the
project status and software quality. A uniform use of metrics
like came into effect which made the suppliers report around
the simple metrics like test definition, resource utilization, test
execution as well as detailed plans for the software coding and
design.
As a result, since the reporting was uniform as well as the
enforcement of these metrics was universal for the of Boeings
vendors, each vendor was now reporting on a bi weekly basis
as which now contained information about completed code,
testing as well as design. This not only lowered the effort on
Boeings part in consolidating the fragmented data (as was
happening previously) but also allowed Boeing to adjust its
own plans in sync with the vendors estimates and hence keep
the project on schedule.
Key Takeaways Boeing realized early enough of the
importance of enforcing a uniform set of metrics. Also vital
learnings from Boeings experience is that done properly,
enforcing software quality in a project ensures that program
risk points can be identified early which would allow a
reasonable time to apply corrective measures without delaying
a project indefinitely. Additional key points are the
implementation of metrics allowed each project point to be
having a check and balance so that the project flows smoothly
without any major roadblocks. A good consequence of the
metrics implementation was the streamlining and the regularity
of communications between Boeing and its vendors, which was
touted as being of equal importance to the metrics as well.
Clear goals, milestones and constant monitoring of the key
metrics around software design coding and testing made sure
the 777 project was a success.
V. COMPARISON OF SOFTWARE METRICS- STRENGTHS AND
WEAKNESSES
The software industry does not have standard metric and
measurement practices. Most of the software metric has
multiple definitions and ambiguous rules for counting. There
are also important subject issues that do not have specific
metrics, such as quantifying the volume or quality levels of
databases, web sites and data warehouses. There is a lack of
strong empirical data on software costs, schedules, effort,
quality, and other tangible elements, which results in metric
problems. [12]
A. Source Code Metrics
Source lines of code or SLOC was the first metric
developed for quantifying the outcome of a software project.
The divergent lines of code or LOC has similar meaning and
is also widely acceptable. Lines of code could be defined
either:
A physical line of code.
A logical line of code.
Physical lines of code are sets of coded instructions
terminated by hitting the enter key of a keyboard. Physical
lines of code and logical lines of code are almost identical for
some languages, but for some languages there can be
considerable differences. Generally, the difference between
physical lines of code and logical lines of code is often
excluded from the software metrics literature.
Strengths of physical lines of code (LOC) are:
It is easy to measure.
There is a scope for automation of counting.
It is used in a verity of software project estimation
tools.
Weaknesses of physical LOC are:
It may include significant dead code.
It may include white spaces and comments.
This metric is vague for software reuse.
It does not function for a few visual languages.
Direct conversion to function points is erroneous.
It is inconsistent for direct conversion to logical
statements.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
140 | P a g e
www.ijacsa.thesai.org
Strengths of the logical LOC are:
It omits dead code, blanks, and comments.
Mathematical conversion of logical statements into
function point metrics is possible.
Logical LOC are used in many software project
estimation tools.
Weaknesses of logical LOC are:
It can be difficult to measure.
These are not comprehensively automated.
These are ambiguous for a number of visual
languages.
This metric is vague for software reuse.
Direct conversion to the physical LOC metric may be
erroneous.
"Measuring software productivity by lines of code is like
measuring progress on an airplane by how much it weighs."
Bill Gates.
It is prudent to focus more on building expertise on
Function Point Analysis and use it effectively.
B. Function Point Metrics
The function point analysis to measure software application
is enumerated from analysis of the requirements and logical
design of the application. Function Point count can be applied
to Development projects, Enhancement projects, and existing
applications as well. [13] There are five key elements of
Function Point Analysis, which capture the functionality of the
application. These are:
External Inputs (EIs),
External Outputs (EOs)
External Inquiries (EQs)
Internal Logical Files (ILFs) and External
Interface Files (EIFs).
First three elements are of Transactional Function Types
and last two are of Data Function Types. Function Point
Analysis consists of performing the following steps:
Determine the type of Function Point count
Determine the application boundary
Identify and rate transactional function types to
calculate their contribution to the Unadjusted Function
Point count (UFP)
Identify and rate the data function types to calculate
their contribution to the UFP
Determine the Value Adjustment Factor (VAF) by
using General System Characteristics (GSCs)
Finally, calculate the adjusted Function Point count
When we examine the patterns of strengths and weaknesses
of function point metrics, we observe that for economic studies
and for studies that include non-coding work such as
specifications, function points are clearly superior to lines of
code metrics. [12]
Strengths of function point metrics are:
It stays stable regardless of programming languages
used.
It can compute non-coding activities such as
documentation.
It can measure non-coding defects in requirements and
design.
These are useful for software reuse analysis.
Function points are used for object-oriented economic
studies.
These are supported by a lot of software cost
estimating tools.
Mathematical conversion of function points into
logical code statements is very easy.
Weaknesses of function point metrics are:
Function Point counting requires good deal of
experience.
Function point counting can be protracted and pricey.
Function point counting automation is of indefinite
accuracy.
Function point counts are unreliable for those projects
that are below 15 function points in size.
Function point variant have no conversion rules to
IFPUG function points.
C. Object-Oriented Metrics
In todays software development environment, Object-
oriented analysis and design concepts are well known. Object-
Oriented Analysis and Design of software provide many
advantages such as reusability, decomposition of problem into
easily understandable object and the aiding of future
modifications. Object-oriented software development requires
a diverse approach from more traditional functional
decomposition and dataflow development methods. But the
OOAD software development life cycle is not easier than the
typical procedural approach. Therefore, it is necessary to
provide dependable guidelines that one may follow to help
ensure good OO programming practices and write reliable
code. Object-Oriented programming metrics is an aspect to be
considered. Metrics should be a set of standards against which
one can measure the effectiveness of Object-Oriented Analysis
techniques in the design of a system. [2]
Strengths of OO metrics are: [12]
The OO metrics are psychologically attractive within
the OO community.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
141 | P a g e
www.ijacsa.thesai.org
The OO metrics come out to be able to differentiate
simple from complex OO projects.
Weaknesses of OO metrics are:
The OO metrics do not support studies outside of the
OO paradigm.
The OO metrics have not yet been applied to testing.
The OO metrics have not yet been applied to
maintenance.
The OO metrics have no conversion rules to lines of
code metrics.
The OO metrics have no conversion rules to function
point metrics.
The OO metrics lack automation.
The OO metrics are difficult to enumerate.
Software project estimation tools do not support the
OO metrics.
OO metrics are not linked to all other known software
metrics. There are no conversion rules between the OO metrics
and any other metrics, so it is complicated to perform alongside
comparisons between OO projects and conservative projects
using the currently available OO metrics.
VI. FUTURE SCOPE
Looking at rising demand for the implementation and
successful case studies of software quality, it is safe to
conclude that in the coming years, software metrics
importance will increase multifold as industry leaders like
embrace newer and more stringent approaches to monitoring,
improving as well as delivering better software quality in
products as well as processes. A number of metrics are
proposed and exercised for measuring the quality of a system
before implementation. Future research directions include
improvement in existing metrics based on the nature and
magnitude of the problem statement. There is a scope for
various tools to support software project development reducing
time, effort and cost of the project in consistent manner.
VII. SUMMARY AND CONCLUSION
With the rapid advancement in software industries,
software metrics have also developed fast. Software metrics
become the basis of the software management and crucial to
the accomplishment of software development. It can be
anticipated that by using software metrics the overall rate of
progress in software productivity and software quality will
improve. If relative changes in productivity and quality can be
determined and studied over time, then focus can be put upon
an organizations strengths and weaknesses. Although people
appreciate the significance of software metrics, the metrics
field still needs to mature. Each of the key software metrics
candidates has broken into many competing alternatives, often
following national restrictions. There is no adequate
international standard for any of the extensively used software
metrics. Absence of firm theoretic background and the
assurance of methods, software metrics are still young in
comparison of other software theories.
REFERENCES
[1] G. Eason, B. Noble, and I. N. Sneddon, On certain integrals of Lips
Software Metrics SEI Curriculum Module, SEI-CM-12-1.1, December
1988
[2] Software Quality Metrics for Object Oriented System Environments,
June 1995, National Aeronautics and Space Administration, Goddard
Space Flight Center, Greenbelt Maryland
[3] Tu Honglei1, Sun Wei1, Zhang Yanan1, The Research on Software
Metrics and Software Complexity Metrics, International Forum on
Computer Science-Technology and Applications, 2009.
[4] http://www.pearsonhighered.com/samplechapter/0201729156.pdf.
[5] http://www.wohlin.eu/Articles/ICGSE11.pdf.
[6] Barbara Kitchenham, Shari Lawrence, Quality: The Elusive Target,
IEEE Software-Vol. 13, No. 1: JANUARY 1996, pp. 12-21.
[7] Henrike Barkmann, Rudiger Lincke and Welf L owe, Quantitative
Evaluation of Software Quality Metrics in Open-Source Projects.
[8] Dindin Wahyudin, Alexander Schatten, Dietmar Winkler, A Min Tjoa,
Stefan Biff,Defect Prediction using Combined Product and Project
Metrics , March 2008.
[9] Nachiappan Nagappan, Brendan Murphy, and Victor Basili, The
Influence of Organizational Structure On Software Quality: An
Empirical Case Study, January 2008.
[10] David I. Heimann,Implementing Software Metrics at a
Telecommunications Company A Case Study,2oo4
[11] Rakesh.L , Dr.Manoranjan Kumar Singh ,and Dr.Gunaseelan Devaraj
,Software Metrics: Some degree of Software Measuremen and Analysis
,(IJCSIS) International Journal of Computer Science and Information
Security, Vol. 8, No. 2, 2010
[12] Capers Jones, Chief Scientist Emeritus, Strengths and Weaknesses of
Software Metrics, Version 5, March 22, 2006.
[13] Kurmanadham V.V.G.B. Gollapudi, Function Points or Lines of Code?
An Insight
[14] Arti Chhikarai, R.S.Chhillar , Impact of Aspect Orientation on Object
Oriented Software Metrics. Vol. 2 No. 3 Jun-Jul 2011, Indian Journal
of Computer Science and Engineering (IJCSE)
[15] Lytz, R., Software Metrics for the Boeing 777: A Case Study, Software
Quality Journal, 4, 1-13 (1995)
AUTHORS PROFILE

1
Ms. Mrinal Singh Rawat is Assitant Professor in the Department of Computer
Science and Engineering in MGMs COET, Noida, UP, INDIA. Her Research
activities are based on Software Engineering and Reliability Engineering.She
is pursuing her M.Tech in Computer Science and Engineering from Amity
University.

2
Ms. Arpita Mittal is working as Assistant Professor in Department of
Computer Science at IIMT Merrut, UP, INDIA. Her Research activities are
based on Software Engineering and Software Tesing. She is pursuing her
M.Tech in Computer Science and Engineering from Amity University.

3
Mr. Sanjay Kumar Dubey is working as Assistant Professor in Department of
Computer Science and Engineering in Amity University Noida, UP, INDIA.
His Research area includes Software Engineering and Usability
Engineering.He is pursuing his Ph.D in Computer Science and Engineering
from Amity Unversity.



(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
142 | P a g e
www.ijacsa.thesai.org
A Cost-Effective Approach to the Design and
Implementation of Microcontroller-based Universal
Process Control Trainer

1
Udeze Chidiebele. C,
3
Uzedeh Godwin,
1,3
R & D Department, Electronics Development Institute
(FMST-NASENI), Awka, Nigeria.
2
Prof. H. C Inyiama,
4
Dr C. C. Okezie,
2,4
Electronics and Computer Engineering Department,
Nnamdi Azikiwe University, Awka, Nigeria.


AbstractThis paper presents a novel approach to the design and
implementation of a low-cost universal digital process control
trainer. The need to equip undergraduates studying Electronic
Engineering and other related courses in higher institutions with
the fundamental knowledge of digital process control practical
was the main objective of the work. Microcontroller-based design
and implementation was the approach used where only one
AT59C81 with few flip-flops were used for the whole eight
processes covered by the trainer thereby justifying its low-cost
and versatility.
Keywords- process control; control algorithm; Algorithmic State
machine (ASM) chart; State Transition Table (STT); Fully-
expanded STT; Control software.
I. INTRODUCTION
The basic objective in process control is to regulate the
value of some quantity which means to maintain that quantity
at some desired value regardless of external influences. The
desired value is called the reference value or set point [2].
Inyiama H. C and Okezie C.C (2007) stated that a process
control is typically a sequential logic system whose control
algorithm can be represented in the form of a flow chart called
an algorithmic State machine (ASM) chart [3] or in the form of
State Transition Diagram (STD) [4]. The ASM chart is a
diagrammatic description of the output function and the next-
state function of a state machine to implement an algorithm and
becomes part of the design documentation when completed.
The symbols covered are the STATE BOX, THE DECISION
BOX, THE CONDITIONAL OUTPUT BOX and ASM
BLOCK.
According to Clare C.R [1973], the ASM chart has three
basic elements: the state, the qualifier and the conditional
output [3]. The need to equip undergraduates studying
Electronic Engineering and other related courses in higher
institutions with the fundamental knowledge of digital process
control practical was the main objective of this work. The
design approach used was the use of Algorithmic State
machine (ASM) charts State Transition Diagrams (STD), State
Transition Tables (STT), Fully Expanded STT, and Link Path
Addressable ROM Structure. The significance of the work
extends to the fact that it presented a standard approach to the
design and implementation of the process control systems that
can be applied to any process controls in the industries and it
also bridges the gap created in the lives of these students due to
their lack of exposure to electronics practical.
The following processes were covered by the trainer:
Temperature Level Control systems
Automatic Liquid Dispenser System
Automatic Water Pump control system
Traffic Light Control System
Upper Tank Water level Control systems
Lower Tank Water level Control systems
Automatic Gate control system
Automatic Street Light Control system
II. METHODOLOGY & DESIGN ANALYSIS
The design of the digital process control system can be
achieved through various methods which include: Gate-
Oriented Design which involves equation-to-gate conversions,
map simplifications, output function synthesis, and next-state
function synthesis, flip flops, state assignment and hazards.
When such discrete logic gates (such as AND, NAND, OR,
NOR, EX-OR, INVERTERS etc) and memory flip flops are
used, what is termed a RANDOM LOGIC system results. Since
such a system involves mostly small scale integration
components, the component count is usually high for a fairly
complex circuit. This implies several interconnections and
many potential sources of error. Modifications and
maintenance are also difficult to achieve when either re-
designing or fault-finding becomes necessary. A random logic
system is therefore very inflexible.
Fortunately however, logic systems can be implemented in
forms more structured than random logic. Such systems
employ structured logic devices such as Multiplexers (or Data
Selectors) (5), and Read-Only-Memories (ROMs). A
multiplexer-based controller requires as many multiplexers as
there are columns in the bit-pattern sequence to be generated
and each of these multiplexers must have at least as many data
input lines as there are rows in the bit-pattern sequence. Thus,
if an 8-bit pattern is to be generated by means of multiplexers, a
total of 64 data input lines would be involved. This is in
addition to other control inputs and outputs necessary in such a
system. The rapid proliferation of data input lines (and hence
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
143 | P a g e
www.ijacsa.thesai.org
A
B
0
0
1

1
potential sources of error) in multiplexer based complex
sequential logic systems is its main disadvantage in such
applications. Multiplexers being structured logic devices do,
however have a considerable advantage over random logic in
that errors in the design can be corrected simply by altering the
logic levels applied to their data inputs. For compact, reliable
and easily maintainable implementation of a complex
sequential logic ROMs have an edge over multiplexers and are
usually preferred.
The microcontroller-based implementation was chosen for
this work due to its numerous advantage over the others which
include its simplicity, and the fact that a simple control
software can be developed which can be used without any
modification in any control system, no matter how complex or
how simple, and even when the numbers of input qualifiers,
state code, size and number of output lines differ from one
control system to the other, provided one input port is sufficient
for Address inputs and one output port for the control pattern
output. The system is made up of the hardware subsystem and
software subsystem. The hardware subsystem is comprised of
the input interface, the control systems and the output interface.
The Structure of the Microcontroller-based Digital Process
Control is shown in Fig. 2 below. It comprise of the
microcontroller with its input and output ports, the input
interface subsystem connected to the input port of the
microcontroller, the output interface subsystem connected to
output port of the microcontroller. The input interface system
comprise of all the sensors that will be used for a particular
process. For example in the temperature level control system,
LM 35 is the sensor, which senses the temperature of the water
container. Also the output interface comprises all the
transducers such as light emitting diodes, LCDs and so on.
Note that buttons and keypads are part of the input subsystems
since they are used for selecting the particular process to be
controlled.









Figure 4. The Structure of the Microcontroller-based Digital Process
Control.
III. THE GENERATION OF CONTROL BIT- PATTERN
SEQUENCE
The procedure that is followed for the generation of the
control bit-patterns include: the drawing of ASM chart for each
of the processes, transforming the ASM charts into a state
transition table, and expanding the state transition table fully so
that the location address and the content address for the process
are generated which is then burned into the flash drive of the
microcontroller.
This procedure was used for all the processes to generate
the control bit-patterns in Table 3 below but the design process
that led to that was illustrated in this paper by one of the
processes which is the temperature level control system of
water or other liquids . The ASM chart of the temperature
control system is shown in Fig. 2 below. These ASM chart is
then transformed into a state transition table of Table1 and then
to fully expanded state transition table of Table2.

















Figure 5. (i)The ASM chart of the temperature control system
2 (iii) STATE ASSIGNMENT 2 (ii) STATE MAP








The labels or names inside some of the state boxes are the
state outputs. The ASM charts of the Fig. 2 above have only
one state output namely: HHEATER. The logic level of the
output signal is high or active when the control system is in that
state. The bit pattern at the top right end of the state box is its
state code. The letters B, A above the state code signify that
two flip flops B and A are used to represents the various states
of the machine.
The state codes are the logical levels at the Q outputs of
these two flip flops respectively. Each rectangular box in the
STATE
NAME
STATE
CODE
ST0 0 0
ST1 0 1
ST2 1 1
ST3 1 0

ST0

ST1

ST3

ST2

INPUT PORT OUTPUT PORT
Input
Interface
Subsystem
(Sensors)
Output
Interface
Subsystem
(Transducers)
MICROCONTROLLER
L

L6
ST3

U
ST1

L3
ST2
L5
HHEATER
CALL DELAY
ERROR STATE
RESET
ST0

L2
01
L1
10

L
11
L4
HHEATER
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
144 | P a g e
www.ijacsa.thesai.org
ASM chart is a state box. The word (e.g ST0) enclosed in a
circle at the bottom left hand corner of a state box is the state
name. Here ST0 stands for state 0, and similar interpretations
apply to the other states, hence ST1 Means state 1, and so on.
Each decision box has one entry path and two exit paths. The
exact value of an input qualifier determines which exit path is
followed out of a decision box. In the ASM chart of Fig. 2 Uth
and Lth are the input qualifiers.
With the help of a K_map in which the state names are
inserted serially in an adjacent cells (Fig. 2ii) and which is
called a state map, the state codes are chosen such that only one
bit changes level as one moves from one state of the control
system to another. This is clearly brought out in the state
assignment of Fig. 2iii. This method of state assignment
facilitates complexity reduction when hardwired logic is the
preferred technique of control system implementation and
helps to prevent race hazards [3].
Every ASM chart has an equivalent tabular representation
known as a State Transition Table (STT) [3]. An ASM chart
can be fully described in terms of the link paths comprising it.
A State machine such as is represented by an ASM chart
attempts to change state (i.e. transits from the present state to
the next) when a clock pulse occurs. A link path is a path
followed from the present state back to itself or to another state,
when the clock pulse arrives. When there is an input qualifier
between the present state and another, the logic level of the
qualifier determines the next state the machine goes to at the
clock pulse. If there is no qualifier between the present state
and the next, the machine must unconditionally transit from its
present state to the adjacent state in the forward direction, when
a clock pulse arrives.
The ASM chart of Fig. 2i has 7 link paths labeled L1 to L7.
L1 is the link path from state 0 back to itself when the input
qualifier is 0. Also L2 is the link path from state 0 to state 1
when the input qualifier Lth is 1. Similarly, L3 represents the
transition from state ST1 to state 2 when the qualifier Lth is 0.
L4 is the transition from state2 back to itself when the input
qualifier Uth is 0 and L5 is the transition from state 2 to state 0
when Lth input qualifier Uth is 1. L6 is the transition from state
1 to state 3 which is the error state where the system stays until
a key is pressed to return it to state 0 through link path L7.
In the STT of Table 1 a number of dashes appear under the
columns headed by the input qualifiers Lth and Uth. A dash (_)
implies that the input qualifier above that column is not
relevant for the transition being made in the link path shown on
the STT row where the dash appeared. That input qualifier may
become relevant in some other link path transition and then its
value would become 0 or 1, rather than a dash. A dash in the
STT also means that the input qualifier that appears as column
heading for that dash may be at logic 0 without affecting the
control process at that material time.
A State Transition Table (STT) is said to fully-expanded
when all the dashes on each row are given all possible
combination of logic values, leading to new rows in the State
Transition Table, one for each combination of the logic values
for the dashes on that row. In this respect, a STT data row with
one row when that (dashed) qualifier is given the logic value 0
and the other when that qualifier is given the logic level 1.
Similarly a STT data row with 2 dashes expands into 4 STT
rows. Assume the (dashed) qualifiers are represented by q1, q2.
Then the first STT row in the expansion will give q1, q2 = 0, 0,
the second would have q1, q2 = 0, 1, the third row would have
q1, q2 = 1, 0, the fourth row would have q1, q2 = 1,1. Three
dashes on an STT data row would in like manner lead to 8 rows
in the fully expanded STT of Table 2 results.
IV. MICROCONTROLLER-BASED IMPLEMENTATION
The high capacity of ROMs relative to the number of
unique bit-patterns in a fully expanded STT suggests the use of
a single ROM to store the fully expanded bit-pattern of all the
processes. The link-path addressable word structure is based on
storing the next state and the output for each link path in the
ASM charts. The next-state portion of the ROM word is called
the LINK PART, while the output portion of the ROM word is
called the INSTRUCTION PART. Each address is a function
of the present state and qualifier inputs and called a link-path
address. In general any process described by an ASM chart can
be implemented with this structure, called a link-path
addressable ROM.
Since a microcontroller-based implementation is used, the
same link path addressable ROM patterns will be programmed
into its ROM or the Erasable and Programmable ROM (i.e.
EPROM) or Flash Drive (now available in newer
microcontrollers). The ADDRESS inputs to the Address
Decoder part of the ROM device would now be input via an
input port of the microcontroller which is then used to locate
the corresponding NEXT STATE and STATE OUTPUT.
When a clock pulse occurs, the NEXT STATE pattern
becomes the present state pattern. This joins the input qualifiers
to comprise the next Address input to be used by the
microcontroller. A power-up one-shot applied to the D flip
flops that feedback the Next State as the present state when a
clock pulse occurs, initializes the system to state 0 at the start
up[6]. Thereafter, the control system behaves as defined by the
ASM chart, with the help of the control software running in the
microcontroller. The user may use a simple switch to stop the
control software run.
A. The Control Software
A very important universal concept that is a natural
outcome of the use of a fully expanded STT in a link path
addressable ROM structure, as part of a microcontroller-based
implementation is that a simple software can be developed
which can be used without any modification in any control
system, no matter how complex or how simple, and even when
the numbers of the input qualifiers, state code, size and number
of the output lines differs from one control system to the other,
provided one input port is sufficient for address inputs and one
output port for control pattern output.
The flowchart for this universally applicable control
software is shown in the Fig. 3 below; with a pseudo code that
explains what it does above it.
The control cycle time is the time delay required for the
system to settle and become ready for the next round of input-
output. Its default value is zero seconds.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
145 | P a g e
www.ijacsa.thesai.org
TABLE 1. THE STT FOR TEMPERATURE LEVEL CONTROL.
LINK
PATHS

INPUT
QUALIFIERS

Lth Uth
PRESENT
STATE
NAME

PRESENT
STATE
CODE
B A
NEXT
STATE
NAME



NEXT
STATE
CODE
B A
STATE
OUTPUT

HHEATER

L1 0 _ ST0 0 0 ST0 0 0 0
L2 1 _ ST0 0 0 ST1 0 1 0
L3 0 _ ST1 0 1 ST2 1 1 1
L4 _ 0 ST2 1 1 ST2 1 1 1
L5 _ 1 ST2 1 1 ST0 0 0 1
L6 1 _ ST1 0 1 ST3 1 0 1
L7 _ _ ST3 1 0 ST0 0 0 0
TABLE 2. FULLY EXPANDED STT TABLE FOR THE SYSTEM
LINK
PATH
LOCATION
ADDRESS
(hex)
ADDRESS PATTERN
Lth Uth B A
CONTENT
PATTERN
B A H
LOCATION
CONTENT
(hex)
L1

0 0
0 4
0 0 0 0
0 1 0 0
0 0 0
0 0 0
0 0
0 0
L2 0 8
0 C
1 0 0 0
1 1 0 0
0 1 0
0 1 0
0 2
0 0
L3 0 1
0 5
0 0 0 1
0 1 0 1
1 1 1
1 1 1
0 7
0 7
L4

0 3
0 B
0 0 1 1
1 0 1 1
1 1 1
1 1 1
0 7
0 7
L5 0 7
0 F
0 1 1 1
1 1 1 1
0 0 1
0 0 1
0 1
0 1
L6

0 9
0 D
1 0 0 1
1 1 0 1
1 0 1
1 0 1
0 5
0 5
L7 0 2
0 6
0 E
0 A
0 0 1 0
0 1 1 0
1 1 1 0
1 0 1 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0
0 0
0 0
0

B. Pseudo Code
BEGIN
Repeat
Input Next Control Pattern Address from Input Port;
Retrieve Next Control Pattern from Link-path
Address Rom Structure;


Output Next Control Pattern Address from Output Port;
Delay for Control Cycle Time.
Until HSTOP = 1
END

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
146 | P a g e
www.ijacsa.thesai.org
OUTPUT ROM PATTERN TO OUTPUT PORT
INPUT ADDRESS FROM INPUT PORT
RETRIEVE ROM PATTERN AT ADDRESS

DELAY FOR CONTROL CYCLE TIME
HSTOP?
EN
D
BEGIN
C. Flow Chart

Figure 6. The Flowchart rep. of the control software
V. RESULTS AND DISCUSSION
A software-based universal digital process control system
was achieved with the program design method presented.
Following the procedure for developing a fully-expanded state
transition table, which begins with representing the process in
an ASM chart, developing a state transition table for the
process from the ASM chart and then translating the state
transition table to a fully-expanded state transition table, the
fully-expanded state transition table for the other processes was
also developed and all of them were shown in Table 3. The
location address is the input address of the processes, which is
used to locate where in the ROM memory the location content
is stored. The location address and the location content will be
burned into the ROM permanently.
The difference in terms of programming effort, between a
software-based controller intended for just one control system
and that designed for several control systems (each with a
different number of input qualifiers) is minimal. Fig. 4
illustrates a typical microcontroller-based digital process
control system. With this trainer the student can perform
several experiments on each of the processes as covered in the
laboratory manual developed for the trainer which enables him
gain the needed practical knowledge on digital process control.
The keyboard (Fig. 4) is used for selecting the particular
process desired to be controlled. It also facilitates the input of
variables or control parameters which make software-based
controllers very flexible, while the LCD display enables the
microcontroller to transmit responses to user commands in
addition to providing the current status of the controlled device
where necessary.

TABLE 3. LOCATION ADDRESS & CONTENT FOR THE
CONTROLLER ROM FROM THE FULLY-EXPANDED STT OF THE 8
PROCESSES.
Location
address
(hex)
Location
content
(hex)
Location
address
(hex)
Location
content
(hex)
0 0 0 0 0 4 0 0 0 0
0 0 4 0 0 4 0 4 0 0
0 0 8 0 2 4 0 8 0 2
0 0 C 0 0 4 0 C 0 0
0 0 1 0 7 4 0 1 0 7
0 0 5 0 7 4 0 5 0 7
0 0 3 0 7 4 0 3 0 7
0 0 B 0 7 4 0 B 0 7
0 0 7 0 1 4 0 7 0 1
0 0 F 0 1 4 0 F 0 1
0 0 9 0 5 4 0 9 0 5
0 0 D 0 5 4 0 D 0 5
0 0 2 0 0 4 0 2 0 0
0 0 6 0 0 4 0 6 0 0
0 0 E 0 0 4 0 E 0 0
0 0 A 0 0 4 0 A 0 0
1 0 4 0 2 5 0 0 0 2
1 0 8 0 2 5 0 4 0 2
1 0 C 0 2 5 0 8 0 2
1 1 0 0 6 5 0 C 0 2
1 1 4 0 6 5 1 0 0 6
1 1 8 0 6 5 1 4 0 6
1 1 C 0 6 5 1 8 0 6
1 0 1 0 4 5 1 C 0 6
1 0 5 0 4 5 0 1 0 4
1 1 0 0 4 5 0 5 0 4
1 1 5 0 4 5 1 0 0 4
1 0 9 0 C 5 1 5 0 4
1 0 D 0 C 5 0 9 0 C
1 1 9 0 C 5 0 D 0 C
2 0 3 0 D 6 0 3 0 D
2 0 B 0 D 6 0 B 0 D
2 1 3 0 D 6 1 3 0 D
2 1 B 0 D 6 1 B 0 D
2 0 7 0 9 6 0 7 0 9
2 0 F 0 9 6 0 F 0 9
2 1 7 0 9 6 1 7 0 9
2 1 F 0 9 6 1 F 0 9
2 0 2 0 8 6 0 2 0 8
2 0 6 0 8 6 0 6 0 8
2 1 2 0 8 6 1 2 0 8
2 1 6 0 8 6 1 6 0 8
2 0 A 0 0 6 0 A 0 0
2 0 E 0 0 6 0 E 0 0
3 0 4 0 0 7 0 0 0 0
3 0 8 0 2 7 0 4 0 0
3 0 C 0 0 7 0 8 0 2
3 0 1 0 7 7 0 C 0 0
3 0 5 0 7 7 0 1 0 7
3 0 3 0 7 7 0 5 0 7
3 0 B 0 7 7 0 3 0 7
3 0 7 0 1 7 0 B 0 7
3 0 F 0 1 7 0 7 0 1
3 0 9 0 5 7 0 F 0 1
3 0 D 0 5 7 0 9 0 5


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
147 | P a g e
www.ijacsa.thesai.org
TABLE 4. COST IMPLICATIONS OF THE PROJECT
ITEM QTY Unit Cost
(Naira)
Cost
(Naira)
1 Sensors 7 400 2800
2 ADC0804 1 200 200
3 Microcontroller, AT89C51 1 350 350
4 4x20 Liquid crystal display 1 2500 2500
8 DC relays (12V & 6V) 2 70 140
12 10k ohm resistors 7 5 35
13 Transistor, BC337 5 5 25
14 Diodes, 1N4001 6 5 30
15 10kohm Variable resistor 2 20 40
16 10uf & 33pf capacitor 2 20 40
19 7805 1 40 40
21 Copper clad board 1 100 100
22 40 pin IC socket 1 40 40
23 Soldering Lead 1 150 150
25 Casing materials - 1000 1000
TOTAL 7490NGN
Table 4 shows the cost implication of the project. The total
cost implication of the project is 7490 NGN (Nigerian Naira).
When compared to the cost implication of the same project
using other methods such gate-oriented designed method,
multiplexer and ROM-based design method as shown in Table
5, it is cheaper thereby justifying the cost-effectiveness of the
approach used in this work.
TABLE 5. COST COMPARISON WITH OTHER METHODLOGIES
Methodology Cost (NGN)
Gate-oriented design method 12,350
Multiplexer-based design method 10,200
ROM-based design method 9,400
Microcontroller-based design method 7,490
VI. CONCLUSION
The use of a single microcontroller to control several
processes, based on storing the fully expanded State Transition
Tables of those processes in its ROM or flash drive makes
possible the realization of a low-cost universal processes
control trainer.
REFERENCES
[1] Inyiama H. C, Okezie C.C, Designing microcontroller-based universal
process control systems, Volume 2 Number 2 (Electroscope), November
2007. Department of Electrical and Electronics Engineering, Nnamdi
Azikiwe University Awka. Pp.11-26.
[2] Curtis D Johnson, Process Control Instrumentation Technology, 8th
Edition, 2006, Prentice-Hall Inc. pp. 1-10.
[3] Clare C.R, Designing logic systems using state machines (U.K,
MacGraw-Hill, 1973) pp 1-108.
[4] Roger L. Tokheim, Digital Electronics, Principles and Applications Fifth
Edition,1999, Glencoe McGraw-Hill. pp269-276.
[5] Williams, G.E., Digital Technology, Principles and Practice. Science.
Research Associats Inc., (1974), pp. 96-105,202-245.
[6] Walter G. Jung IC timer cookbook (4300West 82nd St. Indianapolis
48258 USA, Howard W. Sams & Co. Inc. 1977) pp1-36.
[7] Michael J. Pont, Embedded C, 2002, Pearson Education Limited, pp 17-
34.
[8] Bertram J. E: The concept of state in the the analysis of discrete time
control system, 1962 Joint Autom. Control Conf. New York University
(June 27-29), Paper No. 11-1
[9] Inyiama H.C, Unpublished lecture Notes on Real-Time computing and
control, Department of Electronic and Computer Engineering, Nnamdi
Azikiwe University Awka 2008.
KEYBOARD
INTERFACE
INPUT
QUALIFIERS
FROM VARIOUS
PROCESSES
Figure 4. Microcontroller-based digital process control system
Feedback Bit Pattern to
C I/P port
FEEDBACK
SIGNALS FROM
C
SOFTWARE BASED
CLOCK
LCD DISPLAY
PANEL
LED INTERFACE
MEMORY
FLIP FLOP
OTHER
TRANSDUCERS
SUCH AS MOTORS,
LOUD SPEAKERS
etc.
ROM or EPROM or FLASH DRIVE WITH FULLY
EXPANDED STT AND CONTROL SOFTWARE
INPUT PORTS O/P PORT 1 O/P PORT 2
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
148 | P a g e
www.ijacsa.thesai.org
Self-regulating Message Throughput in Enterprise
Messaging Servers A Feedback Control Solution

Ravi Kumar G
HP, Research Scholar, JNTUH
Bangalore, India
C.Muthusamy
Yahoo
Bangalore, India
A.Vinaya Babu
JNTUH
Hyderabad, India


AbstractEnterprise Messaging is a very popular message
exchange concept in asynchronous distributed computing
environments. The Enterprise Messaging Servers are heavily
used in building business critical Enterprise applications such as
Internet based Order processing systems, pricing distribution of
B2B, geographically dispersed enterprise applications. It is
always desirable that Messaging Servers exhibit high
performance to meet the Service Level Agreements (SLAs).
There are investigations in this area of managing the
performance of the distributed computing systems in different
ways such as the IT administrators configuring and tuning the
Messaging Servers parameters, implement complex conditional
programming to handle the workload dynamics. But in practice
it is extremely difficult to handle such dynamics of changing
workloads in order to meet the performance requirements.
Additionally it is challenging to cater to the future resource
requirements based on the future workloads. Though there have
been attempts to self-regulate the performance of Enterprise
Messaging Servers, there is a limited investigation done in
exploring feedback control systems theory in managing the
Messaging Servers performance. We propose an adaptive control
based solution to not only manage the performance of the servers
to meet SLAs but also to pro-actively self-regulate the
performance such that the Messaging Servers are capable to meet
the current and future workloads. We implemented and
evaluated our solution and observed that the control theory based
solution will improve the performance of Enterprise Messaging
Servers significantly.
Keywords-Feedback control; Message Oriented Middleware;
Enterprise Messaging; Java Messaging Service; JMS Providers;
Adaptive Control
I. INTRODUCTION
Enterprise Messaging also known as Message Oriented
Middleware [1] is a popular asynchronous message exchange
mechanism in heterogeneous distributed applications. It
provides the applications in a distributed environment to send
and receive messages, but still being loosely coupled. Loose
coupling between enterprise class applications and legacy
systems such as business workflow applications, databases, and
data warehouses plays a significant role in Enterprise
Application Integration (EAI) [2]. The Message based
integration provides automation and simplifies the time
consuming integration tasks like create, deploy and manage
integration solutions. There are many such applications such as
Business to Business (B2B) solutions, messaging across
various entities within a business enterprise that are
geographically separate where asynchronous messaging
becomes a major building block [2]. Asynchronous Messaging
is a backbone for many of the Event driven architectures due to
the obvious advantages of asynchronous systems where the
message client need not maintain the connection and session
with the message receiver; no confirmation is required from the
receiving application [2]. As we discussed Enterprise
Messaging is an important element in the business critical
environments, it is always important for the Enterprise
Messaging Servers to exhibit high performance and
availability. Typically there would be Service Level
Agreements (SLAs) [3] defined between the business service
providers and the consumers. Performance is an obvious
Service Level Objective in such SLAs. Any violation of
performance SLOs [4] will affect the business and reputation of
the business enterprise. In this paper we want to discuss the
performance regulation of Java based Enterprise Messaging
Servers. There are different implementations of such Enterprise
Messaging Servers. The Java based Messaging Service is
called as Java Message Service [5], included in the
specification for Java based Enterprise Environments called as
JEE (called as J2EE previously) [6]. There are different
vendors who implemented the JMS Specification and Java
based Enterprise Messaging Servers are referred as JMS
Providers. Hence forth in the document the Enterprise
Messaging Servers are referred as JMS Providers [7].
Typically the performance of JMS Providers is measured
by its message throughput, though CPU and Memory usage [8]
are common metrics to measure the performance of any
computing server. The message throughput will depend upon
various factors such as the number of subscribers, message
size, number of publishers, and number of JMS message
brokers [9]. By tuning these different parameters the desired
performance can be achieved on the JMS Providers. One of
the mechanisms to improve the JMS providers performance is
by following some best practices such as setting non-durable
messages, set the message time to live parameter appropriately,
close message publishers and subscribers when they complete
their jobs [10]. But these kinds of practices will not be able to
address different kinds of JMS environments and applications
limiting the performance improvement. The other mechanism
is to provide the facilities to the administrators to configure
[11] and fix the various parameters values which influence the
JMS Provider performance. Due to the dynamics of messages
flow and workload on the JMS Provider, it will be difficult for
the administrator to tune these values accurately and
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
149 | P a g e
www.ijacsa.thesai.org
periodically. When there are sudden huge loads administrator
may decide to provision additional resources which may be left
unutilized [12] later when there are relatively lesser loads. This
will eventually lead to either not addressing the performance
needs or ineffective utilization of computing server assets.
Another way to manage the performance is to include
conditional programming within JMS Provider implementation
to change the values of the parameters at runtime based on the
workload and deviation from the expected performance. This
method is though useful it is very complex because during
design the workload dynamics have to be accurately estimated.
During implementation the conditional programming is
implemented which is very complex [13] as the conditions
implemented may not be sufficient to meet the run time
dynamics, any spikes in the workload. To summarize, though
there are different mechanisms to adjust the JMS Providers
parameters to regulate and improve the message throughput
either it involves manual intervention, involve complex
conditional programming implementation.
In order to handle such situations, we propose an adaptive
control [14] based solution that regulates the message
throughput according to the pre-defined reference using a
feedback controller. Also, predict the future load on the JMS
Provider, modify the control parameters accordingly. There
may be a case where the future load predicted may demand
additional servers; our controller will actuate a signal to
provision additional resources.
In this paper we will first present a background on the
choosing feedback control systems in Distributed computing
systems, then a brief overview on the Java Messaging Service,
followed by discussing the adaptive controller algorithm that
we have implemented to regulate the performance of the JMS
Provider.
II. BACKGROUND
We have discussed the importance of JMS Providers and
importance of their performance in building business critical
applications and services in enterprise level or at internet level.
There are attempts to predict the performance of JMS Providers
[15], or study and compare the performance of different
vendors of JMS Providers [16]. Additionally there are some
best practices [10] identified to improve the JMS Providers
performance. Manual configuration is one of the most
common approaches followed to tune the JMS Provider
performance. There is a very limited investigation done in
automatic regulation the JMS Providers performance. The
message throughput of the JMS Providers depends upon the
number of subscribers, publishers and number of brokers.
Allowing more number of subscribers on a given JMS Provider
may decline message throughput or having a less number of
subscribers may leave the JMS Provider less utilized. The
control system based solutions provide mechanism to
automatically tune the maximum number of subscribers in an
optimal operating range. In this paper we propose a control
systems based solution for managing the performance by
tuning the maximum number of subscribers that influence the
message throughput.
Control systems theory has been in investigation to address
these kinds of problems related to regulating the performance,
in computing [17]. But the majority of focus is on Web
Servers [18][19][20], Application Server performance
regulation [21], in computer networks such as congestion
control [22]. There are recent investigations to explore the
applicability of control systems in other areas of Java based
Cloud and Enterprise Environments [23], database driver cache
hit ratio improvement [24], spring based software applications
[25]. But in our study we have observed there is no
investigation carried out in applying feedback control system
theory in improving the performance of JMS based servers. We
investigated to apply control systems theory in Enterprise
Messaging server performance improvement and evaluated
how the feedback controllers improve the JMS Providers
performance significantly.
III. THEORITICAL AND PRACTICAL CONSIDERATIONS
The message throughput of the JMS Providers depends
upon various factors such as publishers, subscribers, JMS
brokers. The performance varies based on whether the
messages are persisted are not. The JMS Providers exhibit
higher performance when the messages are not persisted. In
this paper the persistence factor is not considered and the
performance is evaluated with proposed solution. The
subscribers are identified as a significant independent variable
influencing the message throughput. The number of publishers
and the brokers will have a definite impact on the message
throughput to cater huge publisher and subscriber volumes.
When the subscribers and message throughput are depicted in
mathematical model, the accuracy of the JMS Provider model
depends upon the constant values chosen for that model. These
constants can be determined by using different data set values
of message throughput for varying subscribers. These values
may not hold good for different workload conditions on the
JMS Provider, but the best suitable constants can be chosen
before running the experiments.
IV. ENTERPRISE MESSAGING PRIMITIVES
In this section we discuss a brief overview of the Enterprise
messaging [26] also known as Message Oriented Middleware
(MoM). The key concept behind MoM is the asynchronous
messaging. It means that the sender is not required to wait for
the message to be received or handled by the receiver. The Fig
1 shows high level diagram of MoM. The sender can forward
the message and continue the processing. The asynchronous
messages are treated as autonomic units. The message contains
all the data and state needed by the business logic that
processes it.
A. Enterprise Messaging Architectures

Figure 1. Message Oriented Middleware
1) Centralized Architectures :
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
150 | P a g e
www.ijacsa.thesai.org
In Centralized Architectures there will be a Message Server
also called as a message router or message broker that is
responsible of sending messages such that the message sender
is decoupled from the message receiver. This enables the
clients to be added and removed without impacting the system.
In this model, the hub-and-spoke topology is used as shown in
Fig 2 below:

Figure 2. Centralized MoM Architecture
2) Decentralized Architectures
In Decentralized architectures, the IP multicast is used at
the network level. There is no centralized server and some of
the JMS functionality like persistence, transactions, security is
embedded in the client application. The messaging routing is
delegated to the network layer by using the IP multicast
protocol as shown in Fig 3.

Figure 3. Decentralized MoM Architecture
B. Java Messaging Service [26]
The Java Message Service (JMS) is a specification that
proposes programming API for Enterprise Messaging. JMS
supports messaging as a first-class java distributed computing
paradigm. There are many vendors who implemented the JMS
specification and such implementations are called JMS
Providers, which are nothing but Enterprise Message Servers
based on Java.
1) JMS Messaging Models
The JMS provides two types of messaging models, point-
to-point and publish-subscribe models. The intermediate
element that enables the communication between the message
producer and message consumer in JMS is called a broker.
There are two types of JMS brokers as explained below.
a) Point-to-Point
The Fig 4 below shows point-to-point model in which the
producer can send a message to only one consumer. In JMS
Providers such JMS Brokers called Queues.

Figure 4. Point-to-Point Model
b) Publish-Subscribe
The Fig 5 below shows publish-subscribe model in which
the producer can send a message to many consumers. In JMS
Providers such JMS Brokers called Topics.

Figure 5. Publish-Subscribe Model
V. ADAPTIVE CONTROL
The message throughput (T) depends upon various factors
such as the number of subscribers and publishers to the
different brokers of the JMS Provider, The messages size,
number of brokers running. In this paper we have considered
how the number of subscribers of the JMS brokers affects the
message throughput (T). Though there are other parameters
that influence the JMS Provider message throughput, the
maximum number of subscribers allowed on the server will
affect significantly. The number of publishers are considered to
be constant as 1 in our implementation. The Fig 6 is a Single
Input Single Output (SISO) Adaptive control system [27] that
shows how the message throughput is regulated using the
controller and the Predictor.

Figure 6. Adaptive Control of JMS Provider
We explain how the message throughput depends upon the
number of subscribers of the JMS Provider.
The following equation (1) represents the Message
Throughput and its relation with the number of subscribers.
T = bS (1)
Where
T = Message Throughput of JMS Provider measured as
number of messages per unit time
S = Total number of subscribers on the JMS Providers
b = proportional coefficient for the Subscribers
There is a feedback control loop that is implemented which
is used to calculate the error signal of the actual Message
throughput (Ta) and the Reference Message Throughput (Tref).
The error signal is represented by the equation (2)
E = T
ref
- T
a
(2)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
151 | P a g e
www.ijacsa.thesai.org
The controller takes the error signal as one input and the
other input signal to the controller is the predicted values of
number of subscribers and message throughput. The predicted
subscribers will help in estimating the possible future
subscribers volume. The predicted subscribers value is used
to predict the message throughput and the latter one is
important to determine the future resource requirements. The
resources may either be more JMS brokers or additional virtual
machines [28] that can be scaled to cater the future load
requirements on the JMS Providers. There are threshold values
defined for the message throughput based on which the
actuator signals are triggered either to add new virtual
machines or brokers. The following sections explain the
different parts of the solution in detail.
A. Modeling JMS Provider
The JMS Provider, whose message throughput needs to be
controlled, has to be mathematically modeled first in order to
apply the feedback control techniques. There are many ways to
model the compute systems such as difference equations [29],
ARMA models [30] that is based on Least Squares Parameter
Estimation [31]. In our solution we have used the ARMA
model to represent the JMS Provider. The Fig 7 below shows
ARMA based modeling of the JMS Provider.

Figure 7. JMS Provider Model for Feedback control
1) Parameter Estimation
In the JMS Provider the message throughput is defined as a
function
The number of subscribers that the JMS Provider is
supporting.
Though the number of publishers and number of
message brokers also influence the JMS Provider
performance we considered the number of subscribers
as the factor in our paper.
According to ARMA, in a Single Input Single Output
model, for a given sample data set, the next sample of the
output can be predicted using the current and previous inputs.
The same is explained in the equation (3) below:
y(t + 1) = ay(t) + bu(t) (3)
Where
y(t) = The current output
u(t) = The current input
a = The model parameter to be estimated
b = the model parameter to be estimated
y (t + 1) = the output in the next step

The same ARMA model if is applied to model the JMS
Provider, it is represented by the equation (4):
T(t + 1) = aT(t) + bS
max
(t) (4)
Where
T (t) = The current output of message
Throughput
S
max
(t) = the current input of maximum number
of Subscribers
a = the model parameter to be estimated
b = the model parameter to be estimated
T (t + 1) = the output in the next step
The ARMA model is used to estimate the model parameters
a and b. The details of the experiments and the estimated
values are explained in the section VI. Implementation and
Analysis. Based on our experiments the parameter a is
determined as 0.91 and b is 0.12.
2) Input Operating Range
It is important to determine the operating range of the
maximum number of Subscribers (Smax). The training data set
is used again to determine the range of Smax that provides the
desired Tref.
In order to achieve the desired value of the Tref, the
maximum number of subscribers will have to be adjusted. This
value of Smax again will change during runtime due to the
stochastic nature of the load and the controller is useful to
automatically adjust the Smax to meet the Tref. The details are
explained in the Section VI.A Implementation and Analysis
Modeling JMS Provider
B. Adaptive Controller
The adaptive controller is designed and implemented to
self-regulate the message throughput of the JMS Provider for a
pre-defined threshold of message throughput.
We implemented the adaptive control algorithm such that
any changes in the JMS Provider load can be well managed
such that the desired message throughput (Tref) is achieved at
any given point of time. The adaptive control has two different
parts.
Feedback Controller: The feedback controller is
reactive in nature and tunes the controller gain based
on the current measured message throughput, but
cannot handle the future load on the JMS Provider.
This runs a sub-control loop and at the end of each
such loop the controller parameter is tuned such that
the message throughput is in an allowed range of Tref
Predictor: In order to handle the future dynamics of the
loads on the JMS Provider, a predictor is used that
predicts the Smax and Tref. Based on these predicted
values the P-Controller Gain is tuned if predicted
desired message throughput is lesser than a pre-defined
error. We defined a parent control loop that runs
periodically. In each parent-control loop the Smax and
Tref are predicted. After each parent-control loop, the
predicted value of message throughput is compared
with the Tref. If the predicted value is less than Tref
within a pre-defined deviation then controller tunes the
Smax allowed, by adjusting the Controller gain (Kp)
such that subsequent loads on the JMS Provider meet
the Tref. We used the basic P-Controller [32] to tune
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
152 | P a g e
www.ijacsa.thesai.org
the value of Smax. If this deviation is greater than a
pre-defined threshold then it demands additional
resources, then the actuator triggers request to create a
new Virtual Machine.
Now we explain the two different parts of the proposed
Adaptive control solution, Feedback Controller and the
Predictor.
1) Feedback P- Controller
The JMS Provider during its operation will have varying
workloads that may affect its performance. In order to maintain
and regulate the performance in terms of Tref, the maximum
number of subscribers that can be allowed on the JMS Provider
will have to be tuned. We implemented a P-Controller [32] to
adjust the Smax during runtime such that JMS Provider
exhibits desired performance. The lower number of
subscribers will have the possibility of high Tref, but having a
too low value of Smax keeps the JMS Provider under-utilized.
In the section VI.A Implementation and Analysis Modeling
JMS Provider we have discussed optimal operating region of
Smax based on our experiments. In order to keep the desired
Tref, the P-Controller will tune the Smax in the operating
region. But there may be cases where the actual measure
Throughput (Ta) is much lower than Tref. In such scenarios,
the control law will trigger a request to provision additional
compute resource such as more compute power (e.g., Virtual
Machine). The Fig 8 below shows the P-Controller to tune the
Smax.

Figure 8. Feedback Control of JMS Provider
The output of the controller is represented by the equation
(5) below
u(t) = K
p
.E(t) (5)
Where
u(t) = The controller output
K
p
= Proportional Gain
The P-Controller Gain is represented in the equation (6) in
z-Transform
K
p
= E (z)/U (z) (6)

Figure 9. Feedback Control of JMS Provider in z-Transform
The Fig 9 shows the z-Transform [33] of the JMS Provider
adaptive loop control.
The equation (5) represents the controller output is. The
controller output, which is the new Smax becomes the control
input to the JMS Provider. The reactive feedback control runs
for every sub-control loop.
The JMS Provider is represented by G (z) is a first order
system as shown in the equation (7) below
G (z) = b/ (z - a) (7)
2) Predictor
The Predictor is a component in our proposed solution that
predicts the maximum number of subscribers for the future
periods. The Time-Series Triple Exponential Smoothing [34] is
used to predict the Smax that represents the possible future
maximum number of subscribers that could be allowed on the
JMS Provider based on the past history.
The smoothing technique has the ability to forecast up to
m periods ahead. It means that the maximum number of
subscribers that can be supported after m periods from the
current time can be predicted and hence the corresponding
Tref.
The Reference Message Throughput is predicted using the
predicted Smax and the previous value of the reference
message throughput. The equation (8) below shows how the
Tref is predicted
T
ref
(t + 1) = aT
ref
(t) + bS
max
(t + 1) (8)
In the Fig 6, we can notice that the Predictor accepts the
measured throughput (Ta), current Smax and outputs the
predicted Tref. (TrefPred) thus helps in tuning the Kp for the
future period.
C. Controller Algorithm
In this section we explain the controller algorithm
The following are the pre-conditions and Initialization
operations before the controller is executed
The JMS Provider model parameters a and b are
estimated
The parent-loop control and sub-control loop is
initialized
o Sampling time of sub-control loop = m
o Sampling time of parent-control loop = c
times of n
Determine the P-Controller Gain K
p

Initialize subscribers at the beginning = S
i

new VM triggering actuating signal message
throughput threshold = NT
h

Error Threshold range to tune the K
p
= E
r
,
min
, E
r
,
max

Parent-control loop execute threshold for message
throughput = P
T

o During running the sub-control loop if the
message throughput , when T
a
<= P
T
then the
parent-control loop is triggered
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
153 | P a g e
www.ijacsa.thesai.org
ALGORITHM
i. Start the JMS Provider
ii. Start loading the JMS Provider with initial number of
subscribers as S
i

iii. for every m units of time run the sub-control loop as
shown below
a. measure the actual message throughput (T
a
)
b. If T
a
is observed to be less than T
ref


for more than 4
times, then trigger the parent-control loop (step iv.)
c. Get T
a
, compute the Error e
i. If e is between E
r
,
min
and E
r
,
max
where T
a
< T
ref
, then
adjust the P-Controller Gain K
p
to meet T
ref

d. Repeat the steps from a. to d. for m times
iv. For every c x m units of time run the parent-control
loop (i,e for every c sub-control loops)
a. Define the prediction period p determines the number
of parent control loops from the current parent control
loop)
b. Compute or predict S
max
for p periods in advance S
max,
p

c. Compute or predict T
ref
for p periods in advance T
ref, p

d. Feed the predicted values to feedback controller
v. The controller will compare the T
ref, p
and the current
message throughput T
a.

a. If the T
ref, p
is more by NT
h
than T
a.,
then trigger the
actuator to provision new Virtual Machine
b. If the T
ref, p
is less by E
r
,
max
than T
a.,
then tune the P-
Controller Gain K
p

VI. IMPLEMENATATION AND ANALYSIS
The adaptive control discussed is implemented in Java
using an experiment data collected on Apache JMS Provider
ActiveMQ [35] running on Ubuntu Linux 10.04 , i5 Intel 2
GHz CPU, 4 GB RAM, 1 TB Hard disk. A sample custom JMS
application is run to generate the experiment data. A single
JMS topic and a single publisher are used. The subscribers are
increased which read different messages from the JMS Topic
that are published. The data is collected on the explained
experimental setup.
Then the proposed solution is run offline on the
experimental data to examine the improvement in the message
throughput, without running the proposed controller on the
ActiveMQ server online.
The following are the different steps performed for
implementing and evaluating the performance of the proposed
solution.
The model parameters (as in Equation (7)) are estimated
with two different data sets. The parameters with least
error are identified and used for the controller
Operating range of maximum possible number of
subscribers is determined for best possible message
throughput, which is between 60 to 90 subscribers
Based on the operating range, the P-controller Gain (K
p
)
is calculated as 2.67 and the Reference Message
Throughput (T
ref
) is determined as 220.
The Feedback Controller and Predictor are implemented
based on the values of P-controller Gain (K
p
) and
Reference Message Throughput (T
ref
). The improvement
in the message throughput using the proposed controller is
evaluated in comparison with the actual message
throughput.
We explain the implementation details of modeling the
JMS Provider, the controller and discuss the results below.
B. Modeling JMS Provider
The model parameters a and b of the Equation (7) are
estimated using the ARMA model where the actual message
throughput is measured by linearly increasing the number of
subscribers, and predicting the Message throughput. The error
percentage is computed between the measured throughput and
the predicted throughput. The experiments are run with two
different data sets. The Table I shows the estimated model
parameters for both the data sets with their error. We observe
that the values a = 0.91 and b = 0.12 proved to have a lesser
percentage of prediction error.
TABLE I. MODEL PARAMETER ESTIMATION
Data Sets
Model Parameter Estimation
a b
Percentage of
Error
Data Set 1 1 0.28 9.12
Data Set 2 0.91 0.12 8.33

Now using these constants the JMS Provider model in z-
Transform is represented as the equation (9) below, using the
model parameters estimated.
G (z) = 0.12/ (z 0.91) (9)
The Fig 10 and Fig 11 show the parameter estimation with
actual message throughput (Ta) and the predicted throughput
(Tpred). The message throughput in these figures is number of
messages per second. In Fig 10 the number of messages is
plotted against the increasing number of subscribers. There is a
saturation of message throughput after a certain number of
subscribers.
C. Adaptive Control
The Fig 12 below shows the performance evaluation of the
message throughput without Controller and with adaptive
controller proposed in this paper. We observe that the message
throughput using proposed Controller is better by about 25 %
which is a significant improvement in message throughput over
the throughput without controller. We can notice that there are
spikes where there is a sudden increase of the number of
subscribers. The actual message throughput has reduced
suddenly in such cases, but using a P-Controller tuning along
with the predictor, provided the adaptive control and has
regulated the throughput to be in the operating range between
200 and 250.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
154 | P a g e
www.ijacsa.thesai.org

Figure 10. Model Parameter Estimation Data set 1

Figure 11. Model Parameter Estimation Data Set 2

Figure 12. Performance Evaluation of Message Throughput using Adaptive
Control
1) Feedback Controller
From the experimental data the value of P-Controller Gain
Kp is determined as 2.67. This is computed by adjusting the
value of Kp between 2 and 4 and the average Kp is computed.
The reference message throughput is computed from the
operating range average as 220. By tuning Kp the output of the
controller is adjusted which is nothing but the tuning of Smax
to obtain the desired reference message throughput. But our
implementation has shown that the value of Smax is typically
around 59 with a maximum value of 120. The optimal
operating range of Smax is 59 Smax 90. The Table II
shows the operating range limits of Kp and Smax.
2) Predictor
We implemented the Triple Exponential Smoothing
predictor using openforecast Java API [36]. The Table III
shows the different values chosen for Predictor.
TABLE II. OPERATING RANGES
P-Controller Gain (Kp) Range
Kp
Range
Smax
Range
Tref
2.67, 3,
2, 2.4
59-90 220

TABLE III. PREDICTOR PARAMETER VALUES
Predictor Values
Triple Exponential
Smoothing
Coefficients
Er,max
Forecast period
(p)
0.2, 0.6, 0.6 70 1

The Fig 13 below shows the predicted values of the Smax
(Smaxp) and the Tref. (Trefpred). These values are predicted
using Triple Exponential Smoothing with coefficients shown in
the Table III. In our experiment the parent control loop is run
once the Trefpred starts decreasing less than 150, which is less
than the Tref by 70. From the Fig 13, the predicted Tref
(TrefPred) is less than 150, the P-Controller gain Kp is tuned
to a value of 4 such that message throughput is regulated
without fluctuations. The Er,max is set to as 70 ( 220-150). The
predictor adjusts the Kp once the Tref,pred is less by Er,max
(70) than original Tref. In our experimental data we didnt
simulate the condition of the Ta exceeding the threshold to
trigger addition Virtual Machine requests.

Figure 13. Prediction of Smax and Tref
VII. CONCLUSION
We observed using the P-Controller will have a distinct
improvement in the message throughput of the Enterprise
messaging servers. Our experiments are currently limited to
using the P-Controller only which helps in reducing the rise
time [37], but in order to obtain reduce the overshoot and
settling time using the PI-Controller [17] is more helpful.
Additionally, the parameter estimation is done on experimental
data and only two data set samples are used. Our results are
based on a simulation like environment as the P-Controller is
not directly verified online on the JMS Provider. Our
experiments are rather run on the data collected from the JMS
Provider by running a sample application with one publisher
and one JMS topic. We infer that applicability of adaptive
control systems will have significant improvement on the
performance of the Enterprise messaging servers in distributed
computing systems.
0
20
40
60
80
100
120
140
1 7 13 19 25 31 37 43 49 55 61
Tpred
Ta
0
50
100
150
200
250
300
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
1
Tpred
Ta
0
50
100
150
200
250
300
1 1019283746556473
Subscribers
Actual
Throughput
Controlled
Throughput
0
50
100
150
200
250
300
1 9 1725334149576573
Smaxp
Trefpred
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
155 | P a g e
www.ijacsa.thesai.org
VIII. FUTURE WORK
There is a scope of improvement of the solution explained
in this paper. We intend to extend the experiments to adjust the
model parameters during runtime such that model represents
the behavior of the system to be controlled in a real time. Also,
we want to examine the SASO [17] properties of our control
system to determine the controller stability and accuracy. We
also want to verify the solution on the ActiveMQ server with
varying publishers and topics, not limiting to the subscribers
only.
We suggest exploring a hybrid approach where techniques
like fuzzy control [38] can be used in conjunction with the
classic PI controller which can show better performance. The
applicability of fuzzy control enables creating a knowledge
base of rules and can be evaluated against using Triple
Exponential Smoothing for predicting future message
throughput. These rules can be helpful when the Enterprise
Messaging servers are used in massively large distributed
computing systems. We are also studying the different aspects
of Data Mining which can be used to build novel prediction
algorithms there by the adaptive control system is more robust.
REFERENCES
[1] Message Oriented Middleware (MoM):
http://en.wikipedia.org/wiki/Message-oriented_middleware
[2] Matjaz B.Juric, S.Jeelani Basha, Rick Leander, Ramesh Nagappan,
Professional J2EE EAI, Shroff Publishers 2005
[3] Service Level Agreement , http://en.wikipedia.org/wiki/Service-
level_agreement
[4] Service Level Objective,
http://en.wikipedia.org/wiki/Service_level_objectives
[5] JMS Specification:
http://www.oracle.com/technetwork/java/javaee/tech/index.html
[6] JEE Specification:
http://www.oracle.com/technetwork/java/javaee/tech/index.html
[7] JMS Providers: http://en.wikipedia.org/wiki/Java_Message_Service
[8] A. Robertsson, B. Wittenmark, M. Kihl, and M. Andersson, Design and
evaluation of load control in web server systems, IEEE American
Control Conference, 2004
[9] JMS Performance Benchmarks :
http://www.codeproject.com/KB/showcase/PerformanceBenchmarks.as
px
[10] http://www.precisejava.com/javaperf/j2ee/JMS.htm#JMS111
[11] Bruce Snyder, Dejan Bosanac and Rob Davies, ActiveMQ In Action,
Dreamtech Press, 2011
[12] Pradeep Padala, Xiaoyun Zhu, Mustafa Uysal et al. Adaptive Control of
Virtualized Resources in Utility Environments. In the proceedings of the
EuroSys 2007
[13] Evgeny Dantsin, Thomas Eiter, Georg Gottlob, Andrei Voronkov,
Complexity and expressive power of logic programming, ACM
Computing Surveys 2003
[14] Karl J.Astrom and Bjorn Wittenmark, Adaptive Control, Pearson
Education, 2009
[15] Yan Liu, Ian Gorton, Performance Prediction of J2EE Applications
using Messaging Protocols, Proceedings of 2005 Symposium on
Component-based Software Engineering
[16] Michael Menth, Robert Henjes, Christian Zepfel, and Sebastian
Gehrsitz, Throughput Performance of Popular JMS Servers, ACM
SIGMETRICS '06
[17] Joseph L. Hellerstein, Yixin Diao, Sujay Parekh, and Dawn Tilbury,
Feedback Control of Computing Systems, John Wiley 2004
[18] Ying Lu, Avneesh Saxena and Tarek E Abdelzaher, Differentiated
Caching Services; A Control-Theoretical Approach, IEEE
International Conference on Distributed Sysytems, 2001
[19] Keqiang Wu, David J. Lilja, Haowei Bai, The Applicability of
Adaptive Control Theory to QoS Design: Limitations and Solutions,
IEEE Parallel and Distributed Processing Symposium, 2005
[20] Ying Lu, Tarek Abdelzaher and Gang Tao, Direct Adaptive Control of
A Web Cache System, Proceedings of the American Control
Conference, Denver, Colorado, 2003
[21] Giovanna Ferrari, Santosh Shrivastava,Paul Ezhilchelvan, An
Approach to Adaptive Performance Tuning of Application Servers,
IEEE International Workshop on QoS in Application Servers, 2004
[22] Seungwan Ryu, Chulhyoe Cho,PI-PD-controller for robust and
adaptive queue management for supporting TCP congestion control,
Simulation Symposium, 132 - 139 April 2004
[23] Ravi Kumar Gullapalli, Dr.Chelliah Muthusamy and Dr.A.Vinaya Babu
, Control Systems application in Java based Enterprise and Cloud
Environments A Survey , IJACSA, Volume 2, No 8, August 2011
[24] Ravi Kumar Gullapalli, Dr.Chelliah Muthusamy, Dr.A.Vinaya Babu and
Raj N. Marndi, A FEEDBACK CONTROL SOLUTION IN
IMPROVING DATABASE DRIVER CACHING, IJEST, Vol 3, No 7,
July 2011
[25] Dr. Wolfgang Winter , Applying control theory concepts in software
applications, http://www.theserverside.com/feature/Applying-
controltheory-concepts-in-software-applications
[26] Richard Monson-Haefel and David A.Chappell, Java Message
Service, OReilly 2001
[27] Single Input Single Output : http://en.wikipedia.org/wiki/Single-
input_single-output_system
[28] Virtual Machines : http://en.wikipedia.org/wiki/Virtual_machine
[29] Erwin Kreyzig, Advanced Engineering Mathematics, John Wiley and
Sons
[30] ARMA:
http://en.wikipedia.org/wiki/Autoregressive_moving_average_model
[31] MICHAEL L.JOHNSON and LINDSAY M.FAONT Parameter
Estimation by Least Squares Error :
http://mljohnson.pharm.virginia.edu/pdfs/174.pdf
[32] P-Controller, http://en.wikipedia.org/wiki/Proportional_control
[33] Z-Transform: Saed Vaseghi,
http://dea.brunel.ac.uk/cmsp/Home_Saeed_Vaseghi/Chapter04-Z-
Transform.pdf
[34] Triple Exponential Smoothing:
http://itl.nist.gov/div898/handbook/pmc/section4/pmc435.htm
[35] Apache ActiveMQ, http://activemq.apache.org/
[36] Openforecast API, http://openforecast.sourceforge.net/docs/
[37] Jinghua Zhong, PID Controller : A Short Tutorial, Purdue University,
2006
[38] Jan Jantzen, Design of Fuzzy Controllers, Tech report, 1988,
Technical University of Denmark
AUTHORS PROFILE
Ravi Kumar G is working as a Technical Expert in Hewlett-Packard.,
Bangalore, India. He obtained his M.Tech in Computer Science from Birla
Institute of Technology, Mesra,India. He is currently pursuing Ph.D from
JNTU Hyderabad,AP, India.
Dr.Chelliah Muthusamy is Academic Relations Head at Yahoo, Bangalore,.
He obtained his Ph.D from Georgia Tech and M.Sc(Engg) in Computer
Science from Indian Institute of Science(IISc), Bangalore India
Dr.A.Vinaya Babu is a Professor of Computer Science working as Principal,
JNTUH College of Engineering, JNTU Hyderabad, AP, India. He obtained his
Ph.D and M.Tech in Computer Science from JNTU, Hydreabad.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
156 | P a g e
www.ijacsa.thesai.org
Improved Face Recognition with Multilevel BTC
using Kekres LUV Color Space

H.B. Kekre,
Senior Professor
Computer Engineering Department
MPSTME, SVKMs NMIMS
Mumbai, India



Dr. Sudeep Thepade
Associate Professor
Computer Engineering Department
MPSTME,SVKMs NMIMS
Mumbai,India

Sanchit Khandelwal, Karan
Dhamejani, Adnan Azmi,
B.tech Students
Computer Engineering Department
MPSTME, SVKMs NMIMS
Mumbai, India


AbstractThe theme of the work presented in the paper is
Multilevel Block Truncation Coding based Face Recognition
using the Kekres LUV (KLUV) color space. In [1], Multilevel
Block Truncation Coding was applied on the RGB color space up
to four levels for face recognition. The experimental results
showed that Block Truncation Coding Level 4 (BTC-level 4) was
better as compared to other BTC levels of RGB color space.
Results displaying a similar pattern are realized when the KLUV
color is used. It is further observed that KLUV color space gives
improved results on all four levels.
Keywords- Face recognition; BTC; RGB; KLUV; Multilevel BTC;
FAR; GAR.
I. INTRODUCTION
The term face recognition refers to identifying and
verifying a face image. It is basically the process of classifying
a face as known or unknown, based on training set. While
humans can easily identify faces, it is a challenging task for
computer systems. The computer systems store the faces in
such way that the important contents of the face image they
store, can be used efficiently for recognizing the face.
There are many biometric systems such as finger prints,
voice, iris, face and retina. Among these face recognition turns
out to be the most effective system since it requires very less
human interaction [21, 22]. Researchers from the field of
biometrics, image processing, computer vision, pattern
recognition system and neural network give a lot of importance
to face recognition. It is the fastest growing biometric
technology [18]. Some of the applications of face recognition
include physical, security and computer access controls, law
enforcement [12, 13], criminal list verification, surveillance at
various places [15], forensic, authentication at airports[17], etc.
Many algorithms are used to make effective face
recognition systems. Some of the algorithms include Principle
Component Analysis (PCA) [2, 3, 4, 5], Linear Discriminant
Analysis (LDA) [6, 7, 8], Independent Component Analysis
(ICA) [9, 10, 11], Block Truncation Coding (BTC) [1, 15, 19,
22] etc.
The paper presents an approach to enhance the performance
of BTC based face recognition using KLUV color space.
Applying the technique described in [1], using KLUV
color, it is observed that KLUV out performs RGB color space
at each level of Multilevel BTC.
II. BLOCK TRUNCATION CODING
Block truncation coding (BTC) [1, 12, 13, 14] was
developed in the early years of digital imaging more than 29
years. It was first developed in 1979 for greyscale image
coding [14]. It is comparatively a simple image coding
technique. BTC has played a vital role in the history of digital
image coding in such a way that many advanced coding
techniques have been developed, based on BTC.
III. MULTILEVEL BLOCK TRUNCATION CODING [1, 13, 20]
The feature vector in this algorithm is calculated by using
Block Truncation Coding [12, 13 and 14]. In [1], BTC has
been implemented up to four levels on RGB colour space for
face recognition. The feature vector size at BTC-Level 1, BTC-
Level 2, BTC-Level 3 and BTC-Level 4 was 6, 12, 24 and 48
respectively. In the same way BTC on KLUV colour space is
implemented up to four levels for face recognition.
IV. KEKRES LUV COLOR SPACE
It was obvious to extend BTC to multi-spectral images such
as color images. Most color images are recorded in RGB space,
which is perhaps the most well-known color space.
KLUV color space [12] is a special case of Kekre
transform. Where L gives luminance and U and V gives
chromaticity values of color image. Positive value of U
indicates prominence of red component in color image and
negative value of V indicates prominence of green component.
Equation (1) gives the RGB to LUV conversion matrix
which indicates the corresponding L, U and V components for
an image from the R, G and B components.
[

] [



] [

]

(1)

The reverse conversion, that is from LUV color space to
RGB color space is given in Equation (2).
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
157 | P a g e
www.ijacsa.thesai.org

[

] [



] [

]

(2)


V. PROPOSED METHOD
For color space, for each BTC level; the feature vector for
the query image and database set is by using Multilevel BTC.
In each level of BTC, the feature vector of the query image
is compared with the feature vector of each image in the
training set. The comparison (Similarity measurement) is done
by Mean Square Error (MSE) given by equation 3.

( )

( )

(3)

Where,
I & I are two feature vectors of size M*N which are being
compared.
To assess the performance of the different BTC levels
based face recognition techniques, False Acceptance Rate
(FAR) and Genuine Acceptance Rate (GAR) are used.
VI. IMPLEMENTATION
A. Platform
The effectuation of the Multilevel BTC is done in
MATLAB 2010. It was carried out on a computer using an
Intel Core i5-2410M CPU (2.4 GHz).
B. Database
The experiments were performed on two face databases.
1) Face Database [16]
This database is created by Dr Libor consisting of 1000
images (each with 180 pixels by 200 pixels), corresponding to
100 persons in 10 poses each, including both males and
females. All the images are captured against a dark or bright
homogeneous background, little variation of illumination,
different facial expressions and details. The subjects sit at fixed
distance from the camera and are asked to speak, whilst a
sequence of images is taken. The speech is used to introduce
facial expression variation. The images were taken in a single
session. The ten poses of Face database are shown in Figure 1.
2) Our Own Database [1, 20]
This database consists of 1600 face images of 160 people
(92 males and 68 females).For each person 10 images are
taken. The images in the database are captured under numerous
illumination settings. The images are taken with a homogenous
background with the subjects having different expressions. The
images are of variable sizes, unlike the Face database. The ten
poses of Our Own Database are shown in Figure 2.

Figure 1. Sample images from Face database

Figure 2. Sample images from Our Own Database
VII. RESULTS AND DISCUSSIONS
False Acceptance Rate (FAR) and Genuine Acceptance
Rate (GAR) are standard performance evaluation parameters of
face recognition system.
The False acceptance rate (FAR) is the measure of the
likelihood that the biometric security system will incorrectly
accept an access attempt by an unauthorized user. A systems
FAR typically is stated as the ratio of the number of false
acceptances divided by the number of identification attempts.
FAR = (False Claims Accepted/Total Claims) X
100
(4)
The Genuine Acceptance Rate (GAR) is evaluated by
subtracting the FAR values from 100.
GAR=100-FAR (in percentage) (5)
In all 10000 queries (10 images for each of 1000 persons)
are fired on face database and 16000 queries (10 images for
each of 1600 persons) are fired on our own database. For each
query, FAR and GAR values are calculated for respective BTC
level based face recognition technique. At the end the average
FAR and GAR of all queries in respective face databases are
considered for performance ranking of BTC levels based face
recognition techniques.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
158 | P a g e
www.ijacsa.thesai.org
FAR and GAR are calculated for both RGB color space and
KLUV color space.
A. Face Database
To analyze the performance of proposed algorithm, 10000
queries are tested on the database. The feature vectors of each
image for all four BTC levels in RGB color space and KLUV
color space were calculated and then compared with the
database. The FAR and GAR values are calculated by
employing equations 4 and 5.

Figure 3. FAR values at different BTC levels of KLUV and RGB color
spaces for Face Database
Figure 3 gives the FAR values of the different BTC levels
based face recognition techniques tested on face database for
both RGB and KLUV color spaces. Here it can be seen that
the FAR values go on decreasing for each succeeding level of
BTC of respective color spaces. This shows that the accuracy
of face recognition increases with increasing level of BTC and
hence BTC-level 4 gives the best result with the least FAR
value in both the color spaces. Also the FAR values of KLUV
are lesser than the RGB as shown in the figure. Thus, it can be
concluded that the implementation of BTC levels based face
recognition techniques is better when applied in KLUV color
space.
Figure 4 gives the GAR values of the different BTC levels
based face recognition techniques tested on face database for
both RGB and KLUV color spaces. Here it is observed that
with each successive level of BTC the GAR values go on
increasing in respective color spaces and hence a BTC-level 4
gives the best result with the highest value in both the color
spaces. Also the GAR values of KLUV are greater than the
RGB as shown in the figure. Thus, it can be concluded that the
implementation of BTC levels based face recognition
techniques is better when applied in KLUV color space. For
optimal performance the FAR values must be less and
accordingly the GAR values must be high for each successive
levels of BTC. Thus, the performance of KLUV color space
for BTC levels based face recognition techniques is superior to
the performance of RGB color space for Face database.

Figure 4. GAR values at different BTC levels of KLUV and RGB color
spaces for Face Database
B. Our Own Database
In all 16000 queries were tested on the database for
analyzing the performance of the proposed BTC level based
face recognition algorithm for both RGB color space and
KLUV color space. The experimental results of proposed face
recognition techniques have shown that BTC level 4 gives the
best performance in respective color spaces. The efficiency of
the Multi-level BTC based face recognition increases with the
increasing levels of BTC.

Figure 5. FAR values at different BTC levels of KLUV and RGB color
spaces for Our Own Database
0.005
0.01
0.015
0.02
0.025
0.03
BTC Level
1
BTC Level
2
BTC Level
3
BTC Level
4
F
a
l
s
e

A
c
c
e
p
t
a
n
c
e

R
a
t
i
o

Considered BTC levels
K'LUV RGB
96.8
97.1
97.4
97.7
98
98.3
98.6
98.9
BTC Level 1 BTC Level 2 BTC Level 3 BTC Level 4
G
e
n
u
i
n
e

A
c
c
e
p
t
a
n
c
e

R
a
t
i
o

Considered BTC levels

K'LUV RGB
0.2
0.25
0.3
0.35
0.4
0.45
BTC Level
1
BTC Level
2
BTC Level
3
BTC Level
4
F
a
l
s
e

A
c
c
e
p
t
a
n
c
e

R
a
t
i
o


Considered BTC levels
K'LUV RGB
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
159 | P a g e
www.ijacsa.thesai.org
Figure 5 gives the FAR values of the different BTC levels
based face recognition techniques tested on Our Own Database
for both RGB and KLUV color spaces. The FAR values go on
decreasing for each succeeding level of BTC of respective
color spaces. Thus, when BTC based face recognition
techniques is applied on Our Own Database, it gives a result
similar to the Face Database; The BTC level 4 gives the best
result for respective color spaces and KLUV color space is
better than RGB color space for implementing this proposed
algorithm.

Figure 6. GAR values at different BTC levels of KLUV and RGB color
spaces for Our Own Database
Figure 6 gives the GAR values of the different BTC levels
based face recognition techniques tested on Our Own Database
for both RGB and KLUV color spaces. It can be seen from the
above figure that BTC-Level 4 has the highest GAR values and
hence it is better than other BTC-Levels. Also the GAR values
of KLUV color space are greater than RGB color space at all
the levels. Thus it can be concluded that the implementation of
BTC levels based face recognition techniques is better when
applied in KLUV color space.
As seen from the performance of both the databases it can
be concluded that the implementation of BTC based face
recognition techniques on KLUV color space is superior to
RGB color space.
VIII. CONCLUSION
As the Multilevel BTC yields a greatly reduced feature
space, this reduces the processing time required by the system.
Thus, this system can be implemented is real time applications
which generally require fast recognition time and have low
computation power. The proposed face recognition system
using Multilevel BTC has been tested using two face databases.
For experimental analysis in all 10000 queries are fired on Face
Database and 16000 queries on Our Own Database. The
average FAR and GAR values of these queries clearly indicate
that better performance is obtained when Multilevel Block
Truncation Coding is employed using Kekres LUV color
space than RGB color space for face recognition.
REFERENCES
[1] H.B.Kekre, Sudeep D. Thepade, Sanchit Khandelwal, Karan Dhamejani,
Adnan Azmi, Face Recognition using Multilevel Block Truncation
Coding International Journal of Computer Applications (IJCA)
December 2011 Edition.
[2] Xiujuan Li, Jie Ma and Shutao Li 2007. A novel faces recognition
method based on Principal Component Analysis and Kernel Partial
Least. IEEE International Conference on Robotics and Biometrics, 2007.
ROBIO 2007
[3] Shermin J Illumination invariant face recognition using Discrete Cosine
Transform and Principal Component Analysis 2011 International
Conference on Emerging Trends in Electrical and Computer Technology
(ICETECT).
[4] Zhao Lihong , Guo Zikui Face Recognition Method Based on
Adaptively Weighted Block-Two Dimensional Principal Component
Analysis; 2011 Third International Conference on Computational
Intelligence, Communication Systems and Networks (CICSyN)
[5] Gomathi, E, Baskaran, K. Recognition of Faces Using Improved
Principal Component Analysis; 2010 Second International Conference
on Machine Learning and Computing (ICMLC)
[6] Haitao Zhao, Pong Chi Yuen Incremental Linear Discriminant Analysis
for Face Recognition, IEEE Transactions on Systems, Man, and
Cybernetics, Part B: Cybernetics
[7] Tae-Kyun Kim; Kittler, J. Locally linear discriminant analysis for
multimodally distributed classes for face recognition with a single model
image IEEE Transactions on Pattern Analysis and Machine
Intelligence, March 2005
[8] James, E.A.K., Annadurai, S. Implementation of incremental linear
discriminant analysis using singular value decomposition for face
recognition. First International Conference on Advanced Computing,
2009. ICAC 2009
[9] Zhao Lihong, Wang Ye, Teng Hongfeng; Face recognition based on
independent component analysis, 2011 Chinese Control and Decision
Conference (CCDC)
[10] Yunxia Li, Changyuan Fan; Face Recognition by Non negative
Independent Component Analysis Fifth International Conference on
Natural Computation, 2009. ICNC'09.
[11] Yanchuan Huang, Mingchu Li, Chuang Lin and Linlin Tian. Gabor-
Based Kernel Independent Component Analysis on Intelligent
Information Hiding and Multimedia Signal Processing (IIH-MSP).
[12] H.B.Kekre, Sudeep D. Thepade, Varun Lodha, Pooja Luthra, Ajoy
Joseph, Chitrangada Nemani Augmentation of Block Truncation
Coding based Image Retrieval by using Even and Odd Images with
Sundry Color Space Int. Journal on Computer Science and Engg. Vol.
02, No. 08, 2010, 2535-2544
[13] H.B.Kekre, Sudeep D. Thepade, Shrikant P. Sanas Improved CBIR
using Multileveled Block Truncation Coding International Journal on
Computer Science and Engineering Vol. 02, No. 08, 2010, 2535-2544
[14] H.B.Kekre, Sudeep D. Thepade, Boosting Block Truncation Coding
using Kekres LUV Color Space for Image Retrieval, WASET
International Journal of Electrical, Computer and System Engineering
(IJECSE), Volume 2, Number 3, pp. 172-180, Summer 2008.
[15] H.B.Kekre, Sudeep D. Thepade, Image Retrieval using Augmented
Block Truncation Coding Techniques, ACM International Conference
on Advances in Computing, Communication and Control (ICAC3-
2009), pp. 384-390, 23-24 Jan 2009, Fr. Conceicao Rodrigous College
of Engg., Mumbai
[16] Developed by Dr. Libor Spacek. Available Online at:
http://cswww.essex.ac.uk/mv/otherprojects.html
[17] Mark D. Fairchild, Color Appearance Models, 2nd Edition, Wiley-
IS&T, Chichester, UK, 2005. ISBN 0-470-01216-1
[18] Rafael C. Gonzalez and Richard Eugene Woods Digital Image
Processing, 3rd edition, Prentice Hall, Upper Saddle River, NJ, 2008.
ISBN 0-13-168728-X. pp. 407413.S
45
50
55
60
65
70
BTC Level 1BTC Level 2BTC Level 3BTC Level 4
G
e
n
u
i
n
e

A
c
c
e
p
t
a
n
c
e

R
a
t
i
o

Considered BTC levels
K'LUV RGB
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
160 | P a g e
www.ijacsa.thesai.org
[19] Dr.H.B.Kekre, Sudeep D. Thepade and Shrikant P. Sanas, Improved
CBIR using Multileveled Block Truncation Coding, (IJCSE)
International Journal on Computer Science and Engineering Vol. 02, No.
07, 2010, 2471-2476
[20] Dr. H.B.Kekre , Sudeep D. Thepade and Akshay Maloo, Face
Recognition using Texture Feartures Extracted from Walshlet Pyramid
, Int. J. on Recent Trends in Engineering & Technology, Vol. 05, No.
01, Mar 2011.
[21] International Journal on Biometrics and Bioinformatics ,Volume 4 Issue
2 2010, http://www.slideshare.net/CSCJournals/international-journal-of-
biometrics-and-bioinformaticsijbb-volume-4-issue-2.
[22] Dr. H.B.Kekre , Sudeep D. Thepade ,Varun Lodha, PoojaLuthra, Ajoy
Joseph, Chitrangada Nemani,Performance Comparison of Block
Truncation Coding based Image Retrieval Techniques using Assorted
Color Spaces, (IJCSIS) International Journal of Computer Science and
Information Security,Vol. 8, No. 9,December 2010
AUTHORS PROFILE
Dr. H. B. Kekre has received B.E. (Hons.) in Telecomm Engineering from
Jabalpur University in 1958, M.Tech. (Industrial Electronics) from IIT
Bombay in 1960, M.S.Engg. (Electrical Engg.) from University of Ottawa in
1965 and Ph.D. (System Identification) from IIT Bombay in 1970 He has
worked as Faculty of Electrical Engg. and then HOD Computer Science and
Engg. at IIT Bombay. For 13 years he was working as a professor and head in
the Department of Computer Engg. at Thadomal Shahani Engineering.
College, Mumbai. Now he is Senior Professor at MPSTME, SVKMs
NMIMS University. He has guided 17 PhDs, more than 100 M.E./M.Tech and
several B.E./B.tech projects. His areas of interest are Digital Signal
processing, Image Processing and Computer Networking. He has more than
350 papers in National / International Conferences and Journals to his credit.
He was Senior Member of IEEE. Presently He is Fellow of IETE and Life
Member of ISTE Recently ten students working under his guidance have
received best paper awards and two have been conferred Ph.D. degree of
SVKMsNMIMS University. Currently 10 research scholars are pursuing
Ph.D. program under his guidance.
Dr. Sudeep D. Thepade has Received B.E.(Computer) degree from North
Maharashtra University with Distinction in 2003, M.E. in Computer
Engineering from University of Mumbai in 2008 with Distinction, Ph.D. from
SVKMs NMIMS (Deemed to be University) in July 2011, Mumbai. He has
more than 08 years of experience in teaching and industry. He was Lecturer in
Dept. of Information Technology at Thadomal Shahani Engineering College,
Bandra (W), Mumbai, for nearly 04 years. Currently working as Associate
Professor in Computer Engineering at Mukesh Patel School of Technology
Management and Engineering, SVKMs NMIMS (Deemed to be University),
Vile Parle (W), Mumbai, INDIA. He is member of International Advisory
Committee for many International Conferences, acting as reviewer for many
referred international journals/transactions including IEEE and IET. His areas
of interest are Image Processing and Biometric Identification. He has guided
five M.Tech. projects and several B.tech projects. He has more than 130
papers in National/International Conferences/Journals to his credit with a Best
Paper Award at International Conference SSPCCIN-2008, Second Best Paper
Award at ThinkQuest-2009, Second Best Research Project Award at
Manshodhan 2010, Best Paper Award for paper published in June 2011 issue
of International Journal IJCSIS (USA), Editors Choice Awards for papers
published in International Journal IJCA (USA) in 2010 and 2011.
Sanchit Khandelwal is currently pursuing B.tech. (CE) from MPSTME,
NMIMS University, Mumbai. His areas of interest are Image Processing and
Computer Networks and security. He has 1 paper in an international journal to
his credit.
Karan Dhamejani is currently pursuing B.tech. (CE) from MPSTME,
NMIMS University, Mumbai. His areas of interest are Image Processing,
Computer Networks and UNIX programming. He has 2 papers in an
international journal to his credit.
Adnan Azmi is currently pursuing B.tech. (CE) from MPSTME, NMIMS
University, Mumbai. His areas of interest are Image Processing and Computer
Networks. He has 1 paper in an international journal to his credit.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
161 | P a g e
www.ijacsa.thesai.org
Scenario-Based Software Reliability Testing Profile
for Autonomous Control System

Jun Ai
School of Reliability and System
Engineering
Beihang University
Beijing, China
Jingwei Shang
School of Reliability and System
Engineering
Beihang University
Beijing, China
Peng Wang
School of Reliability and System
Engineering
Beihang University
Beijing, China


AbstractOperational profile is often used in software reliability
testing, but it is limited to non-obvious-operation software such
as Autonomous Control System. After analyzing the autonomous
control system and scenario technology, a scenario-based profile
constructing method for software reliability testing is presented.
Two levels of scenario-based profile in the paper are introduced:
system level and software level, and the scenario-based profile
could be obtained through mapping them. With the method, the
testing data for software reliability testing could be generated.
Keywords- software reliability testing; scenario-based testing
profile; autonomous control system.
I. INTRODUCTION
Software Reliability Testing(SRT) is used to ensure and
verify software reliability requirement. The regular method is
to perform a random statistical testing which is based on
operational profile [1][2][3]. Operation profile delegates the
probability distribution of the software usage, and the testing
data could be generated by random sample with it[4][5][6]. But
operational profile is limited to the Autonomous control
system(ACS), in which the operation is not obvious.
Autonomous control system (ACS) usually has autonomy
as it could run without people. The missile and unmanned
plane are two typical examples [10]. It is a kind of non-
obvious-operation system. The traditional operational profile
could not be used in ACS directly. Consequently, it is
necessary to develop a profile constructing method for ACS.
This paper describes a scenario-based software reliability
testing profile (SRTP), which could solve the problem
mentioned above about ACS. At first, the feature of ACS and
the limitation of operational profile in SRT for ACS are
analyzed. And then the conception and constructing method of
scenario-based SRTP are presented combing with scenario
technique. Finally, the feasibility would be validated by an
instance.
II. SRTP REQUIREMENTS OF ACS
ACS does not have obvious operations. The running
process of ACS is affected by system interactive, environment
influencing and activity sequential. Therefore, there are some
new technical requirements for the SRTP of ACS.
A. System-interactive:
System-interactive means the running of ACS software is
affected by the cross-linking systems of ACS. At every
moment, the stimulated inputs from the other cross-linking
systems happen synchronously.
Different system interactions may cause different running
modes and the states of the software are possibly determined by
all inputs from the cross-linking systems. Consequently, all of
the possible systems and their running processes should be
presented in the SRTP.
B. Evironment-influencing:
Environment-influencing means ACS is sensitive to the
change of environment. In other words, different strategies
would be made along with the change by ACS. Autonomy
control means online perceiving situation, making strategies
and executing missions automatically according to definite
assignments and principles [9]. Accordingly, to truly simulate
software running processes, external environment in the life
cycle of ACS should be described in testing profile.
C. Active-sequential:
Active-sequential means the system running periods could
be apperceived by ACS in entire lifecycle, and all of the actives
are sequentially. While working, real-time mission evaluation
should be made, the mission in which period and the
accomplishments should be determined by ACS. In addition,
the next activity should be decided automatically too.
Consequently, testing profile should be able to describe the
mission periods and their time sequence.
According to these characteristics and SRT technical
requirements of ACS, this paper proposes a testing profile
constructing method based on scenario to support SRT for this
kind of system.
III. SCENARIO-BASED SRTP
At first, the concept of usage scenario for software is made
as follow:
usage scenario profile describes the possible time-series of
software states and environments, which occur in the running
process of software and system. For embedded software, the
usage scenario profile should include both the active states of
software and the environmental changes of embedded system.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
162 | P a g e
www.ijacsa.thesai.org
Therefore, for the software of ACS, the usage scenario and
its corresponding probability constitute its scenario-based
SRTP. And the scenario-based SRTP can be express as:
SP={<Sys, So>} (1)
Sys={Sc1, Sc2, Sc3SciScn} (2)
Sc={<Pi, Fj>;i=1,2,3,,m; j=1,2,3,,n } (3)
P={<Ti, pi>,i=1,2,3,,n} (4)
So={< Ai, pi, Lj,>} (5)
A={<IiOi, pi>} (6)
In the formulasSP refers to the scenario-based SRTP; Sys
refers to the system under test(SUT); So refers to the software
under test; Sc refers to cross-linking system; P refers to
parameter of the interactive system; F refers to function of the
interactive system; T refers to equivalence class; A refers to
activity; L refers to layer of the activity; I is a input of the
activity; O refers to output of the activity; p refers to the
corresponding probability.
As mentioned above, a complete usage scenario profile
should include descriptions of both the systems and softwares
running process and their statistical distribution. The
constitution of usage scenario profile is shown in Figure 1. :
Usage
Scenario
System
Scenario
Software
Scenario
System
Process
Associate
Systems
Software
Activities
Data
Scenario Mapping

Figure 1. Constitution of usage scenario profile
In other words, the usage scenario profile consists of
system scenario profile and software scenario profile. System
scenario profile is used to describe the running process of ACS
and its statistical distribution, including system running states,
environments, and external driving sources. Software scenario
profile is used to describe software running states and possible
inputs and their distribution. System scenario profile is looked
as the constraints of software scenario, which define the
possible running condition of the software; while software
scenario is the detail representation of system scenario. The
combination of the two scenario profiles performs an
expression of the software usage.
IV. SCENARIO-BASED SRTP CONSTRUCTION
According to the analysis of scenario-based SRTP, the main
constructing process can be shown in Figure 2. .
System Scenario Software Scenario
Mapping
Profile Construction Testing Data
Limitetion
Manifestations
Property/Function
Software Activities
Software Info
Independent Scenario
Information
Model Tranceformation
Scenario Construction

Figure 2. The constructing process of scenario-based SRTP
At first, the usage scenario should be analyzed and modeled
in system level and software level. In this way, the system
scenario profile and software scenario profile could be
constructed. And then the mapping from system scenarios to
software scenarios according to their properties would be
made. In this step, independent system scenario should be
extracted from the system scenario profile, which would be
used to analyze software scenario profile. After the analysis,
the input and output during the software process could be
obtained, and the corresponding SRT data could be generated.
During the constructing process, some UML standard
diagrams could be used to describe the usage scenario profile.
The details are introduced as follows:
A. System Scenario profile Construction
Scenario-based SRTP includes description of the software
under test and associate systems during the software process.
As a top-to-bottom constructing method, analysis and
construction of the associate systems should be done firstly.
The main purpose of system scenario profile construction is
describing systems states and environments during the
systems running process and all the factors which could affect
the running of system.
Commonly, there are four steps in the system scenario
profile construction:
1) Define associate systems: including the SUT, associate
system, user, and enviroment;
2) Define external interactions of the systems: including
the interactions between SUT and other systems;
3) Define the running input data: including the input data
lead to the change of the state of SUT;
4) Define the factors and their distribution: including all
the factors which could affect the running of system
and the occurrence probability of the factors.
The Class Diagram is used to describe systems states and
environments. The variables in the class are the data which
could be exchanged among the SUT and the state of system is
decided by the variables, associate system, user and
environment. The operations in the class could be used to
describe the relationship among the variables, for example:
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
163 | P a g e
www.ijacsa.thesai.org
Data A is twice of Data B. An example of system and its
interactions is expressed as Figure 3.
+OperationA()
User
-DataA
+OperationA()
+OperationX()
System Under Test
-DataA
-DataX
-Data1
+OperationX()
Associate System
-DataX
Enviroment
-Data1
* * * *
*
* *
*

Figure 3. System and its interactions
In the system scenario, activities of users, interactions of
cross-linking systems, effect of environment, and other
influence from the SUT should be described. The factors which
could affect the running of system should be listed and possible
value range or value trend of the factors should be given. The
factors and its value distribution could be expressed with a
system scenario profile table, such as TABLEI.
TABLE I. EXAMPLE OF FACTOR TABLE
Factor Value
range
Probability Source Target .
Altitude 1~100 0.3 Altitude
meter
System
100~1000 0.7 Altitude
meter
System
Speed <20km/h 0.4 Speed
meter
System

B. Software Scenario Profile Construction
The main purpose of this period is to construct the scenario
model of software running process and its input infomation.
1) Activity dividing: dividing task of the software
according to the main purpose of each activity.
2) Use case modeling: constructing functional use case
models for the software according to the mission and
process.
ActorA
Usecase1
Usecase2
Package
ActorX
*
*
*
*
*
*
*
* *
*
*
*

Figure 4. Use case model example of software function
As shown in Figure 4. , use case model (which is expressed
by Use case Diagram) is used to analyze functional
requirements of the software. However, the detailed description
for each function is not necessary.
3) Scenario flow analyzing: software scenario can be
determined by a path through use cases. A path starts
from a use case and ends with a use case, traversaling
through all of the basic-flow and alternative-flow.
State 1 State 2
/ Basic flow / Basic flow / Basic flow
/ Alternative flow

Figure 5. Software scenario flow example
As shown in Figure 5. , software scenario flow could be
expressed by State chart Diagram. Each path describes a basic-
flow or an alternative-flow with an arrow-line.
Basic-flow: an arrow-line which is the simplest path
through the regular use cases; in this path the program should
run from beginning to the end with no error.
Alternative-flow: an arrow-line which starts from a basic-
flow or an alternative-flow in some specific conditions, and
rejoin the basic-flow or ends with a use case.
Combining the basic-flow and alternative-flows, a scenario
table can be obtained, such as TABLE II.
TABLE II. SOFTWARE SCENARIO FLOW
S 1 BF
S 2 BF AF 1

S 8 BF AF 3 AF 4
4) Scenario flow Refine: analyzing the basic-flow and
alternative-flows according to the stimulation and
sources, the equivalence class of inputs could be
divided.
UserA System SystemX
M1
M2
M3
M4

Figure 6. Software scenario flow refine example
It could be expressed by Sequence Diagram, as shown in
Figure 6. The arrow-line in the diagram presents messages and
data in the interactions.
5) Scenario Splicing: A software scenario with inputs and
output could be obtained by splicing sequence
diagrams in the scenario flow refine process
according to the orders in the basic-flow or alternative-
flow.
Generally, a software scenario is an instance of the use
case. The software scenario flows through some use cases,
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
164 | P a g e
www.ijacsa.thesai.org
defines the relevant data and covers the use case flows through
the path. Data associates with specific scenarios by means of
input/output and some intermediate state. There would be
multiple use cases in a complex scenario. The operations
between use cases are described by execution orders,
controlling conditions, parallels, or loops.
C. Scenario Mapping
System scenario and software scenario describe the running
process of the software under test in macro and micro view.
Using them without connections could not satisfy the testing
requirements mentioned before.
Decision table could be used to map system scenarios with
software scenarios. A decision table example is shown in
TABLE III. , where each row represents software scenario and
each column represent a system scenario. Each software
scenario should contain the scenario ID, conditions (or states),
all data and the expected results elements involved. The
decision table should at least include necessary system scenario
information in executing a software scenario.
TABLE III. SCENARIO MAPPING DECISION TABLE EXAMPLE
S ID S Data X1 Data Y1 Active A1 Result
S1. BF V V V A1
S2. BF-AF2 V V V
Active1-
Active2
S3. BF-AF I V n/a
Active1-
Active3-
Active1
Here V means the basic-flow could be executed only when
the condition is valid; the I means in the invalid condition the
alternative-flow would be activated; and n/a means the
condition is not applicable to this scenario.
By scenario mapping, the activities in software scenario
could be Modular with details, and a scenario combing system
scenario with software scenario could be obtained, which could
be used to generate software reliability testing data.
D. Distributing probability
By distributing probabilities to each data, activity, and
branches in the scenario diagrams and tables, a scenario profile
with both scenarios and probabilities is produced.
Then we could obtain software reliability test cases by
extracting a certain scenario and necessary data in describing
scenario from the profile.
Therefore, scenario-based SRTP construction is an iterative
process of refinement. Describe system and software behaviors
in different points of view.
V. INSTANCE VALIDATION
An ACS software has been taken as an instance, which is a
flight control system for a model airplane. In the issue, we
simplified the detail data of the system. The scenario-based
SRTP is constructed as follow:
A. System Scenario Profile Constructing
By analyzing the embedded ACS, the system scenario
contains the system under test, seeker, fuse, environment and
the carrier aircraft.
+Send Task()
+Guidance Revise()
-Task Info
-Guidance Info
Carrier
+Seeker Control()
+Fuse control()
+Guidance revise()
-Seeker Signal
-Fuse Signal
-Guidance data
SUT
+Guide()
-Guidance Data
Seeker
+Detonate()
-Fuse Signal
Fuse
-Distance
-Hight
Enviroment
* * * *
* *
*
*
*
*
*
*

Figure 7. ACS system scenario construction diagram
We analyzing the input/output and activity of each object,
and construct a system scenario as Figure 7, and the factor table
is as TABLE I. ACS system scenario contains description of
both objects and their factors.
B. Software Scenario Profile Construction
Construct the functional use case model as Figure 8. :
Carrier
Guidance recise
Seeker Control
Con-
Flight
Enviroment
*
* *
*
*
* *
*
Fuse Control
System Under Test
Seeker
Fuse
*
*
*
*
*
*
*
*

Figure 8. ACS functional use case diagram
A complex use case sequence contains guidance revise,
seeker control, and fuse control is selected as an example for
scenario flow analyzing, as in Figure 9. :
Guidance Revise
Send Fuse Control Signal
Target
Intercepted
Fuse Control
Whether be in Fuse
Working Distance

Figure 9. ACS scenario flow diagram
Form the system scenario flow table:
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
165 | P a g e
www.ijacsa.thesai.org
TABLE IV. SIMPLIFY ACS SCENARIO FLOW TABLE
S 1 Guidance Revise
S 2 Guidance Revise Seeker Control Target Lost


The refined scenario flow sequence diagram will be express in
the following process.
C. Scenario Mapping
Decision table based for scenario mapping can be
demonstrated as TABLE V. :
TABLE V. SIMPLIFY ACS SCENARIO MAPPING DECISION
TABLE
S ID Guidance
Revise
Data
Seeker
Working
Distance
Fuse
Working
Distance
Fuse Result
S1. V I n/a n/a Guidance
Flight

S4. V Seeker
Control
Fuse
Control
S 4 Guidance
Revise
D. Scenario Profile and Test Data
By distributing probabilities to each data, activity, and
branches in the scenario, the scenario profile with scenarios and
probabilities is generated. Then a scenario such as figure 12
could be extracted.
Enviroment ACS Seeker
Guidance
Guidance Signal
Seeker Signal
Seeker Control
Fuse
Guidance Revise
Seeker Signal

Figure 10. scenario sequence diagram
According to the system scenario factors in the scenario
mapping table, the properties of system scenario are
determined. By make a secant ling through the ACS scenario
sequence diagram, data across the secant line is the test data in
the very scenario.
A scenario-based SRTP constructing process for the ACS is
demonstrated above. In particular, some data are simplified in
the process.
VI. SUMMARY
A scenario-based SRTP constructing method is proposed in
this paper. With system scenario profile and software scenario
profile, the possible running state and usage of the SUT could
be modeled.
TABLE VI. SIMPLIFY TEST CASE
Test case Id: T01-01
Input:
1) Input Guidance Data XXXXXX
2) In Seeker Working Distance, Input Seeker Control
Signal XX
3) Seeker Guidance Data XX
4) In Fuse Working Distance, Input Fuse Control
Signal XX
5) Input Fuse Signal XX
According to scenario profile, the software reliability test
case could be generated. The problem that traditional operation
profile could not be used in ACS has been resolved. With the
instance of a flight control system for model airplane, the
method in this paper is effective.
REFERENCES
[1] Lu Minyan, Chen Xuesong,Software Reliability Testing and
Practice.Testing and Control Technology, vol. 19, 2000,pp. 509-512.
[2] Musa J D.Operational profiles in software reliability engineering.
IEEE Software , vol.10(2) ,1993, pp.14-32
[3] Ai, Jun ,Lu, Min-Yan, The analysis and modeling for the input space of
real-time embedded software, 8th International Conference on
Reliability, Maintainability and Safety, 2009, pp. 774 - 777
[4] Elbaum, S.; Narla, S. A methodology for operational profile
refinement, Reliability and Maintainability Symposium ,2001, pp. 142 -
149
[5] Saravana Kumar. K., Misra. R.B.and Goyal, N.K., "Development of
Fuzzy Software Operational Profile",Secure System Integration and
Reliability Improvement,Second International Conference on 2008,
pp.195 196
[6] Gittens, M.,Lutfiyya, H.and Bauer, M."An extended operational profile
model",15th International Symposium on Software Reliability
Engineering, 2004, pp.314 - 325.
[7] Xiaofeng L,Chenggang B and Liang S,"Software Operational Profile
Modeling and Reliability Prediction with an Open Environment ",10th
International Conference on Quality Software (QSIC), 2010,pp.227 -
231
[8] Urem, Frane, Mikulic and Zelimir," Developing operational profile for
ERP software module reliability prediction ",2010 Proceedings of the
33rd International Convention , pp. 409 - 413.
[9] Huang Yongfei, Peng Xinjie, Missle Flight Control System Testing ,
Missle and Guidance,Vol.29,2009, pp. 65-67.
[10] Ma Sasa,Chen Zili, Li Bing UAV Flight Control Software Testing
Technique , Radio Engineering,Vol.34(10),2004. pp. 53-57

AUTHORS PROFILE
Jun Ai is the deputy director of Center of Software Dependability in
School of Reliability and System Engineering, BeiHang University, China. He
received a PHD in system engineering from BeiHang University. His research
interests include software reliability and safety, software testing and software
engineering.
Jingwei Shang and Peng Wang are graduate students in School of
Reliability and System Engineering. Their research interests are about
software reliability testing and evaluation.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
166 | P a g e
www.ijacsa.thesai.org
Identification of Critical Node for the Efficient
Performance in Manet

Shivashankar
Asst.Professor
Medical Electronics
Dr.Ambedkar Institute of
Technology,
Bangalore 560 056, India

B.Sivakumar
Professor&HOD
Telecommunication Engg,
Dr.Ambedkar Institute of
Technology

G.Varaprasad
Associate Professor
Computer Science Engg
B.M.S.College of Engineering
Bangalore 560 019, India.



Abstract This paper considers a network where nodes are
connected randomly and can fail at random times. The critical-
node test detects nodes, whose failures are malicious behavior,
disconnects or significantly degrades the performance of the
network. The critical node is an element, position or control
entity whose disruption, is immediately degrades the ability of a
force to command, control or effectively conducts combat
operations. If a node is critical node, then more attention must be
paid to it to avoid its failure or removal of a network. So how to
confirm critical nodes in the ad hoc network is the premise to
predict the network partition. A critical node is the most
important node within the entity of a network. This paper
suggests methods that find the critical nodes of a network based
on residual battery power, reliability, bandwidth, availability and
service traffic type. The metrics for evaluation has been
considered as packet delivery ratio, end-to-end delay and
throughput.
Keywords- Critical node; malicious; residual battery power;
reliability; bandwidth; Mobile Ad hoc Network.
I. INTRODUCTION
Nodes in Mobile Ad hoc Networks (MANETs) are battery
powered and hence have limited lifetime. Due to excessive
utilization a node may die which can result in energy depletion
problem and thereby affect overall network performance. Local
rate of energy consumption based on application can be
monitored. Thus early detection and avoidance of energy
depletion problem is possible based on monitoring of power
and remaining battery lifetime. As and when the rate of
consumption based on the application returns to efficient levels,
the mobile relay is released so that it can be made available for
other critical nodes. This technique can be applied to any
protocol used in MANETs such as DSDV, DSR, AODV,
TORA etc.
In this work, we focus on maximizing avoidance of
network partition, rather than after the network partition, then
taking some remedial measures. We describe the critical nodes
compensation approach to increase the network connectivity
and improve packet delivery rate. First, we detect the critical
nodes which may lead to the network partitioning. Secondly,
propose the compensation algorithm to avoid the network
partitioning. This approach described in this paper is built
around the notion of a critical node in an ad hoc network. Our
definition of a critical node is a node, whose failure or
malicious behavior disconnects or significantly degrades the
performance of the network. Once identified, a critical node
can be monitored in terms of resource. If a node is not
considered critical, this metric can be used to decide if the
application or the risk environment warrant the expenditure of
the additional resources required to monitor, diagnose, and alert
other nodes about the problem. Determining the global network
topology in a MANET gives the time delays of the diagnostic
packets and the mobility of the nodes makes this task futile, but
determining an approximation of this topology or subset of this
topology, within a certain time frame may be useful.
An approximation of the network topology can still provide
useful information about network the density, network
mobility, critical paths, and critical nodes. Even with the
uncertainty associated with correctly reconstructing the
network topology for a given time period, this additional
information can help to reduce the resources consumed to
monitor all nodes in the absence of this information.
The node performing the test is referred to as the testing
node and the node being tested is referred to as the node under
test. Three steps are required to detect whether a testing node
shares a critical link with its neighbor. First step is to
temporarily modify the testing nodes routing table to allow
only one communication link to be operational at a time, while
blocking communication through all others. The enabled
communication link will be between the testing node and a
node other than the node under test. Each communication link
will be tested sequentially in this manner to determine if an
alternative path to the link under test exists.
The responsibilities of a routing protocol include
exchanging the route information, finding a feasible path to a
destination based on criteria such as hop length, minimum
power required and life time of the wireless link. Gathering
information about the path breaks, mending the broken paths
expending minimum processing power, bandwidth and
utilizing the minimum bandwidth. Fig.1 shows the graphical
view of identification of the critical node in the network.
If the critical node find in the network, then alternate path
will be selected for the efficient network. The major challenges
are:
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
167 | P a g e
www.ijacsa.thesai.org
A. Ability to Measure the Resource Availability
In order to handle the resources such as bandwidth
efficiently and perform call admission control based on their
availability, the MAC protocol should be able to provide an
estimation of resource availability at every node.
B. Capability for Power Control
The transmission power control reduces the energy
consumption at the nodes, causes a decreasing interference at
neighboring nodes and increases frequency reuse.
C. Mobility
One of the most important properties of MANET is the
mobility associated with the nodes. The mobility of the nodes
results in frequent path breaks, packet, collisions, transient
loops, stare routing information and difficulty in resource
reservation. A good routing protocol should be able to
efficiently solve all the above issues.
The rest of the paper is organized as follows. In section II,
some previous work related to this paper and some applications
are identified. In Section III, we present the different
mechanisms with algorithms and full connectivity when the
only randomness is due to random connections and there are no
node failures. Section IV contains the simulation results,
analyzing and comparison of different parameters and also it
includes the discussion on the performance by the various
protocols agents in our simulations. Finally, Section V
concludes the paper.

Figure 1. Critical Node representation in network.
II. SOME OF THE RELATED RESEARCH WORK
A. Critical Nodes Compensation Algorithm in Sparse Ad hoc
Wireless network
Sparse ad hoc wireless networks [1] are an active
application in recent years, though sparse environment often
makes the network unconnected. However, most of the
previous researches focus on connected networks where an
end-to-end path exists between any two nodes in the network,
therefore these technologies would not work well in sparse ad
hoc networks, where restricted node communication radius can
result in periods of intermittent connectivity. In this work, we
develop a critical nodes compensation algorithm in order to
prevent network from partitioning, thereby insure network
connectivity and throughput.
But the drawback of previous studies is that they
overlooked the importance of the service traffic flow [2] and
non-articulation nodes. We propose a method that evaluates
how important each node is based on service traffic flow and
articulation node. Based on this evaluation, we can give priority
to each node. In a complex network, it is useful for a network
administrator to focus mainly on some high-ranked nodes,
which are relatively important for security purposes.
B. Determining Node Priority Order Based on Traffic Volume
Assuming that there is a web server with a small amount of
traffic volume and a backup server with huge amount of traffic
volume, it is possible for the network administrator to value
the web server because it needs to be alive all the time
compared to backup server. In this case, the web server node is
more important in spite of its low traffic volume [5].
C. Depth First Search Algorithm(DFS)
DFS was used to detect critical nodes in [3]. DFS is a
centralized algorithm and can be also implemented in
globalized distributed manner. The algorithms in [4] require
that a node should be aware of global topology. In practice, this
method is inefficient and involves a quadratic (in number of
nodes) communication overhead. The node density is the
average number of neighbor nodes of a node in the network.
This value reflects the density of the network. The larger the
value is, the denser the network is [6]. In this paper, we present
a novel critical node detection algorithm for wireless ad hoc
networks. It is an effective distributed localized algorithm,
which greatly reduces communication overheads and the speed
of detection. This is a distributed algorithm which adapts the
dynamic topology adaptively, detects the critical node faster
and more reliably, and decreases the detection overheads
efficiently.
D. Bandwidth Balancing
The bandwidth balancing [7] solves the fairness problems
suffered by long bandwidth in the networks. By constraining
each node to take only a fraction of the available slots,
bandwidth balancing can achieve a fair operation point when
several nodes are performing large file transfers.
As routing protocols exchange routing data between nodes,
as a result, they would maintain routing status in each node.
Based on routing status, data packets are transmitted by
mediated nodes along an established route to the destination
[8]. M.K Rafsanjani, A Movaghar presents a scheme in which
nodes do not need to exchange multiple messages to prove their
identities [10]. However, most of the previous researches[11,9]
focus on connected networks where an end to end path exists
between any two nodes in the network, therefore these
technologies would not work well in sparse ad hoc networks,
where restricted node communication radius can result in
periods of intermittent connectivity.
As routing protocols exchange routing data between nodes,
as a result, they would maintain routing status in each node.
Based on routing status, data packets are transmitted by
mediated nodes along an established route to the destination
[12]. In [13], the authors describe a distributed intrusion
detection system for MANETs that consists of the local
components data collection, detection and response and of the
global components. Whereas their architecture is very
promising and similar to the one we use in our paper, they
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
168 | P a g e
www.ijacsa.thesai.org
neglect the aspect how their local data collection should find
out on incidents like dropped packets, concealed links, etc.
III. DESIGN AND IMPLEMENTATION
A. Bandwidth Constraint in MANET
Since the channel is shared by all nodes in the broadcast
region (any reason in which all nodes can hear all other nodes),
the bandwidth available per wireless link depends on the
number of nodes and a traffic they handle. Thus only a fraction
of the total bandwidth is available for every node. The control
over head involved must be kept as minimal as possible.
Bandwidth efficiency can be defined as the ratio of the
bandwidth used for actual data transmission to the total
available bandwidth. The limited bandwidth availability also
imposes a constraint on routing protocols in maintaining the
topological information. Due to the critical nodes, maintaining
consistence topological information at all the nodes involves
more control over head which, in turn, results in more band
width.
Available bandwidth calculation for the MANET:
Sufficient Bandwidth = BW
available
=

W ------(1) (1)
T=Sampling time window for calculating real time bandwidth.
T
idle
= Time period when a mobile node in idle mode.
W= Maximum bandwidth for data transmission.


B. Location Dependent Contention
The load on the wireless channel varies with the number of
critical nodes present in a given geographical region. This
makes the contention for the channel high when the number of
nodes increases. The high contention for the channel results in
a high number of collisions and a subsequent wastage of
bandwidth. A good routing protocol should have built in
mechanism for distributing the network load uniformly across
the network so that the information of regions where channel
contention is high can be avoided.
C. Quick Route Reconfiguration
The unpredictable changes in the topology of the network
require that the routing protocol be able to quickly perform
route configuration in order to handle path break and
subsequent packet losses because of critical nodes in the
network.
D. Loop Free Routing
This is a fundamental requirement of any routing protocol
to avoid unnecessary wastage of network bandwidth. In
MANET, wireless network due to the random movement of
critical nodes, transient loops may form in the route. A routing
protocol should detect such critical nodes and take corrective
actions.
E. Signal Strength Based Reliability
A node, N can measure signal strength of its active
neighbors. The received signal strength is measured at physical
layer and made available to the access of top layers.
R(node)=P(new)/[P(old)+P(new)] (2)
Where R(node) denotes node reliability and it will be a
value between 0 and 1. Whatever new signal strength is more
than before, R(node) approaches towards one and whatever
new signal strength is less than before, R(node) approaches
towards zero.
F. Reliability
The reliability of the real time data being transmitted can be
enhanced by introducing a buffer of a fixed size calculated
through monitoring the data transmitted by employing self-
healing nodes.
G. Residual Battery Power
Power conservation in wireless ad hoc networks is a critical
issue as energy resources are limited at the electronic devices
used. Therefore to conserve battery energy of the nodes, there
are various routing algorithms and schemes designed to select
alternative routes. These algorithms and schemes are
collectively known as power-aware routing protocols and an
example of a better choice of routes selected is one where
packets get routed through paths that may be longer but that
pass through nodes that have plenty of energy reserves. Some
or all of the nodes in a MANET may rely on batteries or other
exhaustible means for their energy. For these nodes, the most
important system design criteria for optimization may be
energy conservation.
H. Minimize Energy Consumed/Packet
This is one of the more obvious metrics. To conserve
energy, we want to minimize the amount of energy consumed
by all packets traversing from the source node to the
destination node. That is, we want to know the total amount of
energy the packets consumed when it travels from each and
every node on the route to the next node. The energy consumed
for one packet is thus given by the equation:

1
1
1
) , (
k
i
i i
n n T E
-------------- (3)

In equation (1), n1 to nk are nodes in the route while T
denotes the energy consumed in transmitting and receiving a
packet over one hop. Then we find the minimum E for all
packets. However, this metric has a drawback and i.e. nodes
will tend to have widely differing energy consumption profiles
resulting in early death for some nodes. Power aware routing
schemes make routing decisions to optimize performance of
power or energy related evaluation metric.
Excessively, conserving energy neglects power
consumption at individual nodes, which speeds up network
partition by draining batteries of the critical nodes in the
network one by one. In effect, it shortens the network lifetime.
On the other hand, overly conserving power expels energy
consideration, which commits to paths with large number of
hops and longer total distance. Consequently, the total energy
dissipated is high and on average, the battery power decays
faster. In effect, it also shortens the network lifetime. Energy
efficiency is also an important design consideration due to the
limited battery life of wireless node. Since, the network
interface is a significant consumer of power, considerable
research has been devoted to low power design of the entire
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
169 | P a g e
www.ijacsa.thesai.org
network protocol stack of wireless network in an effort to
enhance energy efficiency.
All mobile should drain their power at equal rate as a
minimal set of mobile exist such that their removal cause
network to partition. Such node is called as a critical node. The
route between these two partitions must go through one of
these critical nodes. A routing procedure must divide the work
among these nodes to maximize the life of the network. This
problem is similar to load balancing problem. A packet to be
routed through a path contains mobiles having grater amount of
energy though it is not a shortest path. Delay is minimized as
no congestion and nodes having less number of loads.
I. Availability
MANETs establishing trust relationships between the nodes
in a decentralized fashion has been an important research issue
for a long time. If the sender nodes accurately identify the
legitimate nodes in the network, a robust routing can be
provided while mitigating the effects of malicious nodes.
Further, there is always a mutual interaction between a sender
and its neighbor nodes during the communication. The scheme
guarantees the availability of message as long as a legitimate
path exists. Through simulations, we will show the efficiency
of the scheme with respect to latency, availability and energy
consumption in the presence of adversary.
J. Residual Capacity
Residual capacity at a node is the difference between the
nodes channel capacity and the sum of the bandwidth
consumed by all contending flows of that node. It denotes the
channel capacity that is not used and it constrains the rate that
contending flows can acquire. To measure the residual
capacity, each node in the network monitors the channel
activity. The fraction of channel idle time during the past
measuring period and the channel capacity, determine its
residual capacity. The residual capacity at node k can be
expressed as
R
k
= T
idle
(C
k
)

-----------------------(4)

T
P

In equation(4), Ck is the channel capacity at node k; Tidle
is the channel idle time during the last measuring period Tp.
Larger Tp will give more accurate channel view but also longer
response time for the source node to react to the change in the
network. In this paper, Tp is set to 0.5 second.
IV. SIMULATION SETUP AND RESULT DISCUSSION
Network Simulator (NS-2.33) has a very rich component
library, which is compiled of two languages: C++ or Python is
object oriented extension of TCL. First of all, we define the
simulation in the 1,000 m1,000 m region, random waypoint
mobile model, the node number: 65; node original
communication radius: 40m; according to the distance between
compensation node and critical node, choose compensation
transmission power; in order to maintain low-speed movement
environment, node movement speed is not more than 2 m/s.
We choose two ad hoc network performance parameters:
critical node number and network throughput. For our
simulation, we used NS-2.33 with mobility extension. These
extensions include the modeling of an IEEE 802.11/MAC.
Table 1 shows the simulation parameters used in the
network setup for identifying the critical node and select the
alternate path for maintaining the continuous efficient network
connection in MANET.
TABLE 1. SIMULATION PARAMETERS.
Simulation time 0-50 sec
Traffic type TCP
Packet size 1460
Hello packet interval 2 sec
Node mobility 0 to 10 mts/sec
Frequency 1 Ghz
Channel capacity 2 M bps
Transmit power 2.0 Mw
Receiver power 2.0 Mw
Total number nodes 65
Communication system MAC/IEEE 802.11G
TABLE 2. OVERALL NETWORK INFORMATION WITH CRITICAL NODE AND
WITHOUT CRITICAL NODE IN MANET.



Figure 2. Throughput with and without critical node in MANET.
0
5
10
15
20
25
30
500 700 900 1000 1200 1400
Throughput without critical node
Throughput with critical node
Packet size(Bytes)










T
h
r
o
u
g
h
t
p
u
t


(
k
b
/
s
)

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
170 | P a g e
www.ijacsa.thesai.org


Figure 3. Throughput of receiving packets when there is critical node occurs
during the transmission of data.

Figure 4. Throughput of receiving packets during the transmission of data.

Figure 5. Throughput of generating packets at destination node.

Figure 6. Throughput of dropping bits at destination node.


Figure 7. Average of end-to-end delay.
Figure-2 shows the throughput response of the MANET is
78% difference with and without critical nodes. In this paper,
the performance analysis of MANET for different scenario as
mention in Table-2 has been studied using trace graph. The
first part of the table-2 shows more dropping packets, more
number of lost packets in the presence of critical node as
intermediate node. The second part of the table-2 shows that
the improvements in MANET approximately 94% with more
received packets, less dropped packets and less number of lost
packets.
The figure-3 represents the throughput performance in
MANET. The Y axis shows throughput of receiving packets at
destination node in kilo bytes and the X axis shows simulation
time in sec. The red line of the graph represents the best effort
with 92% of the throughput. The green line of the graph shows
the throughput of critical node presence and it is approximately
52% of the throughput. We start the best effort approximately
started at 12 sec and going on until 50 sec. The presence of
critical node in MANET, results in large delay. Since the delay
has been increased, the throughput is gradually reduced. When
the neighbor nodes are communicating with each other, the
packets will move between source and destination and thus the
throughput will be increased.
Figure-4 plots the receiving packets during the transmission
time v/s simulation time after identified the critical node and
alternate path has been selected. It shows the sum of numbers
of all the intermediate nodes, receiving packets sent by the
source node and number of received packets at the destination
node. There is some packet losses (6% to 12%) for the rates of
78 to 90pkts/sec for both the schemes on some of the nodes.
This may be due to insufficient bandwidth, less residual battery
power and minimized energy consumed per packet. If these
metrics are good, then throughput of the network will be
increased.
Figure-5 shows the throughput of generating packets at any
intermediate nodes v/s simulation time (sec).The graph reflects
the simulation time for which an intermediate node in route
generated packets. In other words, it depicts how long the route
through the intermediate node was valid during the simulation
after selected alternate path for the best efficiency.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
171 | P a g e
www.ijacsa.thesai.org
The throughput for the proposed Networks as shown in
figure-6, are calculated based on the distance. The throughput
has dropped around 45.3%. The number of data packets
dropped at any critical node present in the network. This is an
important parameter because if the number of dropped packets
increases, the throughput would decrease. Therefore the lower
packets drop; lower would be delay in the network.
Figure-7 represents the average of end to end delay for both
the schemes (with and without critical node). Though the nodes
are at same distance from the route, they receive different
bandwidth. This is the average delay of all the data packets.
The delay different is 50% of the throughput of receiving bytes.
It is calculated as the time taken between the generation of data
packets and arrival of last bit of destination. There are possible
delays caused by bandwidth, throughput, average number of
node receiving delay between current and other node etc. This
metric describes the packet delivery time. The lower end to end
delay is the better application performance in MANET.
V. CONCLUSION
In this work, we develop a critical nodes compensation
algorithm in order to prevent network from partitioning,
thereby insure the network connectivity and throughput.
Finally, we give simulation design and analysis of the critical
nodes compensation algorithm using NS-2.33 model. Our
simulation results show that the algorithm can effectively
improve ad hoc networks performance. In addition, some time-
critical applications may not be able to function properly in
disconnected MANET as the end-to-end delay. However, there
are lacks of full proof in theory and in practice. In these
systems, nodes are subject to battery power, efficient
bandwidth constraints and good throughput. The overall
performance of the network is increased after detection of the
critical node and selecting alternate path for the continuation of
the same operation. By considering all the above metrics we
come to a conclusion that while a network is set up, each node
is leveled with a threshold value of energy. The nodes energy
is decreasing gradually as it is used in network connectivity.
So the failure of critical node depends on less bandwidth, less
residual battery power and poor throughput.
REFERENCES
[1] Shoudong Zou, Ioanis Nikolaidis and Janelle J. Harms ENCAST:
EnergyCritical Node Aware Spanning Tree for Sensor Networks, In
Proceedings of the Communication Networks and Services Research
Conference, pp.249-254, 2005.
[2] Michael Dion Building in Reliability (BIR) with Critical Nodes,
http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=00493599
[3] Massimo Franceschetti and Ronald Meester, Critical Node Lifetimes in
Random Networks via the ChenStein Method, IEEE Transactions on
Information Theory, vol.52, no.6, pp.2831-2837, June 2006.
[4] A. Karygiannis, E. Antonakakis, and A. Apostolopoulos, Detecting
Critical Nodes for MANET Intrusion Detection Systems, In
Proceedings of IEEE Workshop on Security, Privacy and Trust in
Pervasive and Ubiquitous Computing, pp.7-15, 2006.
[5] Vahid Tabatabaee and Leandros Tassiulas, MNCM: A Critical Node
Matching Approach to Scheduling for Input Buffered Switches with no
Speedup, IEEE/ACM Transactions on Networking, Vol.17, no.1,
pp.294-304, February 2009.
[6] Daisuke Kasamatsu, Norihiko Shinomiya and Tadashi Ohta, A
Broadcasting Method considering Battery Lifetime and Distance
between Nodes in MANET, IEICE Transactions on Information and
Systems, Vol. J91-B, No.4, pp.364-372, 2008
[7] Min Sheng, Jiandong Li and Yan Shi, Critical Nodes Detection in
Mobile Ad Hoc Network,
ieeexplore.ieee.org/iel5/10777/33944/01620401.pdf.
[8] N.Komnios, D.Vergados and C. Douligeries. Detecting unauthorized
and compromised nodes in mobile adhoc network. Elseviewer Adhoc
network,vol5,n0 3,pp.289-298,2007.
[9] Christian Bravo, Sonia Assa and Andre Girard, Providing Quality of
Service for Critical Nodes in Ad Hoc Networks, IEEE Vehicular
Technology Conference, 2005.
[10] M.K Rafsanjani, A Movaghar, Identifying monitoring nodes with
selection of Authorized nodes in mobile Adhoc network,World Applied
Sciences Journal,vol4,n03, pp.444-449,2008
[11] N.Komnios, D.Vergados and C. Douligeries, Detecting Unauthorized
and Compromised Nodes in Mobile Ad hoc Network, Adhoc
Network(Elsevier), vol.5, no.3, pp.289-298, 2000.
[12] Yongguang Zhang, Wenke Lee, and Yi-An Huang, Intrusion Detection
Techniques for Mobile Wireless Networks, ACM Wireless Networks,
vol.9, no.5, pp.545-556, September 2003.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
172 | P a g e
www.ijacsa.thesai.org
Secret Key Agreement Over Multipath Channels
Exploiting a Variable-Directional Antenna

Valery Korzhik, Viktor Yakovlev,Yuri Kovajkin
State University of Telecommunication
St. Petersburg, Russia

Guillermo Morales-Luna
Computer Science Department
CINVESTAV-IPN
Mexico City, Mexico


AbstractWe develop an approach of key distribution
protocol(KDP) proposed recently by T.Aono et al., where the
security of KDP is only partly estimated in terms of
eavesdropper's key bit errors. Instead we calculate the Shannon's
information leaking to a wire tapper and we also apply the
privacy amplification procedure from the side of the legal users.
A more general mathematical model based on the use of
Variable-Directional Antenna (VDA) under the condition of
multipath wave propagation is proposed. The new method can
effectively be used even in noiseless interception channels
providing thus a widened area with respect to practical
applications. Statistical characteristics of the VDA are
investigated by simulation, allowing to specify the model
parameters. We prove that the proposed KDP provides both
security and reliability of the shared keys even for very short
distances between legal users and eavesdroppers. Antenna
diversity is proposed as a mean to enhance the KDP security. In
order to provide a better performance evaluation of the KDP, it
is investigated the use of error correcting codes.
Keywords-wireless communication; wave propagation;
cryptography; key distribution.
I. INTRODUCTION
The problem of key distribution is still in focus of research
activity especially for wireless LAN systems. This is due to the
severe restriction of asymmetric (public key) cryptography
WLAN implementation entailing a lower processing speed.
In order to solve this problem, quantum cryptography [1]
which allows eavesdropping detection within the key sharing
procedure seems useful. However, this approach does not reach
a practical level due to many technical problems, such as the
requirement of special quantum devices. There are well known
key distribution protocols (KDP) based on the presence of
noise in both legal and illegal channels [2], [3], [4]. But even
though the eavesdropper's channel is less noisy than the legal
ones and the eavesdroppers is passive, it is necessary to have
the knowledge of the eavesdropper's noisy power in order to
guarantee a fixed level of key security. Unfortunately this
condition cannot be taken for granted because an eavesdropper
may be able to get some advantage at the cost of better receiver
sensitivity or a shorter distance of interception that it was
considered by legal parties in the design of the secure KDP.
The most basic assumption on the executed KDP is that the
legal and illegal users have different locations, and this fact has
to be verified by physical means. (For that matter, an existing
special zone surrounding each legal user shall be assumed
where the presence of an eavesdropper is not allowed.)
This assumption is sufficient for secure key distribution if
either the communication channel between the legal users and
the eavesdropper have random parameters or one legal user
generates some randomness, under the condition that this
randomness is transmitted to other legal users over multipath
channels and any eavesdropper is able to receive this
information only on a multipath channel, but with some other
parameters, due to different locations of the legal users and the
eavesdropper.
The first approach is considered in [5], [6] for multipath
channels with random parameters and in [7], [8] for ultra-wide
band channels with random pulse responses. But the
randomness exploiting of the fluctuation of channel parameters
is very questionable because there may be such channel states
in which a temporal variation of propagation characteristics is
slow and small. In order to take for granted some given
randomness level it would be better to create artificially this
randomness by means of legal users.
Let us consider the following mathematical model of the
channels between a source of randomness (the first legal user)
and both the second legal user and the eavesdropper:
where is the
vector randomness, is the coefficient vector of the
multipath propagation to the second legal user, and
is the coefficient vector of multipath propagation to the
eavesdropper. Let us assume for simplicity E()=0, then the
following relation for the correlation coefficient between and
results:

where R

is correlation matrix of the random vector . In a

general case (,) 1. Moreover if x and y are orthogonal,
(e.g.x, y=0) and R

=Id
m
is the (mm)-identity matrix, then
(, )=0.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
173 | P a g e
www.ijacsa.thesai.org
The key bits of the second legal user can be generated after
multiple repetition of the random independent vector () and
the binary quantization of the random values (). The first legal
user can form key bits in a similar manner after a signal
reception from the second user over the same multipath
channel with the same randomness. If the variables and are
Gaussian and non-correlated, then the shared key is provided
secure due to the statistical independence of the variables. (A
more general situation with non-zero correlation is considered
in Section II.)
Common randomness can result from fluctuation of the
cannel characteristics due to communicating object motion.
Such approach has been proposed in [9], [10], [11]. But it still
entails another problem: it is easy to break the secret key under
an environment with small fluctuation of the channel
characteristics or in the case when the communicating objects
are stopped. In order to overcome these defects, a more
sophisticated method, using smart antenna excited randomly by
electronic means [12], has been proposed (although their
results were obtained experimentally without an estimation of
information-theoretic security of the shared keys).
This direction has been developed in many papers, [13],
[14], [15], [16] are among the most important. In [14] some
additional interference signals are simultaneously transmitted
from the auxiliary antenna at legitimate access point. The
authors of [13] change slightly the KDP based on Variable
Directional Antenna (VDA) at the cost of a special selection of
appropriate RSSI values in order to improve the key
distribution security. In [15], an experimental scheme with the
execution of dipole antennas was introduced. Such criterion of
KDP security as Information Mutual Anti-tapping Condition
(IMac) has been proposed in [16] but it was not proved that the
IMac correctly estimates the security of KDP.
It is worth to note that in all above mentioned papers, there
has not been considered the use of Privacy Amplification (PA)
of the raw key bit strings and the application of the Privacy
Amplification Theorem in order to correctly estimate the
amount of Shannons information leaking to an eavesdropper,
although it is a very common technique used in the execution
of different KDP [2], [3], [4], [17], [18], [19].
Our contribution consists first of all in an application of PA
to VDA-based KDP that allows restricting reasonably the
values of the required correlations between samples of legal
users and eavesdroppers which are used for key bit generation.
But in order to come closer to this main problem we have to
solve a number of particular problems. The first attempt of
such approach has been presented in [20].
In Section II, we describe the conditions of the physical
channel and we introduce an exact mathematical model of the
KDP. The results of the VDA simulation are presented in
Section III. Section IV contains an optimization of the KDP in
order to provide both reliability and security. Finally we
conclude the main results and present some open problems in
Section V.
II. KDP BASED ON MULTIPATCH WAVE PROPAGATION
AND RANDOMLY EXCITED VDA
The scheme of the communication system corresponding to
the KDP is presented in Fig. 1.

Figure 1. Scheme of communication system corresponding KDP
The KDP is described in the following steps:
1) The legal user A forms the random antenna diagram
by exciting the VDA with output of truly random generator
(TRG) and fix this diagram for some given time interval [0,T
j
]
of the j-th key bit generation, j=1,2,..., n.
2) A transmits to B a harmonic signal S
j
(t) = cos
0
t,
0 t T
j
/2, with the beam pattern obtained at step 1 over the
multipath channel.
3) B receives a harmonic signal by an omni directional
antenna (ODA) and forms the j-th key bit by comparing some
functional
j
computed with the received signal on the time
interval [0, T
j
/2] with a given threshold, forming the j-th key bit
k
j
.
4) The user B switches its ODA in a regime of radiation
and transmits the same harmonic signal S
j
(t)=cos
0
t within the
time interval T
j
/2 t T
j
.
5) The user A switches its VDA to a receiver and
processes the received signal in the same manner as B did,
forming the j-th key bit k
j
.
6) A and B repeat n times the steps 1-5 with new and
independent outputs of TRG in order to create the desired
number of key bits.
Thanks to the Reciprocity Theorem of radio wave
propagation between uplink and downlink, the key sequences
of A and B should be identical up to a random noise of
receivers. Then the signal received by B at the time interval
T
j
/2 t T
j
can be expressed as:

where with respect to the i-th ray at the j-th time interval,
ij
is
the channel attenuation coefficient,
ij
is the VDA amplitude
gain,
ij
is the phase shift, including both phases in antenna
diagram and phase shift in i-th ray, and m is the number of
paths (rays).
The signal received by E at time interval T
j
/2tT
j
is:
where the primed parameters have the same meaning as the

(1)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
174 | P a g e
www.ijacsa.thesai.org

corresponding parameters in (1) but in possession of E. (We
neglect initially the noise at the legal receivers, and we assume
at this moment a noise absence at the eavesdropper E, in
advantage with the illegal users.)
Later we will show that the probability distributions of the
random values
j
and
j
, which are produced by executing some
functionals from both y
j
(t) and z
j
(t) can have a good
approximation by a zero mean Gaussian law.
It is easy to prove by a series of simple but tedious
transforms that the probability of a bit disagreement between
the j-th bit of the legal users and the eavesdropper key bits is:

where is the correlation coefficient between
j
and
j
,

2
=Var(
j
)=Var(
j
).The dependence of p
e
versus is presented
in Fig. 2. We can see that in contrast to our intuition, the
probability p
e
= 0.1 can be provided even when the correlation
coefficient has a significant value 0.95.

Figure 2. The probability of the key bit disagreement between legal and
illegal users depending on the correlation coefficient
In order to enhance the security of the legal user key string
k shared after completion of the KDP it should be performed a
privacy amplification [3], [17], [18], [19], or more specifically
a mapping of the raw key string k to a shorter key string of
length l<n, using the so called hashing procedure =h(k) taken
from the universal class of hash functions [21]. Then the
amount of Shannons information leaking to E given her
knowledge of the string k satisfies

where t = n+nlog
2
(p
2
e
+(1-p
e
)
2
) is the Renyi information under
the assumption that the errors in the eavesdropper's key bits
occur independently due to the independently generated VDA
on each of the j-th time intervals. Hence in order to select the
parameter l we should calculate the correlation coefficient
depending on the mutual location of the legal user and the
eavesdropper, the properties of VDA and the characteristics of
the multipath cannel. A solution for this problem will be
presented in the next Section.
It is worth to remark that the quantized string k' has no
redundancy and it is senseless to perform its soft decoding. As
far as the use of a list decoding with the cipher text encrypted
through the known key k, it looks as an completely intractable
problem due to its large length (see Tables I and II at the end of
Section IV below).
III. CORRELATION BETWEEN THE VALUES AND
Let us consider as VDA the so called ring antenna (RA)
shown in Fig. 3 having N identical isotropic radiators excited
by their random phases. Then the complex instant antenna
diagram can be presented by the well-known formula [22]:

where
s
is a phase in the s-th radiator; k
0
=2/, is the length
of the wave; R the radius of the RA; is the angle in the
azimuthally plane; and is the angle in the vertical plane.
Both instant amplitude and the phase antenna diagrams can
be obtained from (5) and they are random values providing
random exciting to the RA. It would be possible to find
different statistical characteristics of f(,) theoretically but it is
rather more easy to solve the same problem by simulation.
Since the current paper is limited in space, we present only the
main conclusions based on the simulations for the case of
independent and uniformly distributed phases
s
on (0, 2):

Figure 3. Ring antenna with N identical radiators
the probability distribution of the amplitude antenna
diagram has a good approximation through the Rice
law which can be approximated in its turn by a
Gaussian non-zero mean law;
the probability distribution of the phase antenna
diagram has a good approximation by an uniform law
on the interval (0,2).
Next it is possible to compute theoretically the correlation
coefficients between
j
and
j
for different functionals
producing them and to find their probability distributions by
simulation. However, it is necessary to specify the channel
model and thereafter the functional description. To be more
specific, let us consider a 3-ray channel model and a location of
eavesdropper on the line connecting legal users (Fig. 4).
We select two functionals of y
j
(t) and z
j
(t) producing
j
and

j
respectively. Henceforth the functionals are compared with
some thresholds in order to obtain the key bit k
j
. The
functionals are (see eq. (1)):
(2)
(3)
(4)
(5)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
175 | P a g e
www.ijacsa.thesai.org

Figure 4. Channel model with 3-ray wave propagation
envelope: ,
where

phase difference

In a similar manner, there can be presented the
corresponding functionals for eavesdropper:
j
,
cj
,
sj
,
j
.
We will be interested in a procedure to find the probability
distributions of all functionals and correlations between similar
functionals of any legal user B and the eavesdropper E.
Because it is very hard to compute these values theoretically,
we will find them by simulation for some given channel
parameters.
Let us take distance between AB l
1
=25m; distances to the
first and to the second reflecting surfaces, respectively h
1
=3m,
h
2
=3m, N=6, =12.5cm, R=/2 (see Fig's. 3 and 4). Assume
that E is placed between legal users A and B within the interval
l=3-22m. The dependences of the correlation coefficients r
,

and r
,
versus distance l between the eavesdropper E and
the legal user B are shown in Fig. 5(a) and Fig. 5(b)
respectively.
Similar dependences versus distance l
1
between legal users
A and B where l=4m, h
1
=h
2
=3m are presented in Fig.6 (a) and
6(b). In Fig 7 (a) and 7(b) are shown the same dependences but
versus distances to the first reflecting surface and for other
parameters: l
1
=25m, l=4m, h
2
=3m.
As we can see from these figures, these dependences are
looking very strange because if, for the thing, in some points on
the line connecting A and B the correlation of amplitude is
small enough, nevertheless a small shift of the eavesdropper
location with respect to the locations of legal users results in
strong correlation. Similar property holds also for the
correlation of phase differences but the absolute values of this
correlation are at most 0.8 for any conditions. Since the
correlation between the values
j
and
j
occurs less than the
correlation between
j
and
j
(see Fig. 5, 6, 7), it is reasonable
to select the phase difference functional in order to form
j
and
compare it with zero threshold for the k
j
key- bit generation. (In
order to coincide phases of support generators at users A and B,
it is possible to transmit a special pilot signal and to tune
phases of both users at the initial stage of KDP.)
In Fig. 8 there are presented empirical probability
distributions for these functionals. It is evident that both cases
can be approximated by appropriated Gaussian distributions
(see solid curves). Therefore the relation (3) can be used to find
the probability of disagreement between the key bits of the
legal users and the eavesdropper. But before we address to eq.
(4) in order to calculate security of KDP, it should be taken into
account an opportunity for the presence of noise at the
receivers of the legal users.

(a)

(b)
Figure 5. The dependence of correlation coefficients versus distances
between legal use and eavesdropper.
a) for envelope, b) for phase difference

(a)

(b)
Figure 6. The dependence of correlation coefficients versus distances l1
between legal users for l=4m, h1=h2=3m.
a) for envelope, b) for phase difference

(a)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
176 | P a g e
www.ijacsa.thesai.org

(b)
Figure 7. The dependence of correlation coefficients versus distances to the
first reflecting surface and parameters: l1=25m, l=4m, h2=3m. a) for
envelope, b) for phase difference

(a) envelope (b) phase difference
Figure 8. Empirical probability distribution for chosen functionals
IV. KDP OPTIMIZATION UNDER NOISY LEGAL CHANNEL
From now on we remove our previous assumption that the
multipath channel among legal users A and B is noiseless but
keep such condition for eavesdropper's channel. (Obviously,
the last assumption cannot degrade the security of KDP.)
In this setting it is necessary to use some methods in order
to correct disagreements in key bits of legal users. It is very
reasonable to use firstly a selection of the most reliable key bits
with a public discussion over a noiseless channel between legal
users, and then to apply forward error correction codes (FEC)
by sending of the check bits over the same but noiseless
channel. (It is worth to note that a noiseless public channel
among legal users can be arranged by the choice of special
regime, namely large signal power or omni directional antenna
of the user A that we were unable to use for the execution of
KDP.)
The first method of the most reliable key bit selection is to
take the decision according to the rule:

where
j
is the output of
j
, and a threshold.
After a completion of the KDP including a production of
the erased bits for both legal users it is necessary to mutually
announce the numbers of these bits over public noiseless
channels. In this case, the probability of a key bit disagreement
between legal users and eavesdropper, given by (3), has to be
corrected because an eavesdropper is able to intercept
information about the numbers of accepted key bits over the
public channel. We will take into account this fact later for the
simulation procedure. The second method is to keep only the
most reliable key bits, say M, and to remove the others. This
means that the legal users form variation series of the values |
j
|
on a decreasing order and next to keep (after mutual public
discussion) the first M members of this series to generate the
key bits. Of course in this case the probability of key bit
disagreement p
e
is changed also against (3).
Let us denote by p1 and p2 the probability of legal key bit
errors after the first and the second method, respectively. Next
we use an error-correcting code (n
0
+r, n
0
) sending a sequence
of r check symbols over public noiseless channel in order to
correct eventually errors in the key sequence.
Then the probability of erroneous decoding P
ed
by the
modified Gallager's theorem is [19]: P
ed
2
-noE(Rc)
, where


R
C
= n
0
/(n
0
+r), and no is the number of bits k
j
which have
been kept by legal users after erasing the unreliable bits
following the first or the second procedures, and p is the error
probability for the kept bits. In the case of check symbol
sending, the Privacy Amplification Theorem against (4)
becomes [19]:

KDP optimization problem is to get the maximum key rate

where n
er
is the number of erased symbols after the use of the
method 1 or 2 and given the values I(;k), P
ed
, l, and different
signal-to-noise ratio (S/N) at the receivers of the legal users.
We solve this problem by simulation for the case of Gaussian
noise at the legal receivers.
In Tables I and II there are presented the results of such
optimization for typical conditions for the first and the second
method of unreliable bits removal, respectively, where P
er
is
the probability of key bit erasing.
We can see from these tables that the second method is for
large correlation a little bit better than the first one. However
both methods provide sufficiently reliable and secure key
sharing if eavesdropper is placed on 3-21m away from legal
user B and phase difference is used as key generating method
(see Fig. 5(b)).
A similar conclusion is drawn also for multipath channels
with a greater number of rays and with other reasonable
parameters and eavesdropper locations. In order to enhance the
security of the KDP, antenna diversity can be used when B has
m omni directional antennas and he selects randomly one of
them at each time period T
j
to receive and transmit signal. Then
the relation finding the Renyi information used in (4) changes
for:

(5)
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
177 | P a g e
www.ijacsa.thesai.org
TABLE 1. KEY RATE MAXIMIZATION FOR THE FIRST METHOD GIVEN
I(;K)=10
-9
, PED=10
-5
, S/N=10 AND DEFFERENT
opt pe Per p1 l n0 Rk
0.8 0.18 0.263
0.19
4
0.0087 128 539 0.243
0.19
1
0.0083 256 940 0.272
0.18
9
0.0082 512 1685 0.303
0.95 0.14 0.152
0.18
9
0.0083 128 1528 0.084
0.18
8
0.0082 256 2484 0.103
0.18
7
0.0082 512 4195 0.122
0.99 0.11 0.051
0.18
3
0.0078 128 7405 0.017
0.18
1
0.0075 256 10977 0.023
0.18 0.0075 512 15234 0.033
TABLE 2. KEY RATE MAXIMIZATION FOR THE SECOND METHOD GIVEN
I(;K)=10
-9
, PED=10
-5
, S/N=10 AND DEFFERENT
Mopt pe Per p2 l Rk
0.8
539
0.222
0.24
0.0075
128 0.245
940 0.238 256 0.277
1685 0.236 512 0.306
0.95
1528
0.115
0.236
0.0072
128 0.095
2484 0.235 256 0.105
4195 0.233 512 0.123
0.99
7405
0.049
0.23
0.0069
128 0.017
10977 0.29 256 0.023
15234 0.29 512 0.033

The relation (6) holds with the probability equal to the
probability of the event in which with at least of one of
antennas mutual location of the legal user and the eavesdropper
is got such that *, where * is found by (3) given
e
p%
.
We considered so far a scenario when an eavesdropper uses
the same omni directional antenna as the legal user B. But E
can execute directional antenna to separate all rays and to
process the best of them or even apply joint processing to all of
them. We have performed a simulation of the case with single
ray separation and it has been shown that the correlation
coefficient even decreases in comparison with one presented
before. The case of joint processing of separated rays is
noteworthy. But we can remark that even under the very strong
condition in which the eavesdropper knows exactly all channel
parameters both for E and B, there is still uncertainty about
VDA gains in the direction of E and B. Therefore, generally
speaking, the correlation coefficient occurs even in this case
with a value less than one.
V. CONCLUSION AND FUTURE WORK
We considered a method of key sharing based on the
concept of a VDA under the condition of multipath channel
and we showed that sufficient security and reliability of the
shared keys can be provided even when the eavesdropper's
channel is noiseless. (It is worth to remark that in order to get
such result, the following two conditions are necessary: to
create truly randomness with the help of a VDA and to have
multipath channels.) The results of investigations show that the
security of the KDP (in terms of Shannon's information leaking
to eavesdropper) does not depend only on the distance between
legal users and eavesdropper but also on the eavesdropper's
location. This result somewhat contradicts to a very optimistic
conclusion in [12].
We propose to use the difference-phase functional instead
of either quadrature components or envelope in order to form
key bits. This approach results in less mutual correlation
between legal user and eavesdropper and simplifies a choice of
threshold. The key sequence k is i.i.d if VDA is excited by
independent random phases and threshold is chosen in an
appropriate manner. (This fact has been confirmed by
simulation using statistical tests.) Our contribution consists also
in the proof of relation (3) which allows to connect the
probability of disagreement between the key bits of legal users
and eavesdropper with the correlation of corresponding values.
Unfortunately, a limited space of the paper does not allow us to
show all simulation results for different multipath channels and
mutual location of legal users and eavesdroppers, which we
have got at our disposition. It is pertinent to note that although
some results were obtained by the use of computer simulation,
it does not lead to a loss of generality because this is only an
approach to reach the same goal by a simpler way. As far as the
limitation due to the parameter selection (number of rays,
position of eavesdropper, S/N, etc) it can be explained only by
the space paper limitations. Indeed, following our theory, one
can get the results corresponding to arbitrary parameters.
In the future we are going to investigate: i) the use of
multitone signals in the KDP, ii) the optimal processing of the
eavesdropper rays separation in order to provide the greatest
correlation, iii) the use of real FEC and effective decoding
algorithms with KDP (instead of extended Gallager's
bounds);and, iv) the use of other types of VDA (like ESPAR or
others).
REFERENCES
[1] C. H. Bennett and G. Brassard, "Quantum cryptography: Public key
distribution and coin tossing," in Proceedings of International
Conference on Computers, Systems and Signal Processing, December
1984.
[2] U. Maurer, "Protocols for secret key agreement by public discussion
based on common information." in CRYPTO, ser. Lecture Notes in
Computer Science, E. F. Brickell, Ed., vol. 740. Springer, 1992, pp.461-
470.
[3] U. Maurer and S. Wolf, "Secret-key agreement over unauthenticated
public channels II: the simulatability condition." IEEE Transactions
onInformation Theory, vol. 49, no. 4, pp. 832-838, 2003.
[4] V. Yakovlev, V. Korzhik, and G. Morales-Luna, "Key distribution
protocols based on noisy channels in presence of an active adversary:
Conventional and new versions with parameter optimization," IEEE
Transactions on Information Theory, vol. 54, no. 6, pp. 2535-2549,
2008.
[5] A. M. Sayeed and A. Perrig, "Secure wireless communications: Secret
keys through multipath," in ICASSP. IEEE, 2008, pp. 3013-3016.
[6] Y. Liu, S. C. Draper, and A. M. Sayeed, "Secret key generation through
ofdm multipath channel," in CISS. IEEE, 2011, pp. 1-6.
[7] M. G. Madiseh, M. L. McGuire, S. S. Neville, L. Cai, and M. Horie,
"Secret key generation and agreement in uwbcommuniction channels,"
in GLOBECOM 2008.IEEE, 2008.
[8] M. G. Madiseh, S. W. Neville, and M. L. McGuire, "Time correlation
analysis of secret key generation via uwb channels," in GLOBECOM.
IEEE, 2010, pp. 1-6.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
178 | P a g e
www.ijacsa.thesai.org
[9] A. A. Hassan, W. E. Stark, J. E. Hershey, and S. Chennakeshu,
"Cryptographic key agreement for mobile radio," Digital Signal
Processing, vol. 6, pp. 207-212, 1996.
[10] J. Hershey, A. Hassan, and R. Yarlagadda, "Unconventional
cryptographic keying variable management," Communications, IEEE
Transactions on, vol. 43, no. 1, pp. 3 -6, Jan. 1995.
[11] Ch.Ye, S. Mathur, AReznik, W. Trappe and N.B. Mandayam.
"Information-Theoretically Secrete Key Generation for Fading Wireless
Channels, IEEE Transactions on Information Forensics and Security ,
vol.5, No.2,June 2010,pp.240-254.
[12] T. Aono, K. Higuchi, T. Ohira, B. Komiyam, and H. Sasaoka, "Wireless
secret key generation exploiting reactance-domain scalar response of
multipath fading channels," IEEE Trans. Antennas &ropagation,vol. 53,
pp. 3776-3784, Nov. 2005.
[13] T. Shimizu, H. Iwai, and H. Sasaoka, "Improvement of key agreement
scheme using espar antenna," in Proc. 2008 Int. Symp. Antennas
Propag.(ISAP2008). Taipei, Taiwan, 2008, pp. 1-4.
[14] M. Onishi, T. Kitano, H. Iwai, and H. Sasaoka, "Improvement of
tolerance for eavesdropping in wireless key agreement scheme using
espar antenna based on interference transmission," in The 2009
International Symposium on Antennas and Propagation (ISAP
2009).Bangkok, Thailand: ISAP, October 20-23 2009.
[15] T. Shimizu, N. Otani, T. Kitano, H. Iwai, and H. Sasaoka, "Experimental
validation of wireless secret key agreement using array antennas," in
XXX URSI General Assembly and Scientific Symposium. Istambul,
Turkey: URSI, August 11-20 2011.
[16] T. Yoshida, T. Saito, K. Fujiki, K. Uematsu, and H. U. T. Ohira, "Impact
of direct-path wave on imac in secret key agreement system using espar
antennas," in XXX URSI General Assembly and Scientific Symposium.
Istambul, Turkey: URSI, August 11-20 2011.
[17] C. H. Bennett, G. Brassard, C. Crepeau, and U. M. Maurer, "Generalized
privacy amplification," IEEE Transactions on Information Theory, vol.
41, no. 6, pp. 1915-1923, 1995.
[18] U. Maurer and S. Wolf, "Privacy amplification secure against active
adversaries," Lecture Notes in Computer Science, vol. 1294, pp. 307-
321,1997.
[19] V. Korjik, G. Morales-Luna, and V. Balakirsky, "Privacy amplification
theorem for noisy main channel," Lecture Notes in Computer Science,
vol. 2200, pp. 18-26, 2001.
[20] V.Korzhik V. Yakovlev,,G. Morales-Luna, Yu. Kovajkin. "Wireless
Secrete Key Sharing Based on on the Use of Variable-directional
Antenna over Multipath Channels", in Proc.ELMAR' 2010 ,pp.277-280.
[21] L. Carter and M. N. Wegman, "Universal classes of hash functions," J.
Comput. Syst. Sci., vol. 18, no. 2, pp. 143-154, 1979.
[22] R. E. Collin and F. J. Zucker, Antenna Theory - Part 1.McGraw-
Hill,1969.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
179 | P a g e
www.ijacsa.thesai.org
The Relationships of Soft Systems Methodology
(SSM), Business Process Modeling and e-
Government

Dana Indra Sensuse
Faculty of Computer Science
University of Indonesia
Depok, Indonesia

Arief Ramadhan
Faculty of Computer Science
University of Indonesia
Depok, Indonesia

Abstracte-Government have emerged in several countries.
Because of many aspects that must be considered, and because of
there are exist some soft components in e-Government, then the
Soft Systems Methodology (SSM) can be considered to use in e-
Government systems development process. On the other hand,
business process modeling is essential in many fields nowadays,
as well as in e-Government. Some researchers have used SSM in
e-Government. Several studies that relate the business processes
modeling with e-Government have been conducted. This paper
tries to reveal the relationship between SSM and business process
modeling. Moreover, this paper also tries to explain how business
process modeling is integrated within SSM, and further link that
integration to the e-Government.
Keywords- Soft Systems Methodology; Business Process Modeling;
e-Government.
I. INTRODUCTION
e-Government is the use of Information Technology (IT) by
public sector organizations [1]. The main orientation of e-
Government is the accessibility of information by the public,
rather than financial income, as implied in [1].
Because the target of e-Government is the public sector,
then the e-Government systems are generally built based on the
web technology. This technology is used because it has ability
to reach people quickly and widely. Therefore, the e-
Government can be associated as an attempt to put the public
administration online. This means, e-Government is not simply
replace all the equipment in the public sector, from a typewriter
to a computer. e-Government is more than that. e-Government
move the whole system along with existing business processes
within the public sector to the online world.
Because of its relation with IT, then most people thought
that e-Government is part of computer science. However, e-
Government has become an emergent multidisciplinary field of
research [2]. In addition to computer science, there are many
other scientific fields in e-Government, for example public
administration, management, politics, socio culture etc.
It is revealed in [2], that although theoretical ground is still
under construction, e-Government certainly qualifies as a
legitimate emerging scientific discipline. It also revealed in [2],
that as technological innovations are continuously hitting the
market, the frontiers of the e-government discipline are moving
and its multidisciplinary nature confirmed [2].
Currently the development of e-Government systems have
proliferated in several countries, both in developing countries
and developed countries. It is stated in [3], that e-government is
a useful tool for modernizing the state given that it enables
government to offer higher-quality services to citizens and
provide those services in a more efficient, effective, and
transparent way.
Heeks in [1] says that e-Government is also an information
system, but it is enriched with various aspects, such as the
management aspects, political aspects, economical aspects and
others. These aspects have to be considered by developers
when developing an e-Government system.
Because of many aspects that must be considered, then the
e-Government system development process can be very
complex. These aspects cannot be observed separately, but
should be observed as a whole, where there is interaction in it.
Such characteristics can be solved using systems thinking.
e-Government is a socio-technical system that consists of
soft components and hard components [1]. It could be argued
that the soft component is the people who are involved in e-
Government, whereas the hard component is the Information
Technology (IT) that being used. The management approach of
the soft component is likely inspired by social sciences, it tends
to be subjective, qualitative, and further highlight by the
aspects of humanism [1]. The management approach of the
hard component is inspired by engineering science, it tends to
be objective, quantitative and further highlight by the technical
aspects [1].
It is implied in [1], that the most critical factor to determine
whether an e-Government system development fail or not is the
soft component. Because of the soft component is very
dominating and is tend to be subjective, then we see that one
type of systems thinking, that is soft systems thinking, can be
used in e-Government system development.
Soft systems thinking do not assume that the world is
systemic and well-ordered; on the contrary, it assumes social
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
180 | P a g e
www.ijacsa.thesai.org
reality to be problematical, characterized by multiple angles
of approaches and perspectives [4]. The understanding of
reality is dependent upon the observer, his interpretations and
what he chooses to focus on [4].
Some of the methodologies that can be used in soft systems
thinking is a meta-synthesis approach used in [5], and Soft
Systems Methodology (SSM). In this paper, we choose SSM
and try to find the relationships between SSM and others.
II. SOFT SYSTEMS METHODOLOGY
Soft Systems Methodology (SSM) was proposed in 1981 by
Peter Checkland [6]. As the name implies, SSM is based on
soft systems thinking. The picture of SSM can be seen in Fig.
1. SSM consists of seven steps, i.e. (extracted from [6]):
1) The identification of a problem situation that demands
attention

Figure 1. The seven-step of Soft System Methodology (SSM) [6].
2) Problem situation is expressed. The expression can be
described using the Rich Picture Diagram. The examples of the
Rich Picture Diagram can be seen in Fig. 2.
3) Some relevant human activity systems, potentially
offering insight into the problem situation, are selected and
from these root definitions are built. In this step, CATWOE
analysis is performed. CATWOE stands for Customers, Actors,
Transformation process, World view, Owner, and
Environmental constraints.
4) Construct conceptual models. This is the most important
step in the SSM. Various modes of modeling techniques can be
applied at this step.
5) Comparing the conceptual model with the real world.
The aim is to provide material for debate about possible change
among those interested in the problem situation. This step
shows the social processes within the SSM.
6) Making changes to the model by accommodating the
interests of several actors involved. Changes should be able to
follow the desired model but still possible (feasible)
historically, culturally and politically. Changes may include
changes in attitudes, structures, or procedures.
7) Perform various activities to implement the model and
fix the problem. In this step, the conclusions are drawn and
long-term solution is formulated.
SSM has been amended several times. The first change is
made in 1990 in the form of "two-strands model" [6]. In this
model, were added three types of inquiry, referred to as
Analysis 1, 2 and 3 [6]. Analysis 1 considers the intervention
itself and the roles of client, problem-solver and problem-
owners. Analysis 2 is social system analysis. Analysis 3
examines the politics of the problem situation and how power
is obtained and used [6].

Figure 2. Rich Picture diagram for the system of procurement in the
Department of Agriculture [1].
Subsequent change of SSM is made when the original
seven-step is merged into just four steps. The new four-step is
named as learning cycle of SSM. Four new steps are [7]:
1) Finding out about a problem situation, including
culturally/politically.
2) Formulating some relevant purposeful activity models
3) Debating the situation, using the models, seeking from
that debate both :
a) changes which would improve the situation and are
regarded as both desirable and (culturally) feasible
b) the accommodations between conflicting interests
which will enable action-to-improve to be taken
4) Taking action in the situation to bring about
improvement.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
181 | P a g e
www.ijacsa.thesai.org
Although the SSM has been amended several times and
although Checkland no longer favours it, the representation of
SSM as a seven-step, which appeared in 1981, is still
frequently used today [6]. Some researchers have used SSM in
e-Government, for example can be seen in [24].
III. BUSINESS PROCESS MODELING AND E-GOVERNMENT
Business process is characterized by three key words, i.e.
activities, linked, and objective. It can be seen from some of the
definitions of business process. In [8], it is said that business
process is a set of one or more linked procedures or activities
which collectively realize a business objective or policy goal,
normally within the context of an organizational structure
defining functional roles and relationships. Other definition
about business process is a set of coordinated tasks and
activities, involving both human and system interactions, that
will lead to accomplishing a set of specific organizational goals
[9]. In [10], it is stated that a business process is a set of related
activities or operations which, together, create value and assist
organizations to achieve their strategic objectives.
A business process has a clear beginning and end, creating
outputs by adding value to inputs [10]. It seems that business
process is more likely a function. However, in [11], it was
stated that a business process is not the same as a function [11].
It was said in [11], that the people and operations that include
in a single business process may come from more than one
traditional functional group.
The first stage of the analysis of a business process is
concerned with constructing a model of the business process
[11]. This "constructing" activity is commonly called business
process modeling.
Business process modeling became popular in the context
of enterprise reorganization and modernization in the early
1990s [12]. Business process modeling is the visual
representation of business processes [13]. Visual representation
is usually done in the form of pictures or notations with
specific meanings.
Guizzardi et. al. in [14] stated that business process
modeling is about the description of sequence of business
activities carried out in organizations in order to make them
explicit. It is implied in [15], that business process modeling is
done for better understanding and analysis about the business
process.
It is indicated in [16], that business process modeling can
be used to communicate a wide variety of information to a
wide variety of audiences. Some audiences who become the
main focus are the stakeholders. The graphical nature of
business process models can be used as a medium of
communication between stakeholders (e.g., executives,
developers, and employees) [15]. Business process modeling
capabilities as a medium of communication is also revealed in
[17].
Currently, there exists some software that can be used to
manage the business processes, one of which is SAP [29].
Bider in [30] stated that there are four common views on the
process development, i.e.: input/output flow, workflow, agent-
related workflow, and state workflow.
Business process modeling is essential in many fields
nowadays, and much research and many initiatives have been
proposed in order to facilitate and improve its development
[18]. One example of business process modeling activities
regarding shipment process of a hardware retailer can be seen
in Fig. 3.

Figure 3. Shipment Process of a hardware retailer [16].
From its name, i.e. "Business Process", and from some of
the above explanations, it is impressed that the "business
process" is only needed and used within the business sector
(private sector). However, there are some studies that link the
business process modeling with e-Government (public sector).
Several studies that relate the business processes modeling with
e-Government can be seen in [25], [26] and [27].
IV. SOFT SYSTEMS METHODOLOGY, BUSINESS PROCESS
MODELING AND E-GOVERNMENT
In the section I and II it has been explained that there is a
correlation between the SSM with e-Government, that is based
on its characteristics, SSM can be used in the development of
e-Government systems and that there has been many studies
linking SSM with e-Government. Then, in section III it has
been revealed that many studies has linked business process
modeling with e-Government.
Within this section, we try to describe and propose how
SSM can also be associated with business process modeling. In
addition, at the same time, we also will reconcile these linkages
with e-Government.
In [19], it is shown that the SSM can be associated with
business process modeling. In [19], it is implied that the result
of business process modeling is depicted using Unified
Modeling Language (UML). However, in [19], the process of
making business process model with the UML, is depicted as
standing outside the SSM. In [19], that being compared with
the rich picture diagram is still the conceptual model and not
the business process model. It was further shown, that in [19],
the process of making business process models with the UML,
is conducted separately from the SSM.
In this paper, we propose that the process of making
business process model can be fused within SSM. This
integration process occurs by placing the process of making
business process model into the step 4 of the SSM. So, the
conceptual model that originally resulting from that step can be
in the form of a business process model, which can be depicted
using UML, Business Process Modeling Notation (BPMN) or
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
182 | P a g e
www.ijacsa.thesai.org
the other. Because of this integration, the result of step 1 (e.g.
in the form of surveys results or interviews results), step 2 (e.g.
in the form of rich picture diagram) and step 3 (e.g. CATWOE
analysis) of the SSM, can be considered by business process
modeler to create the business process model in step 4.
Henceforth, in step 5, the business process model will be
compared with the rich picture diagram of the real-world
conditions. It can be seen that in our proposal the making of
business process model is integrated within SSM, and not to
put it separately. The reason for this integration proposal will
be described in the explanation below.
In the conduct of business process modeling there are
several obstacles that may be faced. In [28], it was said that
defining a business process is a taxing, vexing, and iterative
process. Business process modeling is often time-consuming
and sometimes involves a certain amount of redundant work
because of the similarity between modeling objects [20].
Business process modeling is a complicated process [21].
In addition to these constraints, the business process
modeling is also ultimately depends on the modeler who is also
a human being. As stated in [22] that the business process
model is the result of mapping a business process, and that
business process can be either a real-world business process as
perceived by a modeler, or a business process conceptualized
by a modeler. In this case it appears there are two key words
which tend to be subjective and depends on the modelers
perspective, i.e. to perceive and to conceptualize. Because
of the subjective tendencies that are soft and the fact that
business process modeling is very complicated, the SSM is
really suitable to this case.
In [23], it is said that building a model of a real business
process is a challenging task because:
Business processes are not always clearly visible as
they may go through the whole, often functionally
structured organization.
Written information about business processes is often
non-existing or not reliable. The only practical way to
obtain reliable information for creating a model of a
real business process is by interviewing the people
engaged in the process.
From the statement in [23] above, again it is shown that
there is a very common thing carried out in SSM, i.e. the
interview. In the SSM, the interview, usually carried out in the
step 1. This again shows the relationship between SSM with
business process modeling.
In addition, business process modeling involves an
abstraction from the real-world business process [22]. It is
highly compatible with SSM paradigm. Based on Fig. 1, it is
shown that the SSM actually also tried to make abstraction of
the real-world problems. SSM bring the real-world problems
(upper half of the figure) into the conceptual model (bottom
half of the figure). The fact that there are similarities between
SSM paradigms with business process modeling further
strengthen our proposal that the business process modeling can
be integrated in the SSM.
Based on the above explanation and based on our
explanations in section I, II and III, further, we can see that the
SSM, business process modeling and e-Government can be
mixed together. In other words, the making of business process
model in the development of e-Government system can be
done with the SSM.
V. CONCLUSIONS
This paper has explained how the business process
modeling can be integrated into the SSM. Furthermore, this
paper also has described the relationship between the SSM,
business process modeling and e-Government. It can be
concluded that the SSM can be used in the making of business
process models in the development of e-Government system.
Some further research can be conducted to enrich this paper,
for example the research about how to do the real
implementation of the integration of SSM and business process
modeling in a country.
REFERENCES
[1] R. Heeks, Implementing and Managing eGovernment An International
Text, London, England : SAGE Publications, 2006.
[2] S. Assar, I. Boughzala, and I. Boydens, Back to Practice, a Decade of
Research in E-Government, in Practical Studies in E-Government :
Best Practices from Around the World, S. Assar, I. Boughzala, and I.
Boydens, Eds. New York, USA: Springer, 2011.
[3] L. Herrera and J. R. Gil-Garcia, Implementation of E-Government in
Mexico: The Case of Infonavit, in Practical Studies in E-Government :
Best Practices from Around the World, S. Assar, I. Boughzala, and I.
Boydens, Eds. New York, USA: Springer, 2011.
[4] A. Mirijamdotter, A Multi-Modal Systems Extension to Soft System
Methodology, Ph. D. dissertation, Lulea Tekniska Universitet, Sweden,
1998.
[5] J. Gu and X. Tang, Meta-Synthesis System Approach To Knowledge
Science, International Journal of Information Technology & Decision
Making, vol. 6, no. 3, pp. 559-572, 2007.
[6] M. C. Jackson, System Thinking Creative Holism for Managers, John
Wiley & Sons Ltd, England, 2003.
[7] P. Checkland, Soft Systems Methodology: A Thirty Year
Retrospective, Systems Research and Behavioral Science, vol. 17, pp.
S11-S58, 2000.
[8] [Workflow Management Coalition], The Workflow Management
Coalition Specification : Terminology & Glossary, Document Number
WFMC-TC-1011, Document Status - Issue 3.0, 1999.
[9] [Oracle], Oracle Application Integration Architecture: Business Process
Modeling and Analysis, Oracle White Paper, April 2009.
[10] [Australian Government Information Management Office], The
Australian Government Business Process Interoperability Framework,
July 2007.
[11] S. J. Childe, P. A. Smart, and A. M. Weaver, The use of generic
process models for process transformation, in Proceedings of the IFIP
TC5 WG5.7 international workshop on Modelling techniques for
business process re-engineering and benchmarking, pp. 51-60, 1997.
[12] R. von Ammon, C. Emmersberger, F. Springer, and C. Wolff, Event-
Driven Business Process Management and its Practical Application
Taking the Example of DHL, in Proceedings of the 1st iCEP08
Workshop on Complex Event Processing for the Future Internet, 2008.
[13] M. Ramadan, H. G. Elmongui, and R. Hassan, BPMN Formalisation
using Coloured Petri Nets, in Proceedings of the 2nd GSTF Annual
International Conference on Software Engineering & Applications (SEA
2011), 1997.
[14] R. S. S. Guizzardi, G. Guizzardi, J. P. A. Almeida, and E. Cardoso,
Ontological Foundations for Agent-Oriented Organizational
Modeling, in Proceedings of the 3rd International i* Workshop -
istar08, pp. 37 41, 2008.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
183 | P a g e
www.ijacsa.thesai.org
[15] A. Lodhi, V. Koppen, and G. Saake, Business Process Modeling:
Active Research Areas and Challenges, Department of Technical and
Business Information Systems, Faculty of Computer Science, Otto-von-
Guericke University, Tech. Rep., Nr : FIN-001-2011, 2001.
[16] [Object Management Group], Business Process Model and Notation
(BPMN) Version 2.0, OMG Document Number : formal/2011-01-03,
2011.
[17] A. Popovic, M. I. Stemberger, and J. Jaklic, Applicability of Process
Maps for Simulation Modeling in Business Process Change Projects,
Interdisciplinary Journal of Information, Knowledge, and Management,
vol. 1, pp. 109-123, 2006.
[18] A. Koschmider, J. L. de la Vara, and J. Sanchez, Measuring the
Progress of Reference Model-Based Business Process Modeling, in
Proceedings of 3rd International Conference on Business Process and
Services Computing (BPSC 2010), pp. 218 229, 2010.
[19] M. Salahat and S. Wade, Measuring the Progress of Reference Model-
Based Business Process Modeling, in Proceedings of The 5th
International Conference on Innovations in Information Technology,
2008.
[20] C. Ren, W. Wang, J. Dong, H. Ding, B. Shao, and Q. Wang, Towards
A Flexible Business Process Modeling And Simulation Environment,
in Proceedings of the 2008 Winter Simulation Conference, IEEE
Xplorer, pp. 1694-1701, 2008.
[21] R. Lu and S. Sadiq, A Survey of Comparative Business Process
Modeling Approaches, in Lecture Notes in Computer Science,
Springer, Vol. 4439/2007, 82-94, 2007.
[22] J. Mendling, Foundations of Business Process Modeling, in Handbook
of Research on Modern Systems Analysis and Design Technologies and
Applications, M. R. Syed and S. N. Syed, Eds. Hershey, USA: IGI
Global, 2009.
[23] I. Bider, State-Oriented Business Process Modeling: Principles, Theory
and Practice, Ph. D. dissertation, Royal Institute of Technology and
Stockholm University, Sweden, 2002.
[24] M. Alrazooqi and R. De Silva, Mobile and Wireless Services and
Technologies for M-Government Solution Proposal for Dubai
Government, WSEAS Transactions on Information Science And
Applications, Issue 8, Vol. 7, pp. 1037-1047, 2010.
[25] J. Becker, D. Pfeiffer, M. Rackers, and P. Fuchs, Business Process
Management in Public Administrations The PICTRUE Approach, in
Proceedings of 11th Pacific-Asia Conference on Information Systems,
2007.
[26] O. F. Aydnl, S. Brinkkemper, and P. Ravesteyn, Business Process
Improvement in Organizational Design of E-Government Services,
Department of Information and Computing Sciences Utrecht University,
Tech. Rep. UU-CS-2007-041, 2007.
[27] P. Liegl et. al., Modeling eGovernment processes with UMM,
Informatica, Vol. 31, pp. 407417, 2007.
[28] F. Nickols, The Difficult Process of Identifying Processes : Why it isnt
as easy as it sounds, Knowledge and Process Management, Vol. 5, No.
1, 1998.
[29] [www.winshuttle.com]. 2011. How SAP Users hold the key to Business
Process Improvement. http://www.winshuttle.com/White-
Papers/Winshuttle-HowSAPUsersHoldtheKeytoBPI-whitepaper-EN.pdf
[30] I. Bider 2002. Tutorial on: Business Process Modeling as a Method of
Requirements Engineering. http://www.iceis.org/iceis2007/
Hall_Of_Fame/ ibider/ ibider_2002.pdf
AUTHORS PROFILE
Dana Indra Sensuse. B.Sc in Soil Science (Bogor Agricultural University,
Indonesia, 1985), M.Sc in Library and Information Studies (Dalhousie
University, Halifax, Canada, 1994), Ph.D in Information Studies (Toronto
University, Canada, 2004), Lecturer at University of Indonesia, Head of e-
Government Lab at University of Indonesia.
Arief Ramadhan. B.Sc in Computer Science (Bogor Agricultural University,
Indonesia, 2005), M.Sc in Computer Science (Bogor Agricultural University,
Indonesia, 2010), Ph.D Student in Computer Science (University of
Indonesia), Research Assistant at University of Indonesia. Member of e-
Government Lab at University of Indonesia.


(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
184 | P a g e
www.ijacsa.thesai.org
Re-tooling Code Structure Based Analysis with
Model-Driven Program Slicing for Software
Maintenance

Oladipo Onaolapo Francisca (PhD)
Computer Science Department,
Nnamdi Azikiwe University
Awka, Nigeria

AbstractStatic code analysis is a methodology of detecting
errors in program code based on the programmer's reviewing the
code in areas within the program text where errors are likely to
be found and since the process considers all syntactic program
paths; there is the need for a model-based approach with slicing.
This paper presented a model of high-level abstraction of code
structure analysis for a large component based software system.
The work leveraged on the most important advantage of static
code structure analysis in re-tooling software maintenance for a
developing economy. A program slicing technique was defined
and deployed to partition the source text to manageable
fragments to aid in the analysis and statecharts were deployed as
visual formalism for viewing the dynamic slices. The resulting
model was a high-tech static analysis process aimed at
determining and confirming the expected behaviour of a software
system using slices of the source text presented in the statecharts.
Keywords- software maintenance; static analysis; syntactic program
behavior; program slicing.
I. INTRODUCTION
There had been a growing use of static analysis both in
commercial and academic areas in the verification of properties
of software used and locating potentially vulnerable code
in critical sensitive computer systems [1]. Static code
analysis is the analysis of computer software that is performed
without actually executing programs built from that software.
This is in contrast to the analysis performed on executing
programs which is known as dynamic analysis [2]. Empirical
evidence from [3] and [4] showed that software maintenance
consumed between 50% and 80% of the resources in the total
software budget. In addition, according to [5] and [6]; an
estimated 50% to 80% of the time and material involved in
software development is devoted to maintenance of existing
code, hence the main justification for this work was to leverage
on the most important advantage of static code structure
analysis in re-tooling software maintenance for a developing
economy.
The most important advantage of static code analysis lied in
the possibility of considerable cost saving by defects
elimination in a program; this is expected to bring about some
economic benefits to a developing country as technological
innovations were known to do and they can be expected to save
considerable costs in software maintenance. The earlier an
error is determined; the lower the cost of its correction.
According to [7], correction of an error at the testing stage is
ten times more expensive than its correction at the construction
(coding) stage.
This paper presented a model for code structure analysis for
a large component based software system. A program slicing
technique was deployed to partition the code to manageable
fragments to aid in the analysis and statecharts were deployed
as visual formalism to view the dynamism of the static slices.
The model aimed at determining and confirming the expected
behavior of a software system using the source text because
research had shown that during maintenance, the most reliable
and accurate description of the actual behavior of a software
system is its source code [8]. The rest of this paper is organized
as follows: A background to the concepts of static codes
structure analysis and program slicing was presented in section
II; section 3 described the materials and methods adopted in the
research, the resultant models were discussed in section 4 and
the section 5 concluded the paper.
II. RESEARCH BACKGROUND
Identifying errors in software during development is very
important so that the end product can be error free and perform
to its specification. Reference [9] believed that early
identification of bugs in a developing program can be achieved
through the concept of program slicing. Generally, program
slice has a wider spectrum of applications that include
debugging, testing, maintenance, code understanding,
complexity measurement, security etc.
Static analysis is any form of analysis that does not require
a system to be operated. The process complements dynamic
analysis, where system operation is central. When applied to
code, static analysis is typically referred to as white-box, glass-
box, structural or implementation based techniques [10]. In
most cases the analysis is performed on some version of
the source code and in the other cases some form of the object
code. The term is sometimes applied to the analysis performed
by an automated tool, with human analysis being called
program understanding, program comprehension or code
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
185 | P a g e
www.ijacsa.thesai.org
review. Static Analysis is performed without program
execution and the process includes almost everything except
conventional testing. This is because dynamic testing requires
running code, some program properties such as race conditions
are too hard to test for and though it may be impossible to
program correctness, one can easily prove simple properties of
simplified models. Static analysis can be applied earlier in
development because some kinds of defects are hard to find by
testing (e.g., timing-dependent errors) and because testing and
analysis are complementary; each is best at finding different
faults. The sophistication of the analysis performed by tools
varies from those that only consider the behavior of individual
statements and declarations, to those that include the complete
source code of a program in their analysis. Uses of the
information obtained from the analysis vary from highlighting
possible coding errors to formal methods that mathematically
prove properties about a given program [11].
One implementation technique of formal static analysis is
model checking; it includes the consideration of systems that
have finite state or may be reduced to finite state
by abstraction. Data-flow analysis is a lattice-based static
analysis technique for gathering information about the possible
set of values and the abstract interpretation models the effect
that every statement has on the state of an abstract machine. In
this case, the model 'executes' the software based on the
mathematical properties of each statement and declaration. This
abstract machine over-approximates the behaviours of the
system and the abstract system is thus made simpler to analyze,
at the expense of incompleteness since not every property true
of the original system is true of the abstract system. If properly
done, though, abstract interpretation is sound as every property
true of the abstract system can be mapped to a true property of
the original system [12].
The sophistication of the analysis performed by the model
varies from those that only consider the behavior of individual
statements and declarations, to those that include the complete
source code of a program in their analysis. Uses of the
information obtained from the analysis vary from highlighting
possible coding errors to formal methods that mathematically
prove properties about a given program [13]. Static analysis
can find weaknesses in the code at the exact location, it can be
conducted by trained software assurance developers who fully
understand the code and it allows a quicker turn around for
fixes. In addition to all these, it permits weaknesses to be found
earlier in the development life cycle, reducing the cost to fix.
Program slicing was initially proposed to guide
programmers during program debugging, but had been found to
be useful for the process of understanding programs. Dynamic
slicing was used to identify those parts of the program that
contributed to the computation of the selected function for a
given program execution. This can be used to understand
program execution by adopting a commonly used high level
abstraction. Program slicing is the computation of the set of
programs statements, the program slice that may affect the
values at some point of interest, referred to as a slicing
criterion. Program slicing can be used in debugging to locate
source of errors more easily. Other applications of slicing
include software maintenance, optimization, program analysis,
and information flow control. [13]. Slicing techniques have
been seeing a rapid development since the original definition
by Mark Weiser. At first, slicing was only static, i.e., applied
on the source code with no other information than the source
code. Reference [14] introduced dynamic slicing, which
worked on a specific execution of the program; for a given
execution trace. Based on the original definition of [15],
informally, a static program slice S consists of all statements in
program P that may affect the value of variable v at some point
p. The slice is defined for a slicing criterion C=(x,V), where x
is a statement in program P and V is a subset of variables in P.
A static slice includes all the statements that affect variable v
for a set of all possible inputs at the point of interest (i.e., at the
statement x). Static slices are computed by finding consecutive
sets of indirectly relevant statements, according to data and
control dependencies. Dynamic slicing techniques provided a
means to prune unrelated computation, and it may help to
narrow down this part of a program that contributed to the
computation of a function of interest for a particular program
input.
III. MATERIALS AND METHODS
The methodology adopted in the work was a modification
of the Jakstab framework [16]. Jakstab is an Abstract
interpretation-based, integrated disassembly and static analysis
framework for designing analyses on executables and
recovering reliable control flow graphs. In order to make the
framework suitable for the research in this paper, the author
added an extra layer of abstraction to the original framework to
obtain a modified methodology suitable for the task at hand.
While the starting point for the Jakstab framework is binary
source, the approach in this paper performs the analysis on the
program source code.
In addition to the modified Jakstab framework; the
following materials were deployed in building the model-based
high-tech source analysis system.
A bottom-up dynamic slicing technique was defined in
this work and deployed to obtain a hybrid-tech static
analysis model (Fig. 1). A slice was constituted by an
executable portion of the original program whose
behavior is, under the same input, indistinguishable
from that of the original program on a given variable
V at point P in the program. Reference [13] had
showed that bottom-up program slicing techniques
could be successfully deployed to transform a large
component-based program into a smaller one that
contains only statements relevant to the computation of
a given function.
Statecharts notations used in [17] were deployed as a
concise visual formalism that captured the dynamic
behaviour of a system in representing program slices in
this work. Illustrated below is a visualization of a login
module Passwd.pas (Fig. 2).
Goal models- a graph structure representing
stakeholder goals and their inter-dependencies was
deployed to decompose goals into sub-goals through
AND/OR refinements (Fig. 3).
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
186 | P a g e
www.ijacsa.thesai.org
Figure 1. Source model-based slicing for static analysis


Figure 2. Sample statechart notation


Figure 3. Patterns to extract goal models from abstract code [17]
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
187 | P a g e
www.ijacsa.thesai.org

Figure 4. The model-based high-tech source analysis system
IV. RESULTS AND DISCUSSIONS
This work described a model-based high-tech methodology
for static source analysis which consisted of systematically
using slices of code source models as primary source artifacts
throughout the analysis process (Fig.1). This was because,
though proving program correctness may be wrong at static
analysis phase, it is possible to prove simple properties of
simplified models. The process took as input, the program
source text and generated slices based on slicing criterion.
The knowledge schema at this point comprised of the
knowledge of the syntax and semantics of the programming
language (Fig.4). Low-tech static analysis that involved simple
software inspection and manual checking of simple syntactic
standards were first carried out. This was followed by the high-
tech static analysis that included enforced syntactic checks,
check for conformance with respect to specifications, designs,
and the code data flow analysis. The model was bottom-up and
generated slices were analyzed, refined if necessary and
evaluated. The sum of all analyzed parts (slices) gave the
whole (entire source) at the end of the day which may
eventually become a candidate for analysis at a later time.
Generated slices were represented by statecharts and resolved
into goal models.
The Static analysis model was applied to a real-life legacy
application [13]. A procedural application developed for a
commercial bank in Nigeria prior to the consolidation of the
banking sector in the country was chosen due to its high
availability and statecharts were deployed to show the abstract
description of the behaviour of the source system. Fig. 5
showed the statechart that implemented the login page of the
application, passwd.pas.
Visualizations like the one above (Fig. 5) were built for
different slices of the application and the different
visualizations were combined to obtain a high-level meta-
model that contained the entire description of the original
system (Fig. 6).
The top-level statechart above was converted to an
annotated goal model using the conversion process described
earlier (Fig. 3), all the transitions were converted into goals
using the AND/OR decomposition rules. Some tasks in the
goal model contributed to quality concerns modeled by the
softgoals; for example, correctPassword contributed to the
security concern while ErrorMessage contributed to the
usability concern (Fig.7).

Figure 5. Sample visualization of the login module using Statechart
V. CONCLUSION
Previous research had showed that the maintenance of
existing software source code consumed up to of the
resources in the total software budget (time and material), and
that the earlier an error is determined, the lower is the cost of
its correction. In addition, [18] opined that code comprehension
require 47% and 62% of the total time for enhancement and
correction tasks, respectively. The main justification for this
work therefore is to leverage on the most important advantage
of static code structure analysis in re-tooling software
maintenance for a developing economy- cost saving. In this
work, a framework for code structure analysis for a large
component based software system with program slicing was
developed as a re-tooling technique for developing economies
in order to enable an early detection of bugs during a software
development process thereby saving significant costs in
software maintenance.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
188 | P a g e
www.ijacsa.thesai.org

Figure 6. Sample top-level Statechart of the Procedural application

Figure 7. Sample goal model for the application [13]
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
189 | P a g e
www.ijacsa.thesai.org
REFERNCES
[1] B. Livshits, Improving Software Security with Precise Static and
Runtime Analysis Section 7.3 Static Techniques for Security,
Stanford doctoral thesis, 2006. http://research.microsoft.com/en-
us/um/people/livshits/papers/pdf/thesis.pdf
[2] B. A.Wichmann, A. A. Canning, D. L. Clutterbuck, L. A. Winsbarrow,
N. J. Ward, and D. W. R. Marsh, Industrial Perspective on Static
Analysis, Software Engineering Journal vol 10, pp. 69-75, March,
1995.
[3] B. W. Boehm, Software engineering economics, Prentice-Hall,
Englewood Cliffs, NH, 1981.
[4] C. McClure, The three Rs of software automation, Prentical Hall,
Englewood Cliffs, NJ, 1992.
[5] A.F. Ackerman, L. S. Buchwald, and F. H. Lewski. Software
Inspections: An Effective Verification Process, IEEE Software, Vol. 6,
No. 3, May 1989, pp. 31-36.
[6] K. Erdos, & H. M. Sneed, "Partial Comprehension of Complex
Programs enough to perform maintenance," Proceedings of the IEEE
Sixth International Workshop on Program Comprehension, June 24 26,
1998.
[7] S. McConnell, Code Complete, 2nd ed., Microsoft Press: Paperback,
2004, 914 pages, ISBN: 0-7356-1967-0.
[8] R. Klosch, Reverse Engineering: Why and How to Reverse Engineer
Software, Proceedings of the International Conference on
Software Engineering, 2001, pp. 123-132.
[9] K. Thiagarajan, C.Saravanakumar, G. Poonkuzhali, Ponnammal
Natarajan, and S.Jeyabharathi, Static program slicing for composite
data using FSM-Model, World Academy of Science, Engineering and
Technology Journal, Issue 32 Aug. 2009, pp. 820-824. Downloaded
December 2011 from http://www.waset.org/journals/waset/v32.php
[10] A. Ireland, Static Analysis Techniques, Lecture notes on F28SD2:
Software Design, School of Mathematical and Computer Science,
Heriot-Watt University, Edinburgh
[11] Wikipedia the free encyclopedia, Static Program Analysis,
Downloaded November 2011 from
http://en.wikipedia.org/wiki/Static_program_analysis
[12] P. Jones, "A formal methods-based verification approach to medical
device software analysis". Journal of Embedded Systems Design, 2010.
[13] O.F. Oladipo, Software reverse engineering of legacy applications,
Ph.D. Dissertation Computer Science Department, Nnamdi Azikiwe
University, Awka Nigeria, March 2010, unpublished
[14] B. Korel and J. Laski. Dynamic program slicing, Information
Processing. Letters, vol. 29, no 3, pp.155-163, Oct. 1988.
[15] M. Weiser. "Program slicing". IEEE Transactions on Software
Engineering, vol. 10, Issue 4, pages 352357, IEEE Computer
Society Press, July 1984.
[16] Jakstab Framework homepage http://www.jakstab.org/
[17] Y. Yu, Y. Wang, J. Mylopoulos, S. Liaskos, A. Lapouchnian, and J.
Cesar Sampaio do Prado Leite, Reverse Engineering Goal Models from
Legacy Code, In: 13th IEEE International Conference on Requirements
Engineering (RE'05), 29 Aug-2 Sept 2005, Paris, France, Downloaded
October 2009 from www.cs.toronto.edu/~alexei/pub/re05re.pdf
[18] M. L. Nelson, A Survey of Reverse Engineering and Program
Comprehension, Lecture notes on CS 551: Software Engineering
Survey, Old Dominion University, April 19, 1996, Downloaded
November 2011 from arxiv.org/pdf/cs/0503068
AUTHORS PROFILE
Oladipo, Onaolapo Francisca holds a Ph.D in Computer
Science from Nnamdi Azikiwe University, Awka, Nigeria;
where she is currently a faculty member. Her research
interests spanned various areas of Computer Science and
Applied Computing. She has published numerous papers
detailing her research experiences in both local and
international journals and presented research papers in a number of
international conferences. She is also a reviewer for many international
journals and conferences. She is a member of several professional and
scientific associations both within Nigeria and beyond; they include the
British Computer Society, Nigerian Computer Society, Computer
Professionals (Regulatory Council) of Nigeria, the Global Internet
Governance Academic Network (GigaNet), International Association of
Computer Science and Information Technology (IACSIT ), the Internet
Society (ISOC), Diplo Internet Governance Community and the Africa ICT
Network.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
190 | P a g e
www.ijacsa.thesai.org
Transform Domain Fingerprint Identification Based
on DTCWT
Jossy P. George
Department of Computer Science
Christ University
Bangalore, Karnataka, India
Abhilash S. K., Raja K. B.
Dept. of Electronics and Communication Engineering
University Visvesvaraya College of Engineering
Bangalore, Karnataka, India

Abstract The physiological biometric characteristics are better
compared to behavioral biometric identification of human beings
to identify a person. In this paper, we propose Transform
Domain Fingerprint Identification Based on DTCWT. The
original Fingerprint is cropped and resized to suitable dimension
to apply DTCWT. The DTCWT is applied on Fingerprint to
generate coefficient which form features. The performance
analysis is discussed with different levels of DTCWT and also
with different sizes of Fingerprint database. It is observed that
the recognition rate is better in the case of level 7 compared to
other levels of DTCWT.
Keywords-Fingerprint; DTCWT; Euclidean Distance; Preprocessing
I. INTRODUCTION
The term biometric is derived from the Greek word bio
(life) and metrics (to measure). The biometrics identifies the
person, based on feature vectors derived from physiological or
behavioral characteristics. The biometric traits must satisfy
universality, uniqueness, permanence, accessibility,
collectability. The physiological biometrics are Fingerprint,
Hand Scan, Iris Scan, Facial Scan and Retina Scan etc. and
behavioral bio-metric are Voice, Keystroke, Gait, Signature
etc. The physiological biometrics traits are almost constant
throughout the life span of a person [1], even for identical
twins. The behavior biometric trait varies with mood and
environment.
The bio-metric authentication is better than traditional
methods. The existing traditional methods to authenticate are
Passwords, Personnel Identification Numbers (PINs), Tokens
and Smart Cards. The disadvantages of traditional methods are
(i) Passwords and Pins are difficult to remember. (ii) More
chances of losing tokens and smart cards. (iii) The misuse of
traditional methods of authentication by miscreants is very
high, especially in the case of money transaction through
ATMs, access to the unauthorized places, etc.
Biometric traits are identified in different domains, such as
spatial, transformation and hybrid domain. In spatial domain
biometric traits are considered as image itself, wherein features
of image are obtained by computing the area, height, width,
pixel density, mean, variance and standard deviation etc. In
transform domain, the spatial domain image is converted into
other domain using Fast Fourier Transform (FFT), The
Discrete Cosine Transform (DCT), Discrete Wavelet
Transform (DWT), Dual-Tree Complex Wavelet Transform
(DTCWT), Empirical Mode Decomposition (EMD), Singular
Value Decomposition (SVD), Principal Components Analysis
(PCA), Independent Component Analysis (ICA), etc. These
coefficients of transform domain form feature vectors. In
hybrid domain, the combination of spatial and transformation
domain features are obtained. The features of spatial,
transformation and hybrid domains are compared using
distance vector formulae such as Equivalent Distance (ED),
Hamming Distance (HD), Chi-square test, or classifiers such as
Neural Networks (NN), Support Vector Machine (SVM) and
Random Forest (RF).
Fingerprint is a physiological trait which is constant
throughout the life span of a person. The fingerprint is unique
even for the twins. It has been used as the identification
technique for over a century. The fingerprint authentication
officially established as a means of identifying people around
1900s. This authentication got the popularity because; the
devices for accessing the fingerprints are small and
inexpensive. When a biometric verification is to occur, a scan
of the biometric of a person is made and which is to be
compared with the stored data of the same person.
A fingerprint usually appears as a series of dark lines that
represents the high, peaking portion of the friction ridge skin,
while the valleys between these ridges appears as white space
and are the low, shallow portion of the friction ridge skin,
minutiae, or the location and direction of the ridge endings and
bifurcations along a ridge path. The upper most point on the
inner most ridge of the fingerprint image is known as core. The
fingerprints are mainly classified as Arch, Tented Arch, Whorl,
Left Loop, and Right Loop. There are separating point between
pattern area and non-pattern area in many of the fingerprint
images. These points are generally known as delta.
A. Contribution: In this paper the Transform Domain
Fingerprint Identification Based on DTCWT at different
levels is introduced. DTCWT generates complex
coefficients by using a dual tree of wavelet filters to obtain
two parts of the images, i.e., real and imaginary part.
DTCWT is applied on Fingerprint to generate different
levels to obtain Fingerprint features.
B. Organization: The Introduction is given in section I, the
existing research papers are discussed in section II, the
proposed model is explained in section III, the algorithm is
described in section IV, the performance analysis is
discussed in section V and finally, conclusion is given in
section VI.

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
191 | P a g e
www.ijacsa.thesai.org
II. LITERATURE SURVEY
The Fingerprint for individual identification and
verification with existing different technique and applications
are described in this section.
S. Vasuki et. al., [2] proposed a model for segmentation of
a color textured regions of a given images. This is obtained by
the segmentation and by applying DTCWT. This model works
in two levels where in first level, after applying DTCWT, the
image is divided into 16 sub images and from where the
maximum energy sub image is selected as an optimum feature
space. In the second level, K-means spatial refinement
algorithm is applied. The main advantage of this model is the
accuracy. Zhao Song and Liu Yuanpeng [3] gave a novel image
diagnosing scheme by applying 2-D DTCWT to the second
bandlet transform. This is obtained based on the shift
invariance and better directionality of the DTCWT. The
bandlet reconstruction recovers the transitions and directional
textures. This improves significantly image diagnosing results.
Chen Feng and Yu Song nian [4] proposed a model which
retrieves the multi scale image by using a new class of image
features as the image descriptors from DTCWT. From this
work, we can conclude that the performance of DTCWT is
better in the experiment on the stander COREL image database
due to the rotation invariance, translation invariance, robust to
noise and getting the key point according to peoples cognitive
habits. Sathesh and Samuel Manoharan [5] discussed on
advantages and disadvantages of DWT. Also they give the
methods to overcome the limitations of DWT and the
theoretical analysis of complex wavelet transform and its
verification using the simulated images. V. Lulian and B.
Monica [6] proposed a method to design and optimize
separately two channels perfect reconstruction filter banks.
This method ensures the good quality for the two levels. This
method is more useful where the only the few levels of
decomposition is required.
Shahid and Gupta S [7] proposed a novel method to fuse an
image using the DTCWT. This is achieved through the
formation of a fused pyramid using DTCWT coefficients from
the decomposed pyramids of a source image. This method
gives better qualitative and quantitative results than the DWT
methods. Yun and Cho [8] proposed an adaptive preprocessing
method, which extracts five features from the fingerprint
images, analyzes image quality with clustering method, and
enhances the images according to their characteristics. The
preprocessing is performed after distinguishing the fingerprint
image quality according to its characteristics. Brankica M.
Popovic and L. Maskovic [9] used multi scale directional
information obtained from orientation field image to filter the
spurious minutiae. The feature extraction in pattern recognition
system is to extract information from the input data and
depends greatly on the quality of the images. Multi scale
directional information estimated based on orientation field
estimation. F. A. Afar et. al., [10] presented the minutiae based
Automatic Fingerprint Identification Systems. The technique is
based on the extraction of minutiae from the thinned, binarized
and segmented version of a fingerprint image. The system uses
fingerprint classification for indexing during fingerprint
matching. G. Jagadeeswar Reddy et. al., [11] presented
fingerprint diagnosing using both wavelet and Curvelet
Transforms. The search-rearrangement method performs better
than minutiae based matching for fingerprint binary constraint
graph matching since it does not require implicit alignment of
two fingerprint images.
K. Zebbiche and F. Khelifi [12] presented biometric images
as one Region of Interest (ROI) that has the data processed by
most biometric based system. The scheme consists of
embedding the watermark into ROI in fingerprint images.
Discrete Wavelet Transform and Discrete Fourier Transform
are used. Bhupesh Gour et. al., [13] introduced midpoint ridge
contour representation in order to extract the minutiae from
fingerprint images. Color coding scheme is used to scan each
ridge only once. Seung Hoon chae and Jong Ku Kim [14]
proposed Fingerprint Verification in which both minutiae and
ridge information are used to reduce the errors due to
incomplete alignment or distortion. This works gives more
importance to the areas where we get errors in processing the
algorithms. Aparecido Nilcau Marana and A. K. Jain [15]
proposed Ridge Based Fingerprint matching using the Hough
transform. The major straight lines that match the fingerprint
ridges are used to estimate rotation and translation parameters.
This method gives better and accurate results. Anil K Jain et
al., [16] described the use of logistic regression method to
integrate multiple fingerprint matching algorithms. The
integration of Hough transform based matching, string distance
based matching and 2D dynamic programming based matching
using the logistic regression has minimized the False Rejection
Rate for a specified level of False Acceptance Ratio.
Fanglin Chen and Jie Zhou [17] proposed an algorithm for
reconstructing fingerprint orientation field from saved minutiae
and are used in the matching stage to compare with the
minutiae from the query fingerprint. The orientation fields
computed from the saved minutiae is a global feature and the
saved minutia is the local feature, are used to get more
information. Chunxian Ren and Yilong Yin [18] used the
hybrid algorithm based on linear classifier to segregate
foreground and background blocks. The pixel wise classifier
uses three pixel features such as Coherence, mean and
variance. Hartwig Front haler and Klaus Kollreider [19] used a
multi grid representation of a discrete differential scale space
enhancement strategy of fingerprint recognition system. The
fingerprint image is decomposed using Laplacian Pyramid as
relevant information is concentrated within a few frequency
bands. The Fausian Directional Filtering is used to enhance
ridge valley pattern of fingerprint using 1-D filtering on higher
pyramid level. The linear symmetric features are used to
extract the local ridge valley orientation. D. R. Shashi Kumar
et. al., [20] proposed fingerprint verification based on fusion of
minutiae and ridges using strength factors. In this model, to
extract minutiae and ridges, block filter and Hough Transform
are used. This proposed algorithm gives better results than the
many other existing ones. This is achieved by fusing the
minutiae and the ridge parameters using strength factors.
III. MODEL
In this section, definitions of Performance Analysis and
proposed model are discussed.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
192 | P a g e
www.ijacsa.thesai.org
No. of persons accepted from out of database
Total no. of persons in database. (1)
No. of correct persons rejected
Total no. of persons in database. (2)

No. of persons correctly Matched in the database
Total no. of persons in database. (3)

A. Definitons
a) False Accept Rate (FAR): It is the probability that
system incorrectly matches with images stored with input
image database. The FAR can be calculated using Equation 1.

FAR =
b) False Rejection Rate (FRR): It is the probability that
system fails to recognize the correct pattern to match with the
database images. It is the ratio of number of correct persons
rejected in the database to the total number of persons in
database and can be calculated using Equation 2.

FRR =

c) Equal Error Rate (EER): It is the value where both
the reject and accept rates are equal.

d) True Success Rate (TSR): It is the ratio of total
number of persons correctly matched in the database to the
total number of persons in the database and is given by
Equation 3.
TSR =

B. Proposed Model
The proposed model of Fingerprint Identification using
DTCWT is given in the Fig. 1.

Figure 1. Proposed model
a) Fingerprint Database: Database collection is one of
the important works in testing the biometric system. The best
and advisable way of collecting the database is with a different
sensor. There are different databases made available for the
researchers to study on the different biometrics systems. The
Fingerprint database available are the first, second and third
International Competition on Fingerprint verification such as
FVC 2000, FVC 2002 and FVC 2004 [21] respectively.
Four distinct databases for FVC 2004 provided by the
organizers constitute the benchmark: DB1, DB2, DB3 and
DB4. Each database is 110 fingers wide and 8 samples per
finger in depth i.e., it consists of 880 fingerprint images. Each
database is partitioned into disjoint subsets A and B.
The subsets DB1-A, DB2-A, DB3-A and DB4-A, which
contain the first 100 fingers (800 images) of DB1, DB2, DB3
and DB4, respectively, is used for the algorithm performance
evaluation. The subsets DB1-B, DB2-B, DB3-B and DB4-B,
containing the last 10 fingers (80 images) of DB1, DB2, DB3
and DB4 respectively made available to the participants to
allow parameter tuning before executable(s) submission.
The Fig. 2 shows a sample image from each database. The
DB3_A database is considered to test our algorithm due to its
high resolution and size compatibility. A sample of finger print
with eight impressions of DB3 is given in Fig. 3.


Figure 2. One fingerprint image from each database

Figure 3. Fingerprint sample of DB3_A
Source Database: The first seven Fingerprint images
of each person from DB3 _A database of FVC 2004
are stored.
Test Template: The eighth Fingerprint of each person
from DB3 _A database of FVC 2004 are used in the
test template and is compared with source database to
compute FRR and TSR.
Mismatch template database: The DB3_B of FVC
2004 database of 10 fingers are stored in Mismatch
template database and compared with source database
to compute FAR.
b) Pre-processing: The original Fingerprint image is of
size 480 X 300. An observing the DB3_A of FVC 2004, we
crop the input image to the size 401 X 201 in order to remove
the unwanted portion in the image. And then the cropped image
is resized into 512 X 256 for the DTCWT requirement.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
193 | P a g e
www.ijacsa.thesai.org
c) DTCWT: Dual Tree Complex Wavelet Transform is a
recent enhancement technique to the Discrete Wavelet
Transform with some additional properties and changes. It is a
effective method for implementing an analytical wavelet
transform, introduced by Kings bury in 1998 [22], [23], [24].
DTCWT gives the complex transform of a signal using two
separate DWT decompositions ie., tree a and tree b.
DTCWT produces complex coefficients by using a dual tree of
wavelet filters and gives real and imaginary parts which are
shown in Fig 4.

Figure 4. Real and imaginary parts of the complex coefficients

The DTCWT has following properties:
Approximate shift invariance;
Good directional selectivity in 2-dimensions (2-D)
with Gabor-like filters also true for higher
dimensionality: m-D);
Perfect reconstruction (PR) using short linear-phase
filters;
Limited redundancy: independent of the number of
scales: 2:1 for 1-D ( 2m :1 for m-D);
Efficient order-N computation - only.
DTCWT differentiates positive and negative
frequencies and generates six sub bands oriented in
15, 45, and 75 . The different levels of DTCWT
such as levels 5, 6, and 7 are applied on pre-processed
Fingerprint Image.
d) Matching: The Euclidian Distance (ED) is used to
verify the test image with the database images using the
Equation 4.
( )

(4)
Where,
M = the dimension of feature vector.
Pi = is the database feature vector.
qi = is the test feature vector.
IV. ALGORITHM
A. Problem Definition
The physiological trait Fingerprint is used to identify a
person using the features obtained by the coefficients of
DTCWT. The proposed algorithm for the performance analysis
of the fingerprint identification for different levels of DTCWT
is given in the Table1.
The objectives are;
Fingerprint verification to authenticate a person using
DTCWT
To achieve high TSR
To have FRR and FAR very low
While applying the DTCWT in different levels, the number
of features and dimensions are reduces. The Fingerprint images
for level-1, level-2, level-3, level-4 and level-5, level-6, level-7
are shown in Fig. 5.
TABLE 1. PROPOSED ALGORITHM.


1) FVC 2004, DB3_A database is considered.
2) Pre-processing is done by cropping the input
fingerprint image to 401 X 201.
3) Cropped image is resized to 512 X 256 for DTCWT
requirement.
4) DTCWT is applied on Fingerprint with levels 5, 6
and 7.
5) Magnitude and phase resulted from DTCWT are
concatenated and considered as features.
6) Source database is created with the features
obtained by step 5.
7) For the test fingerprint DTCWT is applied and
features obtained using step 5.
8) Test Fingerprint is compared with the database
fingerprint using ED to verify a person
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
194 | P a g e
www.ijacsa.thesai.org

Figure 5. DTCWT images at different levels.
V. PERFORMANCE ANALYSIS
For the performance analysis, DB3_A of FVC 2004
Fingerprint database is considered. The number of persons
inside the database (PIDB) to compute FRR and TSR are
varied from 30 to 90 and the number of persons outside the
database (PODB) are 10 to compute FAR is given in Table 2.
It is observed from the table 2 that the values of EER and
TSR depend on the quality of Fingerprint image than the
number of images in PIDB and PODB. The values of EER and
TSR are better in the case of PIDB: PODB of 40:10. The
performance of recognition rate is better in DTCWT level 7
compared to other lower levels of DTCWT. The TSR and EER
is 85% and 0.15 respectively for DTCWT level 7 with PIDB:
PODB of 40:10.
The graph for FRR, FAR and TSR is given in Fig. 6 and the
variations of FRR and TSR with threshold for POIB: PODB of
40:10 is tabulated in Table 3 and It is noticed that as threshold
increases, the value of FRR decreases, whereas the values of
FAR and TSR increases. The highest success rate of
recognition of 85% is achieved for the threshold value of 2.4.
VI. CONCLUSION
The Fingerprint biometric is used to authenticate a person.
In this paper, Transform Domain Fingerprint Identification
Based on DTCWT is proposed. The Fingerprint is preprocessed
to a suitable size that suit DTCWT. The Fingerprint features
are obtained by applying DTCWT with different levels. The
test image features are compared with Database images using
Euclidean Distance. It is observed that the recognition rate is
better in the case of DTCWT level 7 compared to lower levels
with PIDB: PODB of 40:10. In future, the algorithm may be
combined with spatial domain features such as global and local
features to enhance recognition rate.
TABLE 2. EER AND TSR FOR DIFFERENT LEVELS OF DTCWT




Figure 6. Variations of FRR, FAR and TSR with respect to different thresholds
Level-2
Level-6
Level-3
Level-1
Level-7
Level-5
Level-4
Threshold
F
R
R
,

F
A
R

&

T
S
R

(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
195 | P a g e
www.ijacsa.thesai.org
TABLE 3. VALUES OF FRR, FAR AND TSR FOR DIFFERENT
THRESHOLDS
Threshold FRR FAR TSR %
0 1 0 50
0.3 1 0 50
0.6 1 0 50
0.9 1 0 50
1.2 0.85 0 50
1.5 0.6 0 60
1.8 0.35 0 64
2.1 0.15 0.1 72
2.4 0.1 0.4 85
2.7 0.05 0.4 80
3 0 0.4 74
3.3 0 0.5 72
3.6 0 0.8 60
3.9 0 0.8 55
4.2 0 0.8 50
4.3 0 0.9 50
4.8 0 0.9 50
5.1 0 1 50
5.4 0 1 50
5.7 0 1 50
6 0 1 50
REFRENCES
[1] B. N. Lavanya, K. B. Raja and K. R. Venugopal, Fingerprint
verification based on gabor filter enhancement, International Journal of
Computer Science and Information Security,vol. 6, no. 2, pp. 138-144,
2009.
[2] S. Vasuki, L. Ganesan and R. Florin Raja Singh, DT- WT based
segmentation algorithm for color images, RTCSP Conference
Proceedings, Coimbatore, pp. 207-212, 2009.
[3] Zhao Song and Liu Yuanpeng, A novel image denosing scheme via
combining dual tree complex wavelet transform and bandelets, IEEE
International Symposium on Intellignet Information Technology
Application, pp. 509-512, 2009.
[4] Chen Feng and Yu Song-nian, Content based image retrieval by
DTCWT feature, IEEE International Conference on Computer Research
and Development, vol. 4, pp. 283286, 2011.
[5] Sathesh and Samuel Manoharan, A dual tree complex wavelet
transform construction and its application to image denosing,
International Journal of Image Processing, Vol. 3, pp. 293-300, 2010.
[6] Lulian Voicu and Monica Borda, New method of filters design for dual
tree complex wavelet transform, IEEE International Symposium on
Signals, Circuits and Systems, vol. 2, pp. 437-440, 2005.
[7] Mohd Shahid and Sumna Gupta, Novel masks for multimodality image
fusion using dtcwt, IEEE International Conference, TENCON, pp. 1-6,
2005.
[8] E. K. Yun and S. B. Cho, Adaptive fingerprint image enhancement
with fingerprint image quality analysis, International conference of
Image and Vision Computing, pp. 101110, 2006.
[9] M. P. Brankica and L. Maskovic, Fingerprint minutiae filtering based
on multiscale directional information, FACTA Universitatis-Series:
Electronics and Energetics, vol. 20, pp.233-244, August 2007.
[10] F. A. Afsar, M. Arif and M. Hussain, Fingerprint identification and
verification system using minutiae matching, National Conference on
Emerging Technologies, pp.141-146, 2004.
[11] G. Jagadeeswar Reddy, T. Jaya Chandra Prasad and M. N. Giri Prasad,
Fingerprint image denoising using curvelet transform, Proceeding of
Asian Research Publishing Network Journal of Engineering and Applied
Sciences, vol 3, no 3, pp. 31-35, June 2008.
[12] K. Zebbiche and F. Khelifi Region-based watermarking of biometrics
images: Case study in fingerprint images, Proceedings of International
Journal of Digital Multimedia Broadcasting, pp. 1-13, 2008.
[13] Bhupesh Gour, T. K. Bandopadhyaya and Sudhir Sharma, Fingerprint
feature extraction using midpoint ridge contour method and neural
network, Proceedings of International Journal of Computer Science and
Network Security, vol.8, no.7, pp. 99-103, July 2008.
[14] Seung-Hoon Chae and Jong Ku Kim Ridge-based fingerprint
verification for enhanced security, Digest of Technical Papers
International Conference on Consumer Electronics, pp 1-2, 2009.
[15] A. N. Marana, and A. K. Jain, Ridge-based fingerprint matching using
hough transform, IEEE Proceedings of the Brazilliab Symposium on
Computer Graphica and Image Processing, pp. 112-119, October 2005.
[16] A. K. Jain, S. Prabhakar, and A. Chen, Combining multiple matchers
for a high security fingerprint verification system, Pattern Recognition
Letters, Elsevier Science Direct, vol.20, pp 1371- 1379, 1999.
[17] Fanglin Chen and Jie Zhou, Reconstructing orientation field from
fingerprint minutiae to improve minutiae-matching accuracy, IEEE
Transactions on image processing, vol. 18, no 4, pp 1665-1670, 2009.
[18] Chunxiao Ren and Yilong Yin, A linear hybrid classifier for fingerprint
segmentation, Fourth International Conference on Neural Computation,
pp 33-37, 2008.
[19] Hartwig Fronthaler and Klaus Kollreider, Local features for
enhancement and minutiae extraction in fingerprints, IEEE Transactions
on Image Processing, vol. 17, no 3, pp 354-363. 2008.
[20] D. R. Shashi Kumar, R. K. Chhotaray, K.B. Raja and Sabyasachi
Pattanaik, Fingerprint verification based on fusion of minutiae and
ridges using strength factors, International Journal of Computer
Applications, vol. 4, no.1, 2010.
[21] D. Maio, D. Maltoni, R. Cappelli, J.L. Wayman and A. K. Jain,
FVC2004: Third Fingerprint Verification Competition, Lecture Notes
in computer science, pp.1-5, 2004.
[22] N.G. Kingsbury, The dual-tree complex wavelet transform: A new
technique for shift invariance and directional filters, Proceeding of 8th
IEEE DSP Workshop, Utah, 1998.
[23] N.G. Kingsbury, Image processing with complex wavelets, Philos.
Trans. R. Soc. London A, Math. Phys. Sci., vol. 357, no. 1760, pp.
25432560, 1999.
[24] N.G. Kingsbury, Complex wavelets for shift invariant analysis and
filtering of signals, Appl. Comput. Harmon. Anal., vol. 10, pp. 234
253, 2001.
AUTHORS PROFILE
Jossy P. George currently serves as Assistant Professor of
the Department of Computer Science at Christ University,
Bangalore. He received the B. Sc. In Computer Science,
Bachelor of Philosophy (B. Ph), Bachelor of Theology (B.
Th) and Master of Computer Application (MCA). He has
done his FDPM from IIM, Ahmedabad. He is pursuing his
Ph.D. in Computer Science of Christ University under the
guidance of Dr. K. B. Raja, Assistant Professor, Visvesvaraya College of
Engineering. His research activities focus on algorithms for improved
accuracy in fingerprint and iris biometrics.

Abhilash S. K. obtained his BE in Electronics and
Communication Engineering from DR Ambedkar Institute
of Technology, Bangalore and is pursing his Postgraduate
(ME) in Visvesvaraya college of Engineering, Bangalore.
His research interests include Image processing and
Biometrics.

Dr. K B Raja is an Assistant Professor, Department of
Electronics and Communication Engineering, University
Visvesvaraya college of Engineering, Bangalore
University, Bangalore. He obtained his BE and ME in
Electronics and Communication Engineering from
University Visvesvaraya College of Engineering,
Bangalore. He was awarded Ph.D. in Computer Science
and Engineering from Bangalore University. He has 85
research publications in refereed International Journals and
Conference Proceedings. His research interests include Image Processing,
Biometrics, VLSI Signal Processing, computer networks.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
196 | P a g e
www.ijacsa.thesai.org
Effective Security Architecture for Virtualized Data
Center Networks
1
Udeze Chidiebele. C,
3
Okafor Kennedy .C
1,3
R & D Department, Electronics Development Institute
(FMST-NASENI), Awka, Nigeria.
2
Prof. H. C Inyiama,
4
Dr C. C. Okezie,
2,4
Electronics and Computer Engineering Department,
Nnamdi Azikiwe University, Awka, Nigeria

AbstractThis work presents a candidate scheme for effective
security policy that defines the requirements that will facilitate
protection of network resources from internal and external
security threats. Also, it ensures data privacy and integrity in a
virtualized data center network (VDCN). An integration of Open
Flow Software Defined Networking (OFSDN) with VLAN Virtual
Server Security (VVSS) architecture is presented to address
distinct security issues in virtualized data centers. The OFSDN
with VVSS is proposed to create a more secured protection and
maintain compliance integrity of servers and applications in the
DCN. This proposal though still on the prototype phase, calls for
community driven responses.
Keywords- Infrastructure; Virtualization; VDCN; OFSDN; VVSS;
VLAN; Virtual Server.
I. INTRODUCTION
Recently, data center networks (DCNs) have attracted a lot
of interest in the enterprise networking industry. DCNs are
used to provide data storage and files transfer where end
stations are interconnected as clusters and bladed systems [1].
A data center represents the heart of any organizations
network [2]. Companies rely on the data stored in the data
center to interact with its employees and customers.
The proliferation of the Web-based technologies makes the
data center more vulnerable to security attacks. Any security
attack on the data center can destroy the whole organizations
network and data [2]. Besides throughput and low latency
required in DCNs, the security considerations of enterprise data
centers is also very critical. Several researches were dedicated
to the security issues and the design constraints of large scale
data centers from different points of view [2]. The authors in
[2], [3], [4], [5], [6] discussed on the data center security
problems, technologies, security strategies such as
consolidation, relocation, migration, expansion and review of
asset management policies. The authors of [4] carried out an
overview of the communication network design problems that
arise with large numbers of nodes, links and switch costs.
Some layered security models for addressing complex security
issues are discussed in [5] and [6]. With fast changing
technologies and service demands in DCNs, the need for an
effective open platform secure model becomes very imperative.
In this paper with detailed study on the security proposals
existing in literature, and having considered all the
requirements of network security management for a virtualized
data center model, we propose an effective secured model:
Open Flow Software Defined Networking (OFSDN) with
VLAN Virtual Server Security (VVSS). The design is based on
layered security architecture for virtual servers and open flow
switch architecture. Operational mechanism is presented in
section V with other details. By allowing the MAC controllers
in the virtual open flow switch in our DCN to house the flow
tables for each virtual port, this work creates lines of defense
against any security threat. Unicast, broadcast and multicast
traffic are characterized and monitored by the modeled switch
architecture which serves as an aggregation link buffer.
The paper is organized as follows. In Section II, we
discussed virtualization in data center network, data center
security problems as presented in [2]. In section III, the
proposed security model (OFSDN) is shown with the Virtual
server security system. Section IV gives the experimental setup
for VLAN open flow switch. The paper ends with conclusions
and future directions
II. VIRTUALIZATION IN DATA CENTER NETWORKS
Server virtualization has become popular in data centers
since it provides an easy mechanism to cleanly partition
physical resources, allowing multiple applications to run in
isolation on a single server [7]. Virtualization helps with server
consolidation and provides flexible resource management
mechanisms [7] in DCNs particularly. We quickly add that
Virtualization is not a new technology, but it has regained
popularity in recent years because of the promise of improved
resource utilization through server consolidation. According to
[8], a Data Center is the consolidation point for provisioning
multiple services that drive an Enterprise business. In [2], the
authors enlist the data center hardware and software
components. The hardware components are: firewalls,
Intrusion Detection Systems, contents switches, access
switches and core switches. The software components are:
IPSec and VPN, antivirus software, network management
systems and access control server. However, for effective
security implementation in a virtualized DCN, this work goes
further to propose a more secured data center design that is
programmable, secured with strong isolation, and flexible using
the OFSDN approach in our context.
III. DATA CENTER SECURITY PROBLEMS
Data center networks usually have its security threats. The
work carried out in [8], [9]and[10] discussed some of these
problems, viz: Unauthorized Access, MAC Flooding, ARP
Spoofing, IP Spoofing, Denial of Service (DOS), Viruses,
Worms, Trojans, and internal Security threats. However,
sampled solutions to these problems were given in [2]. We
still argue that these solutions do not completely eradicate
security vulnerabilities in contemporary data center networks.
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
197 | P a g e
www.ijacsa.thesai.org
For a virtualized data center domain, a restructured architecture
which will address the possible lapses in addition to the
outlined remedies in [2], will serve in securing todays
enterprise networks.
IV. DATA CENTER SECURITY TECHNOLOGIES
Information stored at the data center must be protected from
any security threat that may destroy or modify it in any
unwanted way [2]. These security threats can originate from
hackers outside or from inside the data center network.
Different solutions to the security threats can be used together
to achieve the highest possible data protection. Some of these
technologies are:
Firewalls.
Network intrusion detection and prevention systems.
Virtual Local Area Networks (VLAN).
Virtual Private Network (VPN) and IPSec.
Leveraging on these four technologies, our contribution is
shown in the Open Flow Software Defined Network model in
Fig. 2. OFSDN is a layer 2 protocol in the virtual Software
Defined Network (SDN) switch that allows for policy control
via its open flow visor (virtualization layer). This model creates
multiple layers of security for the virtualized DCN controlling
unicast, broadcast and multicast traffics. Section IV and V
discussed in details the security models for highly scalable and
secure virtualized DCN.
V. VLAN VIRTUAL SERVER SECURITY SYSTEM
VLAN Virtual Server Security (VVSS) system proposed in
this work for the server VM provides multi-layered workgroup
segmentation while utilizing the underlying hardware
technology to protect the virtual data center. The VVSS
solution is a generic purpose-built framework proposed for
large scale enterprises. The virtual environment at the core of
the infrastructure is the Vm server running on ESX platform
with its VMware.
Fig. 1 shows the VVSS model while Fig. 5 and Fig. 6 show
the packet tracer simulation. Again, in our architecture shown
in Fig. 2, MAC controllers were assigned to all the network
entities to house their flow tables. For active participation in
the network, the open flow visor must uniquely identify and
authenticate the client node else, the terminal is dropped for
access.
As shown in Fig. 1, VLAN virtual security model was
modeled to be deployed on a virtualized server for various
applications (Vm1Vm5). The kernel utilizes the hypervisor
API to inspect and control the virtual switch network and VM
behavior. Virtual Security Service (VSS) utilizes a subnetted IP
mapping, which is provided as VMsafe for various user groups.
For demonstration in this work, each VM server on virtualized
server is managed and configured through packet tracer
environment.

Figure 1. VLAN Virtual Security System Model.
A VLAN backbone which hosts the Vm server is the
central manager for the applications. VVSS has the following
functions:
Inter-VM migration of Applications and services for
compatibility issues
Virtual Machine generator and monitor
Network Access Control (NAC)
Discovery and Broadcast Isolation
License and Update Management (LUM)
VI. OPEN FLOW SOFTWARE DEFINED NETWORK MODEL
FOR DCN
The Open flow software defined networking switch in
figure 2 is a speed redundant device with isolated MAC
controllers housing the flow tables shown in Fig. 4. An open
flow protocol (OFP) which can be enabled in the switch carries
out control policy (CP), reaction execution (RE) and history
tracking (HT). Once OFP is enabled on the switch, any device
interfaced with the switch is actively monitored as a software
robot, thereby securing the overall network against any form of
threat. This is proposed for virtualized data center in context.
The key security metric is the MAC ID of the interfacing
devices.
The security policy of the flow table in Fig. 4 controls
activities that is handled by conventional VLAN and Access
control list (ACL) such as traffic denial or flow allowance,
routing, broadcast isolation flow, flow detection and
suppression in the OFSDN switch. All servers, etc shown in
Fig. 2 are mapped in the MAC controllers. Fig. 3 shows the
open virtual isolation in the OFSDN switch. This model offers
a highly secured security layer to existing security approaches
in literature.
Vm1 Vm2

Vm3 Vm4 Vm5
Applications Layer
OS
Kernel
Hypervisor
NIC
RAM HDD CPU
I/Os
USER ACCESS
INTERFACE
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
198 | P a g e
www.ijacsa.thesai.org


Figure 2. OFSDN Security Model for DCN










Figure 3. A Virtualized Open Flow Switch









Figure 4. Flow table Ingress in Open Flow Switch Model
VII. EXPERIMENTAL SETUP
The first phase experiment involves virtualizing the server
test bed consisting of one standard HP machine with a dual-
core Intel Xeon processor connected to a rack-mounted disk
enclosure with a Small Computer Scale Interface (SCSI)
backplane running on ESX linux sever.
For the purposes of trace security, six VLANs were created
for the server and simulated with packet trace tool. In the
server, a Seagate model 15,000RPM disks: of size 1TB was
considered with a RAM of 6GB. The server was connected via
a switched (OFSDN) 1Gbps Ethernet link.
This work provides three fundamental security services:
Data confidentiality: protecting against unauthorized
access to data being transmitted.
Data integrity: protecting against alteration or future
replay of traffic.
Source authentication: network addresses are
authenticated as part of the protocol.

Server
Virtualization
13s:36v:79d
IDS
System&
Proxy
Server
OFSDN
Switch
Network
Management
Server, LOG,
VMs
Firewall/Gate
way/ Load
balancer
MAC
Controller#4

MAC
Controller#3

MAC
Controller#2

MAC
Controller#1

OPEN FLOW VISOR- Virtualization Layer
Controller#5
MAC #5
Controller#6
MAC #6

Controller#7
MAC #7

Controller#8
MAC #8

Controller#1
MAC #1

Controller#2
MAC #2

Controller#4
MAC #4

Controller#3
MAC #3

VLAN#4

VLAN#3

VLAN#2

VLAN#1

VLAN#5

VLAN#6

VLAN#7

VLAN#8

POLICY REACTION HISTORY
Switch. Port MAC IDMAC DST ID Ethtype VLAN ID IP SRC IP DSTNIP PROTOCOLTCPSRC PORT TCP DST PORT


#1.Flood frame Packets to Physical Interfaces
# 2. Encrypt, Encapsulate and forward to
MAC Controller # 3. Drop &Queue Frame
Packet #4. Migrate to Process Pipeline
.
Frame packet Byte count CRC
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
199 | P a g e
www.ijacsa.thesai.org
We deployed Classless Inter-domain Routing (CIDR)
approach to generate usable IP for the VM server and users on
the network. For valid IP range for 200 users with 128Vm
servers, we used a class valid host range: 192.168.10.1 to
192.168.10.199 with a subnet mask of 255.255.255.0. For
effective security and broadcast isolation, virtual IP mapping
on the Vm server enables the hosts, guests and clients to
communicate with each other. Fig. 5 and Fig. 6 show the
packet flow in the packet tracer integrated development
environment (IDE).
TABLE 1: DATA CENTER VM SERVERS (13 SERVERS, 36
VOLUMES, 79 DISKS)
TABLE 2: AVERAGE UTILIZATION RATES.
Resource Utilization
CPU 6%
MEMORY 40%
NETWORK I/O <5%
DISK I/O <5%

Figure 5. DCN VLAN workgroup Model with ESx server

Figure 6. Simulated Packet Flow for secured DCN
VIII. CONCLUSION
The effective security architecture discussed in this paper is
conceived to achieve the best possible solution for virtualized
data center networks. Owing to advancements in virtualization
technology, the security methodologies for traditional data
centers which includes: firewalls, intrusion detection
system/intrusion protection system, virtual local area network
(VLAN) and virtual private network (VPN) cannot effectively
handle security implications of a virtualized data center
networks. This work presents an effective open flow software
defined network switch with VVSS model and with emphasis
on VLAN virtualization on ESX server to ensure total security
of the critical data in the virtualized data center network.

VmServers VLAN Volumes IP Mapping
UserV
M
10 3 192.168.10.2
ProjectV
M
10 3 192.168.10.3
PrtrV
M
20 4 192.168.10.4
HrdmV
M
20 5 192.168.10.24
RDVm 20 1 192.168.10.20
PrxyV
M
30 2 192.168.10.22
ScrV
M
30 3 192.168.10.50
WebVm 40 2 192.168.10.24
MdSVm 40 4 192.168.10.23
ERPVm 40 2 192.168.10.68
NACVm 50 4 192.168.10.70
E-ComV
M
30 2 192.168.10.58
IntrantVm 60 1 192.168.10.78
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 3, No. 1, 2012
200 | P a g e
www.ijacsa.thesai.org
The analytical model and validation of the proposed
models in Fig. 2 and 3 will be clearly shown in the future
work; however this work seeks to use the presented
approaches to enhance the security design of a virtualized data
center network.
REFERENCES
[1] Jinjing. jiang and R.Jain, Analysis of backward congestion notification
(BCN) for ethernet in datacenter applications. IEEE communications
Society INFOCOM 2007 proceedings.
[2] Jalal Frihati, Florica Moldoveanu, Alin Moldoveanu , General
guidelines for the security of a large scale data centre design, U.P.B.
Sci. Bull., Series C, Vol. 71, Issue 3, 2009.
[3] Data centre services, URL,
http://www.sun.com/service/storage/datacenterdatasheet.pdf
[4] Practical Large-Scale Network Design With Variable Costs for Links
and Switches,
URL:http://whitepapers.silicon.com/0,39024759,60304468p,00.htm
[5] Mitchell Ashley LAYERED NETWORK SECURITY 2006: A best-
practices
approach,URL:http://www.stillsecure.com/docs/StillSecure_LayeredSe
curity.pdf.
[6] Juniper networks layered security solution,
URL:http://cn.juniper.net/solutions/literature /white _papers/2005.pdf
[7] Timothy Wood, Improving data center resource Management,
deployment, and availability with virtualization, PHD thesis
June,2009,(Unpublished).
[8] http://www.cisco.com/application/pdf/en/us/guest/netsol/ns107/c649/cc
migration_09186a008073377d.pdf
[9] Data center: infrastructure architecture
SRND,URL:http://www.cisco.com/application/pdf/en/us/guest/netsol/ns
304/c649/cdccont_0900aecd800e4d2e.pdf
[10] Data Center: Securing Server Farms ,
URL:www.cisco.com/application/pdf/en/us/guest/netsol/
ns304/c649/ccmigration_09186a008014edf3.pdf
[11] Data center security topologies: www.cisco.com/application/pdf/en/ us/
guest/netsol/ns376/c649/cdccont_0900aecd800ebd1d.pdf