International Journal of Scientific Research & Engineering Trends

Volume 3, Issue 6, Nov.-2017, ISSN (Online): 2395-566X

An Approach for Trusted Computing Of Load

Balancing In Cloud Environment
Manoj Kumar Selkare*, Vimal Shukla
KNP College of Science & Technology, RGPV, Bhopal, M.P
*Email: selkaremanoj@[Link]

Abstract – Cloud computing is a novel approach in order to use the resource of computing where these resources may be
hardware or software. This facility is delivered as a service in the communication network. This facility known as cloud, which
occurred from the use of a service as a cloud, which is an abstraction for the complex infrastructure system containing
diagrams. Services of cloud computing involve trusted remote user data, and computer software. This paper proposed to an
approach to efficient load balancing in cloud computing.

Keywords – Cloud, Load Balancing, Trusted Computer.

I. INTRODUCTION The key concept of the application of cloud computing is

to reduce the economic burden on computing resources.
Cloud computing can be defined as a type of distributed To develop and maintain these services the cloud service
and parallel system consisting of a compilation of provider can apply the rent on the consumers. Although
virtualized and inter-connected computers that are most of the cloud service are the responsibility of the
dynamically provisioned, and presented as one or more manager of cloud to provide the resources requested by
integrated computing resources based on service-level end users as soon as possible in an organized and efficient
agreements established through negotiation between the manner[1].
service provider and consumers [1]. The term cloud Cloud computing or cloud, is a keyword in order to
computing can also be defined as the ability to use express concepts of computation that have the large
computing resources such as applications, storage and number of resources for computing and they are connected
processing power over the internet. These computing with some different real time system like network or in
resources are hosted and managed by “someone else” i.e. other words the Internet. Cloud computing is a synonym
the cloud service provider. of distributed system in which there is a network having
The cloud computing is pay-per use model. Resources the ability to execute multi program at one computer
can be accessed on-demand by everyone at anytime from simultaneously. [13] The term cloud computing is used to
anywhere via internet. The cloud is measured by a group refer to network-based services seems to be provided by
of resources available in the system, including hardware the actual server hardware, where these services are
such as CPU, memory, storage and software size. In actually accessed by the virtual machine by a software
addition to these, resources from other computing which is executing on one or more actual machines. These
resources are available. These resources are used to virtual machine does not exist physically and but it might
provide end-users and consumers. This service is available be possible can be moved and scaled up or down on the
regarding service plans. fly without affecting the end user, without doubt, a bit like
a cloud. The Cloud Computing logical diagram is shown
in figure 1.
Infrastructure
II. ADVANTAGES OF CLOUD COMPUTING
Computer is now a fundamental part of human life. We
need computers everywhere, whether for work or in
research. As the use of computers in our daily life
increases, so is the computing resources needed. For big
Cloud companies like Microsoft and Google, to mobilize
Computing
resources as and when they need is not a problem but for
small businesses, accessibility of resources becomes a
challenge. With the massive failure of the resources such
Application Platform
as machines, hard disk, software bugs, etc., this can be a
problem for such communities. Cloud Computing offers a
solution to this situation.
Fig.1. Cloud Computing Logical Diagram

© 2017 IJSRET
217
 International Journal of Scientific Research & Engineering Trends
Volume 3, Issue 6, Nov.-2017, ISSN (Online): 2395-566X

The progress of Cloud Computing is huge with respect services. A private cloud is alleviation or a data centre
to business and personal uses. Cloud computing users can network that offers hosted services to a limited number of
utilize the online resources on rent. Among many customers or group of customers [17]. A hybrid cloud is a
advantages or benefits, some are discussed here [3] . cloud computing environment that combines both public
A. Scalability: and private clouds. An organization provides and manages
Scalability is the capability of a system to enhance the the resources of its own house and of others, provided
output under an augmented load when resources are externally, in the hybrid cloud. Ideally, the hybrid cloud
added. Resources can be servers, hardware, storage and approach allows a company to take advantage of the
networking. The user can use the cloud computing scalability, efficiency, flexibility and profitability
resources according to his needs without having to buy
resources quickly. IV. LOAD BALANCING
B. Virtualization:
In cloud computing, virtualization is a conception where In general, load balancing [4] provides the ability to
users have a unique view of available resources, regardless avoid the situation where some resources are overloaded
of their position in the physical device. Therefore, it is system while others remain idle or under load. Load
advantageous to provide users with services with less balancing aims to optimize resource use, minimize
physical resources. response time, maximize throughput and avoid single
C. Mobility: resource from being overloaded. This section provides a
Cloud computing means mobility since the users have summary of the work related to load balancing techniques
full access to the applications at any instance of time by offered. Specifically, the main features are described in a
the use of internet. wide range of approaches to load balancing.
D. Reduced infrastructure costs: Load balancing [4] is the process of reallocation of the
The pay per use model is supported in the cloud. In fact, total expenses of the individual nodes of the collective
it allows an organization to pay for the resources they system, to gain a better response time and also a good use
need, do not make the investment in the resources of resources. Cloud computing is an Internet computing in
available in the cloud. In addition, the supplier requires no which load balancing is one of the difficult tasks. Several
infrastructure maintenance or upgrade costs. Increased methods are used to make a better system that affect the
storage: users or customers of cloud computing can store loads on nodes in a balanced way but due to network
more data in the cloud than in private computer systems, congestion, use of bandwidth, etc, problems had occurred.
they use regularly. It not only relieves the purchase of These problems were solved by some existing techniques.
additional storage space, but also improves the A load balancing algorithm is dynamic in nature it means
performance of your regular system, as it is less busy. On it does not taken into account the previous state or
the other hand, data or programs can be accessed at any behavior of the system i.e., this algorithm depends only on
time via the Internet, as they are available in the cloud. the actual behavior of the system. There are several
objectives related to load balancing to improve
III. CLOUD COMPUTING performance significantly, to maintain system stability,
etc. According to the current state of the system,
Cloud computing [2] is an emerging paradigm of algorithms of load balancing can be classified into two
business computing infrastructure viable information on types, which are static algorithms and dynamic. The static
the Internet that can be accessed from a Web browser by algorithm was necessary prior knowledge of the system
customers to meet their needs. This is a type of computing and does not depend on the current system. In the case of
that is characterized by the availability of resources in a dynamic algorithm, this is based on the current system,
scalable and dynamic way of application, where the and is more efficient than the static algorithm.
resource is used to represent the infrastructure, platforms,
software, and storage. It promises to abolish the need for V. PROPOSED METHODOLOGY
costly computer equipment companies and institutions
alike. Cloud computing is, generally, defined for all The user requests for service which may include, request
matters related to the provisioning of hosting services over for storing the data at cloud data enter, processing of some
the Internet. It is based on the notion of computing useful information and to run application (software) at
resources, networking, storage and shared applications cloud. The request is passed to the cloud controller policy
provided by a third party. There are different types of from the broker through the web browser provide as
clouds. These can be deployed as private, public, and interface. The requested user communicates to the broker
hybrid clouds. [2] In the private cloud, infrastructure is of the service provider through Internet. Broker sends the
operated and managed solely by a particular organization. task (cloudlet) to the cloud controller. The virtual machine
It can be managed by a third party also, on site or off site. manager of cloud controller will create the list of virtual
In other words, a public cloud sells services to anyone on machines according to individual user request
the Internet and its infrastructure will be made available to specification. Each VM has different description a
the public and are owned by an organization selling cloud different task to execute.

© 2017 IJSRET
218
 International Journal of Scientific Research & Engineering Trends
Volume 3, Issue 6, Nov.-2017, ISSN (Online): 2395-566X

The simulation tools which can be used in solving the Let us assume the utilization of each host (in terms of
problems or fulfilling the objectives came across during percentage).
the survey for the dissertation work. They are: List of host utilization [23,65,10,75,50,84,15,30,90,12]
  After sorting [10, 12, 15, 23, 30, 50, 65, 75, 84, 90]
CLOUDSIM Median = (30+50) / 2 = 40
Cloudsim is a new, general, and extensible simulation First half [10, 12, 15, 23, 30], Second half [50, 65, 75, 84,
framework that allows modeling, simulation, and 90]
experimentation of emerging cloud computing After selection of VM from VMmigrationList1 check
infrastructures and application services. In cloud best suitable host from Most Likely Over loaded Host List
computing case, the simulation tools like Cloudsim offers and if its utilization after allocation is greater than
significant benefits to the customers and providers. For previous utilization on the host along with power after
customers, it allows them to test their services in allocation (Power after Allocation) is less than min Power
controllable environment with free of cost and to check then select this host for VM migration.
the performance before publishing to the real clouds. In algorithm 2 we replace host list by Most Likely Over
 
ECLIPSE: loaded Host List which is defined as follows:
Eclipse is a Java-based open source platform that allows The proposed algorithm finds over loaded host first and
a software developer to create a customized development then choose the virtual machine for migration from
environment (IDE) from plug-in constituents fabricated by overloaded hosts and place this VM using VM placement
the Eclipsemembers. Though Eclipse Platform is written in algorithm for Overloaded Host. After load balancing, finds
Java language, it allows the developers to develop and test an under loaded host and select all the virtual machines
code written in other languages too through the plug-in from that host. Place these VMs over least loaded host by
constituents it supports. using VM placement algorithm for Under loaded Host. To
 
ANEKA: evaluate the performance of the proposed scheme, the
Aneka is a Cloud Application Development Platform CloudSim toolkit was used for simulation and forming of
(CADP) for developing and running compute and data cloud computing environments and evaluation of resource
intensive applications. As a platform it provides users with provisioning algorithm [13]. CloudSim is a new
both a runtime environment for executing applications simulation framework for designing, simulation, and
developed using any of the three supported programming testing of emerging Cloud computing infrastructures and
models, and a set of APIs and tools that allow the users to application services. It provides built in java classes to
build new applications or run existing legacy code. simulate datacenter, host machine and various strategies.
User can evaluate the new strategies (policies, scheduling
A. Inter quartile Range (IQR) method for finding algorithms, mapping and load balancing etc.) using these
dynamic threshold all above together. The classes of the libraries can be
The inter quartile range (IQR) is an estimate of replaced or extended, new policies can be defined and new
variability, based on dividing a data set into quartiles. It is scenarios for utilization can be ciphered
the difference between the upper and lower quartile in a
data set. VI. EXPERIMENTAL RESULTS
Steps for finding Inter quartile Range: We have analyzed different scenarios by taking 10 hosts,
1. Sort the data set in increasing order. 20 virtual machines and various numbers of tasks (cloudlets)
2. Find the median for the ordered set (Q2). i.e. load to evaluate the performance of proposed algorithm. It
3. Divide the data set into two halves. is essential to use workload traces from real system. In this
4. Find the mean for the first half of the ordered data set simulation we have taken different work load. We have
(Lower Quartile Q1). plotted different graphs based on different work load between
two strategies, proposed method and the existing method
5. Find the mean for the second half of the ordered data set
(Static threshold based VM provisioning). The performance
(Upper Quartile Q3). of the proposed approach is evaluated for various parameters
6. IQR = Upper Quartile – Lower Quartile. (i.e. Energy Consumption, Number of VM migrations,
Here data set describes set of the host utilization. We Number of SLA violation and Average SLA violation). These
propose a method based on two threshold values, lower results can vary according to the different environment
threshold and upper threshold. The mean for the first half setups.
of the ordered data set (host utilization) is used to
calculate the lower threshold value, while mean of second
half of the ordered data set (host utilization) is used to
calculate the upper threshold value. This is shown in
example as follows:

© 2017 IJSRET
219
 International Journal of Scientific Research & Engineering Trends
Volume 3, Issue 6, Nov.-2017, ISSN (Online): 2395-566X

Energy Used SLA Violation

9 90
8

Systerm
80
7 70
6 60
Load

5 50

Load in
4 40
3 30
2 20
10
1

0 0.1 0.2 0.3

0 1000
2000 3000
Energy
Existing Method Proposed Method Value
Existing Method Proposed Method
Fig.2. Energy Consumption Comparisons between
Proposed and Existing Method Fig.4. SLA Violation Comparisons between Proposed and
Existing Method
Energy consumption is compared for proposed method
and existing method as shown in figure 2 from graph we Number of SLA violation is compared for proposed
have analyzed that proposed method; consumes less method and existing method in figure 4. From graph we
energy than existing method. have analyzed that proposed method; incurs less number
of SLA violations than the existing method.

Virtual Machine VII. CONCLUSION

Migrations In Energy Efficient Dynamic Virtual Machine
90 Consolidation algorithm. Host overloading may cause
increase in virtual machine migration and SLA violation.
80 The proposed method for resource allocation and load
70 balancing is to minimize the number of virtual machine
migrations and SLA violation.
60
Load

50 REFERENCES
40 [1] R. Buyya, S. Venugopal, J. Broberg, and I. Brandic,
Cloud Computing and Emerging IT Platforms: Vision,
30
Hype, and Reality for Delivering Computing as the 5th
20 Utility. Future Generation Computer Systems, 25(6): 599-
616, Elsevier Science, June 2009.
10 [2] Rodrigo N Calheiros, Rajiv Ranjan, Anton Beloglazov,
Cesar AF De Rose, and Raj Kumar Buyya. CloudSim: A
0 20 40 60 toolkit for modelling and simulation of cloud computing
environments and evaluation of resource An Efficient
Value VM Scheduling Technique in Cloud Computing
Existing Method Proposed Method Environment Page 59 provisioning algorithms. Software:
Practice and Experience,41(1):23 50, 2011.
Fig.3. Number of VM Migration comparisons between [3] Eeraj Jan Qaisar, “Introduction to Cloud Computing for
Developers Key concepts, the players and their
Proposed and Existing method offerings,” Information Technology Professional
Conference, IEEE©2012.
Number of VM migrations is compared for proposed [4] [Link]
method and existing method in figure 3. From graph we [5] Komal Singh Patel and A. K. Sarje, “VM Provisioning
have analyzed that proposed method; incurs less number Method to Improve the Profit and SLA Violation of
of VM migrations than existing method. CloudServiceProviders”,IEEEInternational

© 2017 IJSRET
220
 International Journal of Scientific Research & Engineering Trends
Volume 3, Issue 6, Nov.-2017, ISSN (Online): 2395-566X

Conference, Cloud Computing in Emerging Markets

(CCEM) 11-12 Oct. 2012.
[6] YatendraSahu, R.K. Pateriya, and Rajeev Kumar Gupta,
(2013) “Cloud Server Optimization with Load Balancing
and Green Computing Techniques Using Dynamic
Compare and Balance Algorithm,”5th International
Conference on Computational Intelligence and
Communication Networks© IEEE.
[7] R. Buyya, A. Beloglazov, and J. Abawajy, “Energy-
Efficient Management of Data Center Resources for
Cloud Computing: A Vision, Architectural Elements, and
Open Challenges”, in proceedings of the 2010
International Conference on Parallel an d Distributed
Processing Techniques and Application s (PDPTA 2010),
Las Vegas, USA, July 12-15, 2010.
[8] Beloglazov and R. Buyya, “Adaptive Threshold-Based
Approach for Energy-Efficient Consolidation of Virtual
Machines in Cloud Data Centers”, in Proc. 8th
International Workshop on Middleware for Grids, Clouds
and e-Science, Ney York : ACM, 2010.
[9] Jing Huang, Kai Wu, and Melody Moh, “Dynamic virtual
migration algorithms using enhanced energy consumption
model for green cloud data centers,” IEEE ©2014.
[10] Mohammadalaulhaquemonil, Romasaqasim,and
Rashedur M rahman, “Energy-aware VM consolidation
approach using combination of heuristics and migration
control”, IEEE ©2013.
[11] K. Yang, J. Gu,T. Zhao and G. Sun ,“ An Optimized
Control Strategy forLoad Balancing based on Live
Migration of Virtual Machine ”, in Proc.6th Annual
China grid Conference ( China Grid), Liaoning : IEEE,
2011.
[12] P. GtzijabaLeelipushpam and [Link], “Live VM
migration techniques in cloud environment –A Survey,”
proceedings of 2013 IEEE Conference on Information
and Communication Technologies (ICT 2013).
[13] Khiyaita, A., et al. "Load balancing cloud computing:
state of art." Network Security and Systems (JNS2), 2012
National Days of. IEEE, 2012.
[14] Anton Beloglazov, jemalAbawajy, and RajkumarBuyya,
“Energy aware resource allocation heuristics or efficient
management of data center for cloud computing,” future
generation computer systems (FGCS), volume 28, Issue
5, Elsevier science, the Netherlands 2012.
[15] R.N. calheiros, [Link], [Link], and [Link],
(2011) “CloudSim a toolkit for modelling and simulation
of cloud computing environments and evaluation of
resource provisioning algorithms,” Software-practice and
Experience.
[16] Cloud computing demystifying SaaS, PaaS and IaaS.
[Link] [Link]/2010/05 /cloud computing-
demystifying-saaspaas-and-iaas/, 2010.
[17] Manishaben Jaiswal, "Cloud Computing And
Infrastructure", International Journal of Research and
Analytical Reviews , E-ISSN 2348-1269, P- ISSN 2349-
5138, Volume.4, Issue 2, Page No pp.742-746, June
2017.

© 2017 IJSRET
221