International Journal of Emerging Technology and Advanced Engineering

Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)

An Efficient Wormhole Prevention in MANET Through Digital

Signature
Anil Kumar Fatehpuria1, Sandeep Raghuwanshi2
1 2
M.Tech Scholar, Assistant Professor, Samrat Ashok Technological Institute,Vidisha (M.P.), India
Abstract— Wireless networks are gaining popularity to its On the other hand, MANET utilizes an untrusted
peak nowadays, the desire to be connected anytime and environment for data transmission, and therefore it is
anywhere has led to the development of wireless network, subjected to various kinds of security attacks [2].
opening new vista of research in pervasive and ubiquitous
computing. Mobile Ad hoc Network (MANET) has distributed
mobile wireless nodes, which do not have pre-determine
topology and pre-existing infrastructure mobile nodes that
can arbitrarily change their geographic locations and random
mobility with constrained resources, ad hoc networks are
vulnerable due to their structure less property. During the
wormhole attack, a malicious node captures packets from one
location in the network, and “tunnels” them to another
malicious node at a distant point, which replays them locally.
We identify two types of wormhole attacks. In the first type,
malicious nodes do not take part in finding routes, meaning
that, legitimate nodes do not know their existence. In the
second type, malicious nodes do create route advertisements
and legitimate nodes are aware of the existence of malicious
nodes, just do not know they are malicious. Some researchers
have proposed detection mechanisms for the first type. In this
paper we represent a mechanism which is helpful for
prevention of wormhole attack, through observing the delay
of different path to receiver and verification of digital Figure1. Mobile Ad Hoc Network
signature. Our mechanisms detect pinpoint location of
wormhole and prevent them. This method requires neither Mobile ad hoc networks have following features [3]:
synchronized clocks nor special hardware equipped mobile  The wireless link between the nodes is highly
nodes. vulnerable. This is because nodes can continuously
move causing the frequent breakage of the link. The
Keywords- MANET, Wormhole Attac Detection &
Prevention, Digital Signature , Multipath Hop Count power available for transmission is also strictly
Analysis. limited.
 The topology of the network is highly dynamic due to
I. INTRODUCTION the continuous breakage and establishment of wireless
link. Nodes continuously move into and out of the
Mobile Ad Hoc Networks are formed by a set of hosts
radio range. This gives rise to the change in routing
that communicate with each other over a wireless channel.
information.
Each node has the ability to communicate directly with
 These is a bandwidth constraint in this wireless
another node (or several of them) in its physical
networks.
neighbourhood. Such mobile Ad Hoc Networks have many
 MANETS need energy - efficient operation because
attractive features including automatic self-configuration
and self-maintenance, quick and inexpensive deployment, all the nodes depend on battery power which is highly
and the lack of the need for fixed network infrastructures or limited.
centralized administration [1],[2].These features lead to The open nature of the wireless medium makes it easy
important applications that can not be Performed by for outsiders to listen to network traffic or interfere with it.
traditional wired networks. The importance of mobile Ad These factors make ad hoc network potentially vulnerable
Hoc Networks is increasing rapidly with advances in to several different types of malicious attacks. Like
technology that result in smaller, cheaper, and power- Wormhole Attack, Block Hole Attack, Rushing Attack etc
efficient devices . [1] and [5].
360
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
Mobile ad hoc Network are becoming more and more Now, let us discuss that how wormhole attack can be
popular because of their important applications ranging launched in wireless network: it can be launched by four
from health care and logistics, through agriculture, forestry, ways [6].
civil and construction engineering, to surveillance and 1) Packet Encapsulation: In which one malicious node
military applications [2]. encapsulates the route request and sends it to colliding node
The remainder of the paper is organized as follows. We which decapsulate it and forwards the route request
first present wormhole attack mode, classes and types. In (RREQ) packet.
section II, Section III describes the relate work. Section IV
2) Out-of-Band: In Out-of-Band, two malicious nodes
explain proposed schema. Section V describes simulation
sends route request (RREQ) between them by using the
and result. Section VI describes the conclusion and future
long range directional wireless link or direct wired link.
work.
3) High-Power-Transmission: In high power
II. THE WORMHOLE ATTACK transmission a malicious node get a route request (RREQ)
and broadcast that request with high power level. Any other
In this section we explain the wormhole attacks modes node that hears the high power broadcast must be a
and classes while pointing to the impact of the wormhole malicious node so it receives that route request and again
attack and the efforts that have been done in the literature rebroadcast towards the destination.
to detect and prevent this attack.
During the attack a malicious node captures packets 4) Packet Relay: In packet relay two malicious nodes
from one location in the network, and tunnels them to relay packet between two nodes which are far apart from
another malicious node at a distant point, which replays each other and convenIence these nodes that they are
them locally. neighbor.
A. Wormhole Attack in Routing Protocol.
Many routing protocols are available for MANET. In
this section, some of the frequently used routing protocol
are reviewed and the threat of wormhole attack to such
protocols is considered. Theses routing protocols can be
categorized into two types: table- driven/proactive and on
demand/reactive [6], [7].
Table driven routing protocol:
In these protocols all nodes consists up-to date routing
information, so they update their table with in a fixed time
period. The OLSR is the example of these proactive routing
protocols.
On-Demand routing protocol:
Figure 2 wormhole attacks
In these routing protocols, the route is created only if
Figure 2 shows the scenario of the wormhole attack. required. Nodes do not update their information within a
There are two malicious nodes which are far apart from fixed time interval. These protocols find route on demand
each other in a similar network or may be in different by flooding the network with route request packet from
networks which are connected with each other through a source to destination. AODV and DSR are the example of
tunnel and pass the data packets through tunnel where they reactive routing protocols.
are replayed [4]. The tunnel is either the wired link or a
B. Wormhole Attack Classification:
high frequency links.
The wormhole attack can actually be helpful if used for In a wormhole attack, two attackers work together. One
forwarding all packets. It can be launched without having receives the packets, tunnels the packets to its partner, and
the cryptographic keys. Some harmful effects of wormhole then the partner replays them into the network. There are
are as follow: selectively drop data packets, routing two kinds of wormhole attacks. In the first type, malicious
disruption in which attackers prevent discovery of nodes hide the fact that they forward a packet, meaning
legitimate route and traffic analysis for information leaking that, legitimate nodes do not know their participation in
[5]. packet Forwarding.

361
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
In the second type, legitimate nodes are aware of the fact Then the RREQ packet is tunnelled to M2 and M2
that the malicious nodes are forwarding packets, just do not performs the same setup procedure and broadcasts the
know they are malicious. For the case of discussion, we RREQ packet to R. R finds its previous hop is M2 with hop
refer the first type as hidden attack while the second types count equals to 3. The same thing happens in the reverse
as exposed attack [7] and [8]. path. When S receives the RREP packet, it finds its
Hidden attack: The attackers do not modify the content of pervious hop is M1 with hop count equals to 3. And the
the packet and the packet header, even the packet is an route is setup as {S, Ml, M2, R}.
AODV advertisement packet. Instead, they simply tunnel C. Metric to Detect Wormhole:
the packet from one point and reply it at another point. This
There are different metrics to measure the strength of
kind of wormhole attacks makes the sender treat the
wormhole present in the network. In previous considered
receiver as its immediate neighbour. Suppose that S wants
several metrics for measuring the capability of the nodes
to establish a route to R using AODV.S would broadcast a
involved in wormhole attack. These include strength,
RREQ message. Any node that receives the RREQ should
length, attraction and robustness [6]. These are defined
check whether it knows how to get to R. if not ,it should
below.
continue to broadcast RREQ if it receives RREQ for the
first time. It should also update the hope count information 1) Strength: It is the amount of traffic attracted by the false
and put its identity in the packet header. However, in the link advertised by the colluding nodes.
hidden attack, malicious node do not update the packet 2) Length: Larger the difference between the actual path
header as it should. As shown in fig. ( ) the packet from S and the advertised path, more anomalies can be observed in
is received by M1, then M1 tunnels the packets to M2 and the network.
replies them to R, without modifying the packet header. 3) Attraction: This metric refers to the decrease in the path
Since M1 and M2 do not include themselves in the header, length offered by the wormhole. If the attraction is small
what R will find is that the pervious hop is S. the same then the small improvements in normal path may reduce its
observation can be obtained in the reverse path, such that S strength.
finds R as its immediate neighbor, and the path found is 4) Robustness: The robustness of a wormhole refers to the
{S,R}.This is obviously not correct since S and R are ability of the wormhole to persist without significant
Separated by M1,M2 and other node that are in the tunnel. decrease in the strength even in the presence of minor
topology changes in the network.
Besides these, the packet delivery ratio which is the
number of packet of delivered divided by the total number
of packets dispatched forms a basic metric to quantify the
impact of intrusion.
III. RELATED WORK
The problem of security and cooperation enforcement
has received considerable attention by researchers in the ad
hoc network community. In this section, some of these
contributions are presented.
Packet Leashes is the excellent mechanism for detection
of wormhole attack [9]. There are two types of packet
Figure 3(a) Hidden Attack and (b) Exposed Attack
leashes: One is temporal leashes which are related to time
of sending and receiving packets from one node to another
Expose attack: In this kind of attacks, the attackers do not node. Another one is geographical leashes which are
modify the content Of the packet, but include themselves in related to geographical location of nodes.
the packet header following the route setup procedure. a) Geographical Leashes: It requires location determination
Other nodes are aware that the malicious nodes lie on the (eg. GPS hardware) and all nodes contains a clock which is
path but they would think that the malicious nodes are loosely synchronized. In geographical leashes when one
direct neighbours. Let's consider the situation where S node sends a packet to another node then it includes its own
wants to establish a route to R. As illustrated in Fig. l(b), location ps and time on which it sends a packet ts. The
when M1 receives the packet, it modifies the pervious hop receiving node compare the value of sending packet with
field to M1 and increases the hop count by 1. its own location pr and time at which it receives packet .
362
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
b) Temporal Leashes: In temporal leashes all nodes must All sensors send this distance information to the central
required a tightly synchronized clock. The time controller, which calculates the network’s physical
synchronization can be achieved now with off-the-shelf topology based on individual sensor distance
hardware based on LORAN-C [10], WWVS [11], and GPS measurements.
[12] and [13]. Lazos et al [8] have proposed a graph-theoretical
But the problem faced in packet leashes is that it is approach to wormhole attack prevention based on the use
inaccurate due to unpredictable processing time and of Location-Aware Guard Nodes (LAGNs).
channel availability. It does not prevent DoS attack against
route establishment. IV. PROPOSED SOLUTION
Lite Worp [14] also provides a defense mechanism In Our mechanism is inspire form DelPHI wormhole
against wormhole attack; it uses secure two hop neighbor detection method, the advantages of our mechanism is that
discovery and local monitoring of traffic by using guard it does not require clock synchronization and position
node. I t also has additional features that provide a information ,and it does not require the mobile node to be
technique to isolate the malicious node from the network. equipped with some special hardware. In this method we
But there is some restriction in liteworp that it requires are providing solution for both kinds of wormhole attack,
extremely accurate clock, assuming no delay in network Hidden attack or Exposed attack. There are two phases in
apart from propagation delay and exact measurement of our mechanism.
angle of reception. In the first phase delay/hop count and verification of
Womeros [15] is the framework for defending against digital signature information is collected. And in the second
wormhole attack which contains two phase: one is phase analyzes the collected information obtained in first
suspicious and another is conformation. The first phase phase to detect whether there is any wormhole attack
applies inexpensive techniques and utilizes local present or not. The reason behind is that under normal
information that is available during the normal operation of situation ,the delay a packet experiences in propagating one
wireless nodes. Advance techniques in the second phase are hop should be similar along each hop along the path.
adapted only when wormhole attack is suspected. If there i However, under a wormhole attack the delay may
s n o presence o f malicious node in the network after unreasonably high or low, since there are in fact many or
applying suspicious phase technique then there is no need no hops between them. Therefore, if we compare the delay
to waste consumption and communication resources by per hop of a legitimate path with the delay per hop of a path
applying conformation technique. that is under wormhole attack, we should find that the delay
Directional antennas [17] are used to prevent wormhole of the legitimate path is smaller. Therefore, if path has
attack. Each node in the network shares a secret key with distinguishable high or low delay value, it is likely to be
every other node and broadcasts HELLO messages to subjected to a wormhole and another technique we using
discover its neighbor using directional antenna in each for pin point detection of wormhole digital signature
direction. technique, it is assumed that each legitimate node shares
A Round Trip Time (RTT) mechanism [16] is proposed the digital signature of every node in the network and the
by Jane Zhen and Sampalli. The RTT is the time that malicious node does not have its own digital signature. Our
extends from the Route Request (RREQ) message sending mechanism is designed specifically for AODV routing
time of a node A to Route Reply (RREP) message protocol in mobile ad hoc network.
receiving time from a node B. A will calculate the RTT If a sender want to send the data to the destination, then
between A and all its neighbors. Because the RTT between it firstly broadcast the route request (RREQ) packet in the
two fake neighbors is higher than between two real network. The route request (RREQ) packet header contains
neighbors, node A can identify both the fake and real the information of visiting node (node-id) in the
neighbors. In this mechanism, each node calculates the information column and hop count column which contains
RTT between itself and all its neighbors. This mechanism the number of visiting nodes used in path and digital
does not require any special hardware and it is easy to signature column. When the sender broadcast a route
implement; however it can not detect exposed attacks request then it adds its node-id in node information column
because fake neighbors are created in exposed attacks. and starts its hop counter with one and every nodes of the
Wang et al [18] have proposed an approach in which network share their digital signature with each nodes. All
each sensor estimates the distance to its neighbors using the the intermediate node add its node-id and increment the
received signal strength. number of hop count by one until it reached at destination.

363
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
The destination node used a scheme called delay/hop In the second phase we analysis the data and detection
count in which destination node received all route request the delay/hop. used the Round Trip Time (RTT)
which reached at destination following different path. To mechanism.
enhance reliability of the information collected, the data Suppose that the sender initiates the detection, i.e.
collection procedure is repeated three times. It is possible broadcasts the RREQ packet, at time ts, and receives a
that the hope counts of three RREQ packets from the same RREP packet from a neighbor node i at time ti, then the
neighbour are different. Now the destination node analysis round trip time (RTT) of the path through node i is given
the number of hops used by different path and select the by RTTi, = ti,− ts. If the hop count field in the RREP from
route for unicast route reply packet (RREP). node i is hi, then the delay/hop value (DPH) of the path to
In this case, we select the delay/hop of the shortest path the receiver through node i is given by
for analysis. It is because a path that is under wormhole
DPHi = RTTi = ti-ts
attack tends to be shorter. Among all the shortest path
2hi 2hi
RREPs, we take the average of the delays for wormhole
detection. Because the minimum or maximum number of In normal situation, a smaller h provides a smaller value
hop count may contains malicious node due to their of RTT. It can be explained by the fact that a shorter path
encapsulation and tunneling property. should have a smaller round trip time. Hence the DPHs of
To check the authentication of selected path. Destination normal paths should have similar values independent to h.
node unicast the (RREP) packet. Whose header contains However, it is not the case in the paths which suffer from
the column of node-id, which contains the id of the all wormhole attacks. Recall that a tunnel is formed by two
nodes used in that path and the digital signature column in malicious nodes. No matter how long the tunnel is, the
which each visiting node add its digital signature. When the malicious pair M1 and M2 advertise to others that they are
receiving node received route reply (RREP) packet. It 1 hop away. Therefore, the longer the tunnel, the larger the
compare the digital signature of previous node, which are RTT, but the hop count remains the same. The resulted
in the signature column of RREP header, from its database DPH value will be larger than normal path.
which contains the signature of all nodes in the network. If We performed some simulations and observed that the
sending node is legitimate then the digital signature of DPH values of normal paths usually appear as small values
sending node should be identical to the digital signature when compared with those of tunneled paths.
which are in database of receiving node, digital signature of
two nodes in signature column of packet header should not V. SIMULATION AND RESULT
be identical and there is blank space in place of signature in
In this section, we evaluate the performance of our
signature column of packet header. If all condition is
proposed solution by OPNET modeler 14.0[]. In our
satisfy then sending node is a legitimate node so receiving experiments, the ad-hoc network includes 25 mobile nodes
node also add its signature in signature column of header placed randomly in square field campus of 100 square
and again unicast route reply (RREP) packet to next node.
kilometer area. For different scenarios of simulation,
This process is repeated again and again till source node.
Constant position mobility and random walk 2D mobility
When the RREP packet reached at source node, source
model are used.
node also verifies the signature of previous nodes, if the We have various simulation parameters along with their
route reply reached at source is legal then source node values are listed in table1.
creates a secure and authenticated path between source and
destination. If there is presence of any malicious node in Parameters Values
the path which received route reply (RREP) packet and Routing protocol AODV
unicast it to next node, the node that received packet from Simulation time 1000sec
malicious node found that signature column of packet Simulation area 10*10 kilometer
header either contains duplicate digital signature of any Numbers of mobile nodes 15
previous node or a blank space in place of digital signature Data packet size 1024Bytes
because the malicious node does not have their own digital Data rate 11Kbps
signature. So the node discard the reply and inform to all Speed of node 10Km/h
node in the network about the malicious node and all other Number of malicious nodes 2
nodes update their database.

364
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
According to figure the delay in second when there are
not present any wormhole in the network the value of the
graph lying between .8 to 1.6 and when the attack present
in the network the value is goes down .4 to .6 then apply
proposed schema the result is more efficient and the value
is between .1 to .6.
f

Figure 4 Average Route Discovery Time

In the above result graph showing result in three

condition of network one when there is not any attack in
network another is when attack present in network and last
when apply proposed schema figure 1 is graph between
average route discovery time for AODV protocol. We can
see that there is much better performance of network for Figure 6 Average Delays per sec
the route discovery time.

Figure 5 average number of hopes per route Figure 7 Average Traffic Sent packet per second

Based on figure we are showing average number of hop According to the simulation result of MANET average
per route is here when attack came in the network then the traffic sent packet per second is near about to the normal
hopes per route is goes down and in the secure hops per situation of the network when applying proposed schema.
route have good result which is near to normal network.

365
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
ACKNOWLEDGMENT
I am thankful to Samrat Ashok Technological Institute
Vidisha (M.P.) India for providing facilities and resources
for this paper.
REFERENCES
[1] Tran Van Phuong, Ngo Trong Canh, Young-Koo Lee, Sungyoung
Lee and Heejo Lee, “Transmission Time-based Mechanism to Detect
Wormhole Attacks” IEEE Asia-Pacific Service Computing
Conference 2007.
[2] Subir Kumar Sakar, T. G. Basavaraju, C.Puttamadappa “Ad HOC
Mobile Wireless Network Principals, Protocols, And Applications”
Auerbach publications.
[3] H.yang, H. Luo, F. Ye, S. Lu and L. Zhang, “ Security In Mobile Ad
Hoc networks: Challenges and Solutions. IEEE Wireless
Communications, 11(1):38-47, Feb 2004.
Figure 8 Average Traffic Received packets per second. [4] Pallavi Sharma, Aditya Trivedi, “An Approach to Defend against
According to the simulation result in figure the graph Wormhole Attacks in Ad Hoc Network Using Digital Signature.
IEEE 2011.
between the packed received per second is about to near to
[5] Radhika Saini, Manju Khari, “Defining Malicious BehavIor of Node
the normal condition of network, we can clearly show that and its Defensive Methods in Ad hoc network” International jurnal
the packet received is high when wormhole present in the of Computer Applications(0975-8887) Volume 20- No.4, April
network. 2011.
[6] Reshmi Maulik and Nabendu Chaki,”A Study on Wormhole Attacks
in MANET” International journal of Computer Information Systems
VI. CONCLUSION AND FULURE WORK
and Industrial Management Applications ISSN 2150-7988 Volume 3
In this paper, we present an efficient mechanism which (2011) pp. 271-279.
is helpful for detection and prevention of wormhole attack, [7] Marianne Azer, Sherif EI-Kassas and Msgdy EI-Soudani “ A
through Delay Per Hop Indication (DelPHI) by observing [8] Full Image of Wormhole Attacks Towards Introducing Complex
the delay per hop of different path to receiver and Wormhole Attacks in Wireless Ad Hoc Networks” international
journal of Computer Science and Information Security, Vol.1, No.1,
verification of digital signature. Our mechanism is able to May 2009.
detect both hidden and exposed wormhole with pinpoint [9] Hon Sun Chiu and King-Shan Lui. “DelPHI: Wormhole Detection
location of wormhole and prevent them. Mechanism For Ad Hoc Wireless Networks” IEEE 2006.
The performance of our mechanism is also evaluated by [10] Y.C Hu, A. Perrig and D. Johnson, “Wormhole Attack in Wireless
simulation using OPNET modeler. Simulating result network” IEEE JSAC, Vol. 24, No.2, Feb, 2006.
showing better performance of network for secure data [11] David L. Mills, “ A Computer Controlled LORAN-C Receiver for
transfer with accurate detection of wormhole attack Precision Time Keeping” Technical report 92-3-1. Department of
Electrical and Computer Engineering, University of Delaware, DE,
Table 1 March 1992.
Evaluation of some related work [12] David L. Mills, “ A Precision Radio Clock for WWV Transmission”
Exposed Hidden Pinpoint No special Technical Report 97-8-1. Department of Electrical and Computer
Attacks Attacks Location HW required Engineering, University of Delaware, DE, August 1992.
Packet Leashes No Yes Yes No [13] Tom Clark,”Tom Clark’s Totally Accurte Clock FTP Site. Greeblelt,
SAM Yes Yes Yes Maryland”. ftp://aleph.gsfc.nasa.gov/GPS/totally ,accurate.clock/.
EWPTDS Yes Yes Yes Yes [14] Defense Advanced Research Projects Aency
NNT No Yes No Yes www.mil/ato/solicit/bao01-01 faqv4.doc, October(2000).
WHOP Yes Yes Yes Yes [15] I.Khalil, S. Bagchi, N.B.Shroff,”Lite Worp: Detection and Isolation
of the Wormhole in Static Multihop Wireless Network.journal.”
In table we make a comparison between efficient Acm: The International Journal of Computer and
wormhole prevention in MANET Through digital Telecommunication Network Archive, Vol 51,Issue 13,September
signature and others in term of ability to detect exposed 2007.
attacks, hidden attacks to pinpoint wormhole location and [16] H.Vu, A, Kulkarni, N. Mittal, “WOMEROS: A New Framework for
Defending Against Wormhole Attacks on Wireless Ad hoc
avoid using special hardware. Some future work also needs Newtwork.” In WASA 2008,LNCS 5258,pp. 491-502, 2008.
to be done to extend our mechanism to work in complex [17] Zhen J, and Srinivas S. (2003) “ Preventing replay attacks for secure
wormhole attacks. routing in ad hoc networks”, In ADHOC-NOW, LNCS 2865,
pp.140-150.

366
 International Journal of Emerging Technology and Advanced Engineering
Website: www.ijetae.com (ISSN 2250-2459, ISO 9001:2008 Certified Journal, Volume 3, Issue 3, March 2013)
[18] L.Hu and D. Evans, “ Using Directional Antennes to Prevent [20] Ranjan Kaparti,Dan Likarish,”OPNET IT GURU: A Tool for
Wormhole Attacks” in Proc. Network and Distributed System Networking Education”, MSCIT Pr Racticum Paper REGIS
Symposium (NDSS), San, Diego, USA, Feb, 2004. University.
[19] Wang W., Bhagrava B. (2004), “ Visualization of Wormholes in [21] Xinjie Chang, “Network Simulations With OPNET”
Sensor Networks”, Proceedings of the 2004 ACM workshop on PROCEEDINGS Of the 1999 Winter Cinference.P.A. Farrington,
Wireless Security,pp.51-60. H.B. Nembhard,D.T. Sturrock, and G.W. Evans, eds., 1999, pages
307-314.

367