25-06-2013, 12:58 PM
Evaluation of Impact of Wormhole Attack on AODV
Evaluation of Impact.pdf (Size: 96.25 KB / Downloads: 46)
ABSTRACT
Mobile Adhoc Networks (MANET) are self organizing, decentralized networks and possess dynamic topology, which
make them attractive for routing attacks. Wormhole attack is a network layer attack observed in MANET, which
completely disrupts the communication channel. This paper focuses on study of wormhole attack , its behavior and
the performance impact of wormhole attack on Adhoc On Demand Distance Vector (AODV) routing protocol. The
NS2 network simulator is used to evaluate the wormhole attack impact on AODV.
INTRODUCTION
A Mobile Adhoc network (MANET) [1] is composed of a
collection of independent mobile hosts connected by
wireless links without any fixed or centralized
administration. MANET is characterized by its dynamic
topology, multi hop routing, energy limited operations and
network scalability. Malicious nodes carry out both active
and passive attacks [2] due to the open and adhoc nature of
MANET. Basic routing protocols [3] used in MANET,
table-driven/proactive, demand-driven /reactive or hybrid
variants, have not met the security requirements such as
confidentiality, availability, integrity, authentication and
non repudiation.
WORMHOLE ATTACK IN AODV ROUTING PROTOCOL
AODV [5] is an on demand routing protocol, which creates
the routes on demand. During the route discovery phase, the
route request message, RREQ are broadcasted to its
immediate neighbors. This process is repeated till the
RREQ message reaches the destination. Upon receiving the
first RREQ at destination, reply message RREP is sent back
by destination to the source following the reverse path. All
intermediate nodes set up forward route entries in their
table. Route error message are forwarded upon detecting an
error in link. Periodic hello messages check for the
neighbor node link connectivity.
SIMULATION STUDY
Scope of study
In this work, wormhole attack is simulated in ns2 [14] by
using encapsulation of packet approach in AODV routing
protocol. At one end of the wormhole tunnel, the packets
are encapsulated and at the other ending end of tunnel,
packets are decapsulated. Here, wormhole peers are far
apart but this tunnel creates an illusion that wormhole peers
are one hop count apart as shown in Fig. 1. However the
latency of the wormhole link is very high. Once wormhole
tunnel is created, wormhole peer nodes would drop the
packets.
Network Throughput
Network throughput is measured as the total number of
packets received at the destination over a period of time
and is expressed in kbps. In the first scenario, 50 node
MANET is considered with a wormhole link (two
wormhole peers) simulated and number of network
connections is increased from 0 to 5. The throughput
comparison for this first scenario is depicted in Fig 2.As
shown in Fig 2, the AODV throughput decreases when the
wormhole link is present compared to normal AODV
throughput. Maximum throughput difference observed
between normal AODV and wormhole infected AODV is
around 40kbps. In the Table 2, the throughput observed in
normal AODV and wormhole infected AODV is
represented. Also the percentage of decrease in AODV
throughput due to the presence of one wormhole link is
depicted. The simulated second scenario involves
increasing the wormhole links from 1 to 5; total 10
wormhole malicious nodes are present. Again the network
throughput is calculated in this scenario and depicted in
Fig3.
CONCLUSION AND FUTURE WORK
In this paper, the study of wormhole attack launched in
AODV routing protocol in MANET is conducted and the
simulation study depicts the performance degradation in
terms of parameters like network throughput, average end
to end delay, packet delivery ratio, drop rate.
In future, a novel multi layer approach to detect
wormhole attack in MANET would be proposed and the
simulation results for same would be captured to show the
effectiveness of the proposed detection mechanism.