Improving the reliability of ad-hoc on demand distance vector protocol

Ad-hoc networks are a collection of mobile nodes that can be deployed without the need for any centralized management infrastructure. As such, the very basic operation of these networks is dependent on the cooperation of their nodes to provide communication routes. Furthermore, such networks are prone to several security risks. Consequently, when some nodes behave maliciously or in a selfish manner, the operation of the whole network can be severely affected. The behavior of such nodes can result in degradation of the performance of the network or even disruption of its operation altogether. This work reports the results of studying the Ad-hoc network performance in the presence of malicious nodes. It also proposes an approach that takes the behaviour history of member nodes into account, to improve network reliability. The reported approach, namely Behavior Ad-hoc On Demand Distance Vector approach, is based on discovery of communication paths with minimal number of selfish or malicious nodes. Using OPNET simulator, several situations of practical significance have been studied and analyzed. The results of these studies show that by applying the proposed approach, marked improvements in the performance and reliability of the Ad-hoc networks depend on AODV protocol can be achieved.