A HYBRID OUTLIER DETECTION APPROACH WITH MULTI DIMENSIONAL FEATURES TO PREVENT BLACK HOLE ATTACK IN MANET

Abstract

An increase in the technological upgrades in the mobile communication technology has led to the tremendous increase in the usage of mobile networks. Mobile Adhoc Network is an instantaneous network where the requirement of the infrastructure is not mandatory. This feature has raised a lot of issues in the security aspects of the MANET. The mobility of the nodes, frequent topological changes, weak communication link and lack of infrastructure are the critical factors creating security issues in MANET. The security issues leads to lot of routing attacks, which disturbs the network communication and completely collapse the system. One such attack is the black hole attack. The AODV routing protocol is more vulnerable to the black hole attack. This paper proposes an efficient approach based on outlier detection techniques to detect and prevent the black hole attack in MANET. The results have been simulated using the NS2 simulator and performance of the algorithm has been evaluated based on various performance metrics like packet delivery ratio, routing overhead, detection efficiency ratio and end to end delay time .