Design of genetic algorithm based IDS for MANET

Abstract

A MANET (Mobile Adhoc Network) is an infrastructure-less self configuring wireless networks of routers. It has potential applications in totally unpredictable and dynamic environment. Routing protocol used here is a form of reactive routing called AODV. AODV (Adhoc On Demand Distance Vector) routes based on demand. The major benefit of AODV is minimum connection setup delay and assignment of sequence numbers to destination to identify the latest route. The route updates are done by periodic beaconing. This network is susceptible to various hazards. In specification based Intrusion Detection system, certain characteristics of vital objects are analyzed and any abnormality is detected. We propose a technique to analyze the exposure to attacks in AODV, specifically the most common network layer hazard, Black Hole attack and to develop a specification based Intrusion Detection System (IDS) using Genetic Algorithm approach. The proposed system is based on Genetic Algorithm, which analyzes the behaviors of every node and provides details about the attack. Genetic Algorithm Control (GAC) is a set of various rules based on the vital features of AODV such as Request Forwarding Rate, Reply Receive Rate and so on. The performance of MANET is analyzed based on GAC.