Agent Based Intrusion Detection System for Manet

Abstract

Mobile Ad hoc Network (MANET) is an autonomous, wireless network of mobile devices with dynamic infrastructure. Despite its wide-ranging applications, MANETs are prone to various network attacks due to lack of centralized authority.For intrusion detection, agent-based detection develops a method for selecting a node in the network as a monitoring node. These helps to reduce the energy consumption of sensor nodes as the load of information collection is shifted to the mobile agents. newlineFor intrusion detection in MANET, one approach proposed is to take an extra one hop in the packet forwarding process when the malicious node launch a cooperative attack in the network and an auditor method is combined with the one hop method to detect the single malicious attacks. This can eventually lead to the depleted number of malicious nodes present in the network when comparing with the existing protocols. This method therefore focuses on increasing the delivery ratio. newlineDistributed agent based selective forwarding attack detection and attack response technique is one of the agent-based IDS approach proposed. For detecting the selective forwarding attacks, the detection agents perform downstream monitoring and end-to-end monitoring. It uses a two-hop acknowledgment technique for monitoring and detecting. The collector agents residing at the source initiate the attack determination process by collecting all the attack information at every involved intermediate node. The detection agent residing at the intermediate node works based on the information received from the collector agents and finds out any possibility of attack. newlineAnother agent-based Intrusion Detection System (IDS) proposed is to design a distributed agent-based detection system for network layer attacks using particles warm optimization and neural network in MANET. For eand#64259;cient detection of the intrusion, the network is divided into small zones. In each zone one detection agent is present to manage and detect the irregularity in node activities.Detection agent is fou