Mobile agent intrusion detection system for Mobile Ad Hoc Networks: A non-overlapping zone approach

Abstract

A Mobile Ad Hoc Network (MANET) is a group of wireless nods that can be dynamically organized as a multi-hop packet radio network without using any existing infrastructure or centralized administration. However, MANET is venerable to attacks due to its mobile and ad hoc natures. The security issue is becoming a main concern in the applications of MANET. We consider the problem of decreasing the number of false positives generated by cooperative Intrusion Detection Systems (IDS) in Mobile Ad hoc Networks. In order to overcome the shortcomings of existing intrusion detection techniques, a hierarchical distributed model of multi-level detection intrusion is proposed. The model utilizes the intelligent, lightweight mobile agent as alternative to the client server distribution model and nonoverlapping zone framework. So it can effectively detect complicated attacks. The data collection, data analysis, alert message and alarm message can be achieved by using mobile agents, which may reduce the data transmission to save the bandwidth resource in the MANET. The model can adjust itself dynamically to adapt to the change of the external environmental. Also the model is robust and scalable.