The Research And Implementation Of Ad Hoc Cooperative Intrusion Detection System And Its Cooperative Algorithm

With the development of communication technology, the application of Ad Hoc Network is expanded to commerce and civil areas rapidly. At the same time, the security issues caused by the structural properties of Ad Hoc Network are concerned by the people increasingly. Ad Hoc Network has no fixed network borders, the traditional firewall technology can not be applied in it, and the intrusion detection technology plays an important role in its network security. At present, the common security strategy of Ad Hoc Network is combining defense strategy with intrusion detection. Constructing an intrusion detection system which is consistent with the characteristic of Ad Hoc Network has great significance for the development of Ad Hoc Network.Based on the framework of the tradition distributed intrusion detection system, this paper presents a Distributed Cooperative Intrusion Detection System(DCIDS) which combines subjective trust theory and feedback control theory. The system introduced subjective trust level to weigh the influence of nodes in the cooperative process, which reduced probability of errors caused by using balanced voting, and improved the detection efficiency of the system. Through result feedback, the system can automatic updates trust levels of the cooperation nodes, which improved self-adapting ability of cooperative nodes.While presenting the architecture and implementation of the DCIDS in detail, this paper designs the cooperative detection algorithm and the trust level auto-update algorithm, proposes a transmission protocol between nodes which based on C-S structure, and introduces a method which uses DMS(data multiplexing system) and multi-slot to dealing with collaborations happened at one time. In the end, the system was tested by simulated intrusion. The result analysis shows that the system completed the cooperative detection properly, and the system real-time updated the trust level of nodes perfectly, having a strong self-adaptive ability.