Distributed and Collaborative Intrusion Detection Systems

Abstract

Intrusion detection systems help improve the security of networks by providing early warning and response. To improve the detection of attacks, sharing data among distributed nodes or terminals and collaborating on a decision is key. This paper presents a distributed and collaborative intrusion detection (DaCID) system that relies on Dempster Shafer theory of evidence for fusing data from multiple nodes. In this approach the detection is done collaboratively and the decision is distributed among all nodes. DaCID is more robust than other systems since it is completely distributed and the decision is made autonomously at each node. Simulation results demonstrated that DaCID's performance approaches that of a centralized method.