Design of a Fault-Tolerant Mechanism for Multi-Agent Based Intrusion Detection System

摘要

To cope with the various intrusion patterns, intrusion detection system based on multiple independent autonomous agents working collectively was proposed recently. However, most of intrusion detection system does not include fault tolerance mechanism when autonomous agent was failed. This is a critical problem for intrusion detection system relying on autonomous agents. Intrusion detection system must be fault tolerant by being able to recover from system crashes, either accidental or caused by malicious activity. Upon startup, the intrusion detection system must be able to recover its previous state and resume its operation unaffected. In this paper, we propose a novel fault tolerance mechanism for agent based intrusion detection that able to detect agent stop when agent was failed and maintain audit information from failed agent.