A Markov-Based Intrusion Tolerance Finite Automaton

Abstract

It is inevitable for networks to be invaded during operation. The intrusion tolerance technology comes into being to enable invaded networks to provide the necessary network services. This paper introduces an automatic learning mechanism of the intrusion tolerance system to update network security strategy, and derives an intrusion tolerance finite automaton model from an existing intrusion tolerance model. The proposed model was quantified by the Markov theory to compute the stable probability of each state. The calculated stable probabilities provide the theoretical guidance and basis for administrators to better safeguard network security. Verification results show that it is feasible, effective, and convenient to integrate the Markov model to the intrusion tolerance finite automaton.