Intrusion Tolerance System Model Based on Adaptive Probability Voting Mechanism

摘要

Current majority voting mechanism usually used in intrusion tolerant system have some limits. To solve this problem, a novel adaptive probability voting mechanism based on double threshold values is proposed. The intrusion tolerance system based on this voting mechanism can provide correct and reliable services continually in case of existing fault or intrusion and detect the faulty server replicas, thus achieving the intrusion tolerance target. The output of voting is produced by account coherence rate among response results of server replicas and voting reliability probability' history record of server replicas and dynamically setting the double threshold values, so that the voting results are more correct, reliable and adaptive.