A Novel, Adaptive Replica Control Protocol for Distributed Intrusion-tolerant Environments

Replication is a key technique to provide high dependability, fault tolerance, and desirable performance in a distributed intrusion tolerant system. But existing replica control protocols cannot keep the dependability and the consistency of the replicated data at the same time, especially when some failure occurs. By virtue of the concept of virtual partition, this paper presents a novel, adaptive replica control protocol, which can dynamically adapt to the changes of underline communication topology in the presence of failure, especially partition failures, so not only provides the ability to tolerate any number of omission, performance and Byzantine failures, but also guarantee consistency of the protocol.