Proactive service migration for long-running Byzantine fault-tolerant systems

A proactive recovery scheme based on service migration for long-running Byzantine fault-tolerant systems is described. Proactive recovery is an essential method for ensuring the long-term reliability of fault-tolerant systems that are under continuous threats from malicious adversaries. The primary benefit of our proactive recovery scheme is a reduced vulnerability window under normal operation. This is achieved in two ways. First, the time-consuming reboot step is removed from the critical path of proactive recovery. Second, the response time and the service migration latency are continuously profiled and an optimal service migration interval is dynamically determined during runtime based on the observed system load and the user-specified availability requirement.