Fault detection and recovery coverage improvement by clock synchronized duplicated systems with optimal time diversity

摘要

Fast fault detection and recovery are indispensable to improve the availability of fail-safe and fault-tolerant systems. Fast fault detection and recovery can be realized by clock-synchronized duplicated systems with optimal time diversity, which have fast fault detection and recovery features. The experiments presented in this paper compare fault detection coverage and retry recovery coverage of a clock-synchronized system without optimal time diversity, a clock-synchronized system with optimal time diversity, and an I/O synchronized system. The experimental results show that clock-synchronized systems with optimal time diversity realize the highest fault detection and recovery coverage.