A Solo-Checkpointing Co-Recovery Mechanism for Reliability Improvement of Cyber-Physical Systems

The application of cyber-physical systems (CPSs) in health and health-care industry demands a high level of reliability in the design and manufacturing of these systems. This paper addresses the reliability of local and remote surgical systems as the real-world examples of networked cyber-physical systems. We introduce a solo-checkpointing co-recovery mechanism to minimize checkpointing overheads while keeping the system’s reliability as high as possible. We have developed a method to find consistent checkpoints among components of the CPS in a timely and efficient manner. We have evaluated this mechanism on a prototypical surgery training and assistive system developed in our research laboratory. We have injected a total of 560 faults into the system while working in the operational mode. Results show that our mechanism can mask almost all injected faults with a negligible fault detection latency.