A communication-induced checkpointing algorithm using virtual checkpoint on distributed systems

Checkpointing is a fault-tolerant technique for restoring faults and restarting jobs quickly. The algorithms for checkpointing on distributed systems have been under study for years. These algorithms can be classified into three types: coordinated, uncoordinated and communication-induced algorithms. In this paper we propose a new communication-induced checkpointing algorithm that has a minimum checkpointing count equivalent to the periodic checkpointing algorithm, and relatively short rollback distance at fault situations. The proposed algorithm is compared with the previously proposed communication-induced checkpointing algorithms with simulation results. In the simulation, the proposed algorithm produces better performance than other algorithms in terms of task completion time in both fault-free and fault situations