A distributed first and last consistent global checkpoint algorithm
Yoshifumi Manabe · 2002
Distributed coordinated checkpointing algorithms are discussed. The first global checkpoint for a checkpoint initiation is a set containing the checkpoint for each process in which any checkpoint before the element is not consistent with the initiation. The last global checkpoint for a checkpoint initiation is a set containing the checkpoint for each process in which any checkpoint after the element is not consistent with the initiation. This paper presents distributed algorithms that make the first and last global checkpoints consistent with a minimum number of check points taken in each process.