Optimistic Message Logging for Independent Checkpointing in Message-PassingSystems

摘要

Message-passing systems with communications protocol transparent to theapplications require message logging to ensure the consistency between checkpoints. This paper describes a periodic independent checkpointing scheme with optimistic logging to reduce the performance degradation during normal execution while keeping the recovery cost acceptable. Both time and space overhead for message logging can be reduced by detecting messages that need not be logged. A new checkpoint space reclamation algorithm is for any possible future recovery. Communication-trace-driven simulation for several hypercube programs is used to evaluate the techniques. Independent checkpointing, Recovery lines, Optimistic message logging, Message systems.