Abstract：Checkpointing and rollback recovery of Unix process is the underlying technique of fault tolerance of distributed system and parallel environment.To save and restore the state of active file of the process is an important aspect of checkpointing and rollback recovery.This paper presents an approach called Write Operation Buffering to support this capability.WOB approach buffers all the write operations after a checkpoint until the next one,making all the operations between two checkpoints atomic as a whole.By choosing a suitable size dynamically for memory buffer,and by hiding the latency of flushing the buffer,the WOB approach achieves an overhead lower than other approaches.
2000, Vol. 28 
