On selecting rollback points for error recovery

A quasi-synchronous checkpointing algorithm that prevents contention for stable storage

“…In distributed computing systems, checkpointing and rollback recovery are well-established techniques for handling failures [3,4,7,8,10,1,16,22,25,28,19,29,26,13,9,24,14,11]. Existing checkpointing algorithms can be classified into three main categories -asynchronous, synchronous and quasi-synchronous [23].…”