Checkpoint restart is a facility offered by some database management systems (DBMSs) and backup-restore software. Checkpoints are taken in anticipation of the potential need to restart a software process.

Many ordinary batch processes on impersonal computers are time-consuming, as are backup and restore operations. They consist of many units of work. If checkpointing is enabled, checkpoints are initiated at specified intervals, in terms of units of work or of processing time. At each checkpoint, intermediate results and a log recording the process's progress are saved to non-volatile storage. The contents of the program's memory area may also be saved.

The purpose of checkpointing is to minimize the amount of time and effort wasted when a long software process is interrupted by a hardware failure, a software failure, or resource unavailability. With checkpointing, the process can be restarted from the latest checkpoint rather than from the beginning.

Checkpoint Frequency

Checkpoints should occur frequently enough to minimize wasted effort when a restart is necessary but not so frequently as to prolong the process unduly with checkpoint overhead. Optimal checkpoint frequency depends on the mean time between failures (MTBF), among other factors.