Optimum Checkpoint Interval for MapReduce Fault-Tolerance

MapReduce is the efficient framework for parallel processing of distributed big data in cluster environment. In such a cluster, task failures can impact on performance of applications. Although MapReduce automatically reschedules the failed tasks, it takes long completion time because it starts from scratch. The checkpointing mechanism is the valuable technique to avoid reexecution of failed tasks in MapReduce. However, defining incorrect checkpoint interval can still decrease the performance of MapReduce applications and job completion time. In this paper, the optimum checkpoint interval is proposed to reduce MapReduce job completion time when failures occur. The proposed system defines checkpoint interval that is based on five parameters: expected job completion time without checkpointing, checkpoint overhead time, rework time, down time and restart time. Therefore, because of proposed checkpoint interval, MapReduce does not need to re-execute the failed tasks, so it reduces job completion time when failures occur. The proposed system reduces job completion time even though the number of failures increases and the performance of this system can be improved 4 times better than the original MapReduce