Allocation of Phase-Based Scheduler for MapReduce Job Scheduling

Abstract: Hadoop MapReduce is effective user interface design classic for large scale data handling. MapReduce has two levels: Task-level and Phase level. In existing system, that focuses on scheduling at task level which tasks can have changing resource requirements. There are some difficult to efficiently apply accessible resources to reduce job implementation time. To report this limitation, this project proposes a Phase-Based Scheduler. Map Reduce which allocates resource information about status of every Phase the phase-based to executed job scheduling. The job scheduling of phase based is executed by the MasterNode, which handle & service lots of list of jobs in the system. EachNodeManager (slave node) from time to time getting a heartbeat message to the scheduler. Getting the status message from a NodeManager running on machine, the scheduler divides the use for fixed of phases for the tasks using the jobs phase-based resource requirement. This improves to reduce job implementation time. This is achieving high job performance and resource utilization.

Keywords: Big Data, Hadoop, Scheduler, MapReduce, Phase-Based Scheduler