A MapReduce Based Job Scheduling Framework to Improve Performance of Processing over Distributed Resources

Abstract

Job scheduling is a process where more number of jobs are assigned with particular deadline, and then they are executed according to their schedule with in the given deadline. Sometimes job execution depends on the process which we may select and the resource availability. Earlier there exists so many job scheduling algorithms like FIFO, SJF, EDF etc. All these algorithms execute the jobs depending on first come first serve or shortest time basis. In these approaches some jobs may not get a chance to execute when a fixed deadline was given and then penalty will take place. Whenever there is lack of required resource to execute due to others are using it, may extend the waiting time of other jobs. So to conquer these issues we propose a MapReduce based job scheduling frame work with deadlines for each job. In this a process related to a particular job will be Mapped and Reduced to accomplish before or within the deadline. This will also release the ongoing resource promptly to utilize by other jobs. Compared with existing approaches the anticipated one yields better results over fast and time. These implementations enhanced the job scheduling process for fast and better utilization of distributed resources.