A SURVEY ON SCHEDULING IN HADOOP FOR BIGDATA PROCESSING

Abstract

Scheduler is a way of assigning jobs to various available resources. Job assignment is done in such a way that it should minimize starvation of jobs and maximum utilization of resources. Scheduler efficiency is dependent on realistic workloads and clusters. Here, we are introducing a scheduler technique for real, multi node, complex system as a Hadoop. Based on size, priority is assigned to make scheduling efficient. This makes our algorithm different than conventional scheduling algorithm. Performance of the proposed scheduling technique can be increased by assigning Deadline constraints for local optimality of data. Map Reduce framework is used for execution of jobs.