Parallel and Distribute Processing for Virtual MapReduce Clusters by using improvised Hybrid Job Scheduling Algorithm

Abstract

MapReduce is a programming model that defines a MapReduce job for instance, a map perform task then reduce perform task. This model splits the job into several map perform tasks and reduce perform tasks at run time. It also accomplishes these tasks in parallel on a MapReduce cluster. We have researched a resourceful and appropriate scheduling scheme called as hybrid job-driven scheduling scheme (JoSS) which source higher map and reduce data-locality. But, in this existing JoSS scheduling scheme, virtual MapReduce cluster does not provide flexibility over multiple workloads for load balancing. For this purpose, we have enriched native JoSS with advanced JoSS by adding a new functionality called virtual MapReduce cluster to provide flexibility to JoSS. This enhanced work achieves load-balancing and also improves job performance.