A framework for scheduling of heterogeneous workloads in cloud management system

Abstract

Cloud computing is predominantly utilized in many emerging applications like economics e governance e commerce and so on Despite of its on demand resource provisioning and pay per usage model lots of naïve users prefer Cloud Service Providers CSPs for processing their jobs effectively in a cloud environment In a cloud environment naïve users may submit their jobs with varied requirements and so the jobs are considered as heterogeneous jobs or heterogeneous workloads The heterogeneous jobs encompasses of independent jobs dependent jobs and parallel jobs The focus of this research work is to process heterogeneous jobs of type independent and parallel jobs The independent jobs may either demand for a single Virtual Machine VM or multiple VMs for their computation depending upon the demand of naïve users In the perspective of naïve users their ultimate need is to process more number of jobs at minimal their processing cost On the other hand the ultimate goals of CSPs are to increase and effectively utilize the available resources in a data centre Job scheduling and resource scheduling play a vital role in fulfilling the general needs of the naïve users and in meeting the goals of CSPs A cloud environment has several CSPs for offering services to naïve users simultaneously The spotlight of this research work is to meet the goals of CSPs but for detail study it considers only a single CSP and how to effectively utilize it newline