An Enhanced MIN-MIN algorithm for workflow scheduling in Cloud Computing

Abstract

In Cloud Computing environment, scheduling of Cloudlets and resource allocation of Cloudlets are two major issues that need to be considered. To improve the system performance, it is necessary that the Cloudlets scheduled on the resources available must be efficiently executed. For scientific applications, workflow scheduling is a significant issue. To handle this issue, a new mechanism called Enhanced Min-Min (E-Min-Min) is introduced inspired by the Min-Min algorithm. This scheduling algorithm not only schedules the jobs according to their completion time but considers the data transfer time too with equal priority and makes an aggregated function for scheduling of jobs to the best suitable resources. During the scheduling process, the main focus is on the minimization of Makespan, maximize the resource utilization and load balance level and also minimize the cost of scheduling. Workflow scientific applications such as Montage, Cybershake, InSpiral, SIPHT are considered. In workflow the tasks are represented as nodes and their dependencies represents the relationships among them. WorkflowSim simulator with Net Beans IDE is used to create the simulation environment and the results are validated over it. The proposed algorithm shows the better results as compared with the basic Min-Min and Modified Min-Min algorithm.