International Journal of Engineering and Technology

In this paper, an efficient Batch-Mode scheduling heuristics have been proposed for balancing the load in the Desktop Grid environment. The proposed heuristic works in two phases: In first phase, we are making a schedule of Expected Execution Time (EETuv) for all tasks w.r.t. resources by following Max-Min for m tasks & Min-Min and Max-Min alternatively for remaining n-m tasks, where n & m are the number of tasks (Tu) and resources (Rv) respectively. Then scheduling is being done as per the minimum EETuv taken by the tasks w.r.t. the resources without considering the load imbalance on resources. In second phase, to remove the load imbalance, tasks will get transferred from maximally loaded to minimally loaded resources. The concept has tested experimentally by using GridSim 5.2, and results proves that proposed heuristic performs well on comparing with Min-Min, Max-Min and LJFR-SJFR heuristic for minimizing Makespan, Flowtime and Average Completion Time value.