Forward Load Aware Scheduling for Data-Intensive Workflow Applications in Cloud System

Scientific workflows and other large complex problems are benefited from cloud infrastructure for processing, storage and communication. Workflow scheduling is recognized as a well-known NP-complete problem. In this paper, we propose a load-balanced scheduling technique for workflow applications in a cloud environment. The proposed algorithm works in two phases. In the first phase, priorities of all the tasks are calculated in bottom up fashion while virtual machine selection and scheduling take place in the second phase. This technique also considers the overall load to be executed immediately after the execution of current task node. We compare the simulated results with the benchmark scheduling heuristic named as heterogeneous earliest finish time (HEFT) and a variation of the proposed technique. All the simulations are done by using the benchmark scientific workflow applications. We show that our proposed method remarkably display the performance metrics i.e., minimization in makespan and maximization in average cloud utilization.