Aggregation Measure Factor-Based Workflow Application Scheduling in Heterogeneous Environments

With the development of heterogeneous distributed computing environment, workflow application scheduling has become an important and challenging problem while Quality of Service (QoS) guarantees are ensured for science workflows. In this paper, we first introduce an aggregation measure factor to balance the execution time and cost of workflow applications. Then, we propose an aggregation measure factor-based scheduling algorithm (AFSA) for workflow applications in a heterogeneous distributed environment. The proposed algorithm through allocating the sub-budget and sub-deadline for each task to choose available processors takes into account of the budget and deadline aggregation to select the processor for the science workflows. Furthermore, we introduce both a planning success rate and normalized deadline (ND) as performance metrics to evaluate workflow application scheduling algorithms. Furthermore, we use both a randomly generated data set and a real-world workflow data set in our experiments for the performance evaluation. Moreover, our experimental results demonstrate that the proposed AFSA has a higher balance factor and an almost equal or higher planning success rate under different workflow application structures compared to the existing algorithms, BHEFT, HBCS, WMFCO, and DBCS.