Cost and Makespan Trade-off Management for Scheduling Workflow on Utility Grids

Abstract

Reaching a high performance in Utility Grids, among other things, requires an efficient workflow application scheduling on shared resources. The shared resources among users are autonomously managed in these environments. Application tasks are submitted to a scheduler; next the scheduler selects a resource for executing the task.Finally, the resource schedules the assigned tasks according to its own allocation policy. Therefore, users have no explicit control on allocating resources to application task. As a result Quality of Service cannot be guaranteed. In this paper, a heuristic algorithm known as Minimum First-fit Cost-Makespan Trade-off (MinFCMT) is developed in order to effectively schedule an application in Utility Grids so that the application makespan and allocation-cost can be minimized. To evaluate the MinFCMT heuristic algorithm a widespread simulation of a real parallel workload models plus the synthetic workflow is used according to the existing application scheduling algorithm. The results demonstrated that the MinFCMT algorithm is far more effective than the existing algorithms in terms of time-complexity of algorithm.