Generating high quality schedules for scientific computation on a computational grid is a challenging problem. Many scheduling algorithms in grid computing are for independent tasks. However, communications commonly occur among tasks executed on different grid nodes. In this paper, an eXtended Communication-Inclusion Generational Scheduling (XCIGS) algorithm is proposed to schedule dependent tasks of an application with their DAG. During scheduling, those ineligible tasks are momentarily ignored, and a Buffer Set of Independent tasks (BSI) is conducted to leverage the utilization of grid resources. The predicted transferring time, the machine ready time and the expectation completion time of all predecessors are taken into consideration while an alternative auxiliary algorithm dynamically makes the schedule. Corresponding experimental results suggest that it betters resource utilization of grid experiments and improves execution performance.
展开▼
