Abstract

The computational grid has emerged as an attractive platform to tackle various science and engineering problems. One of the challenging issues in the grid associated with the effective utilization of the heterogeneous resources is scheduling. This paper designs and implements a task-scheduling algorithm considering the dynamicity of the resources and the tasks. We explain the concept of queue’s virtual time and combine it with virtual finish time and the service time error to allocate resources to the tasks for improved fairness and better throughput. The detailed performance evaluation of virtual finish time driven scheduling algorithm is carried out through a series of simulations by varying the number of tasks and processors of different capacities to optimize the cost and execution time of the tasks to achieve fairness.

Key words: Computational grid, heterogeneous resources, dynamicity, task-scheduling