Adaptive Task Scheduling Strategy for Economy Based Grid.

In this paper, we propose an adaptive task scheduling strategy in an economy based grid. The proposed strategy is meant to ensure consistent performance, despite of variation in individual resource performance (i.e. meeting user QoS requirement like budget and deadline). The proposed strategy maintains performance index of grid resources, which depicts resource past performance. The performance index shows resource vulnerability towards variance in their past performance. Further, over the time it dynamically updates the resource performance index based on completion result received of the assigned task from the grid resource. Later, when grid resource broker have tasks to schedule on grid resource(s). It makes use of the performance index and dictated by this performance apply different intensity of heuristics like dummy task allocation and penalty/reward heuristic and/or both. In the presence of the variance in resource performance, the proposed strategy can effectively schedules grid jobs. And executes more jobs successfully, within the user specified deadline and allotted budget. It can also improve the overall execution time and minimizes the execution cost of grid jobs. Hence, can help in upholding trustworthiness of grid environment.