Incorporating Energy-Aware Mechanism into Workflow Scheduling Policy in Heterogeneous Distributed Systems

In the past decades, plenty scheduling policies have been proposed for improving the execution performance of workflow applications. However, few of them have addressed the issue of energy conservation. In this paper, a novel scheduling metric, namely Minimal Energy Consumption Path, is introduced to reducing the energy consumption when scheduling large-scale workflows. The proposed scheduling metric can be incorporated into current schedulers so as to enhance their capability of energy conservation as well as maintain their scheduling performance. A lot of experiments are performed by using different scheduler and experimental settings to evaluate the proposed scheduling metric and the results indicate it is effective for reducing the storage-related energy consumption, especially when the target workflows are data-intensive.