An Energy-Aware Optimization Model Based on Data Placement and Task Scheduling

Recently, technologies on reducing energy consumption of data centers have drawn considerable attentions. One constructive way is to improve energy efficiency of servers. Aiming at this goal, we propose a new energy-aware optimization model based on the combination of data placement and task scheduling in this paper. The main contributions are: (1)The impact of servers´ performance on energy consumption is explored. (2) The model guarantees 100% data locality to save network bandwidth. (3) As tasks involved in cloud computing are usually tens of thousands, in order to solve this large scale optimization model efficiently, specific-design encoding and decoding methods are introduced. Based on these, an effective evolutionary algorithm is proposed. Finally, numerical experiments are made and the results indicate the effectiveness of the proposed algorithm.