Affiliations: School of Computer Science and Technology, Xidian University, Xi'an, Shaanxi, China
Correspondence:
[*]
Corresponding author: Yuping Wang, School of Computer Science and Technology, Xidian University, Xi'an 710071, Shaanxi, China. E-mail: [email protected].
Abstract: Soaring power consumption and limited communication bandwidth are the most critical issues involved in cloud computing. Reducing energy consumption will not only cut down the operational costs of data centers, but also reduce the amount of greenhouse gases emissions. In order to improve energy efficiency of servers, a new multi-objective bi-level programming model based on MapReduce is proposed in this paper. In the model, first, the relationship between performance and energy consumption of servers is taken into account. Second, data locality can be adjusted dynamically according to current network state. Third, data placement policies and task-scheduling strategies are considered simultaneously as a whole. In order to solve the model efficiently, specific-designed encoding and decoding methods are introduced. With these, a new effective multi-objective genetic algorithm on the basis of Multiobjective Evolutionary Algorithm Based on Decomposition (MOEA/D) is presented. As tasks involved in cloud computing are usually tens of thousands, a local search operator is designed in order to accelerate convergent speed of the proposed algorithm. Finally, numerical experiments are made and the results indicate the reasonableness of the model and the effectiveness of the proposed algorithm.
Keywords: Energy-aware, data locality, multi-objective optimization, bi-level programming, MapReduce
