A Joint Energy Efficiency Optimization Scheme Based on Marginal Cost and Workload Prediction in Data Centers.
Kaixuan Ji,Fa Zhang,Ce Chi,Penglei Song,Biyu Zhou,Avinab Marahatta,Zhiyong Liu
DOI: https://doi.org/10.1016/j.suscom.2021.100596
IF: 4.923
2021-01-01
Sustainable Computing Informatics and Systems
Abstract:With the widespread development of cloud computing, the dramatic increase in the number and size of data centers (DCs) has resulted in substantial energy consumption and serious environmental problems. This creates a challenge for the further development of DCs. It is imperative to improve the overall energy efficiency of DCs. According to the statistics, servers and cooling systems are the main energy-consuming components of DCs. Recently, several energy-efficient strategies have been developed to address these problems. However, most of these works only consider the energy optimization of servers or cooling systems separately. Therefore, a Joint Energy Efficiency Optimization Scheme (JEES) is proposed in this paper, where the energy consumed by servers and the cooling system is jointly considered, and coordinately optimized. JEES includes a dynamic online task scheduling algorithm based on marginal cost evaluation, a resource management strategy that integrates the workload prediction technique to manage resources, and a task migration method using marginal cost evaluation. By using the proposed techniques, the total energy consumption of DCs can be reduced. Extensive experiments have been conducted based on real-world workload traces, and the results demonstrate that compared with other techniques, the proposed scheme effectively improves the overall resource utilization and reduces the total energy consumption of DCs.