Distributed coordination of internet data centers for energy cost minimization

摘要

Internet data center (IDC) has been built to meet the increasing requirements of computing, storage and network. However, a large number of servers will bring a sharp increase in energy consumption. It is essential for cloud providers to manage servers' safe and reliable operations based on smart grid techniques to reduce the energy cost. Therefore, this paper builds a new energy cost model that considers servers' energy cost in different state, lifecycle of server, workload balance and time-of-use price. And a distributed optimization algorithm based on Predictor Corrector Proximal Multiplier (PCPM) method is proposed. The proposed algorithm is also compared with other methods, such as traditional distributed strategy and Particle Swarm Optimization algorithm. The simulation results demonstrate the scalability, effectiveness and fast convergence of the proposed distributed optimization algorithm.