A Load Adaptation Scheme For Renewable Energy In Data Centers

The computing capacity and scale of datacenters are increasing to meet the demands for cloud computing.Mega datacenter(such as those of Google,Amazon and Apple)can host thousands of servers and require up to tens of megawatts of electricity.The large electricity bills and the negative impacts on environments accelerate the use of renewable power to supply data centers.However,the volatility and intermittency of renewable power often cause the challenges in the stability of the grid and data center systems as well as decrease the utilization of renewable power.In order to address these problems about renewable energy management in data centers,existing proposed schemes unfortunately fail to alleviate the impact of renewable power fluctuation on the stability of the grid and data centers.This paper proposes a load adaptation scheme for renewable energy in data centers,which consists of the renewable power smoothing scheme and a load adaptation scheme.First,the renewable power smoothing scheme carries out the optimal charge/discharge operation via computing the minimum variance of the renewable power that is ultimately supplied to data centers per hour to smooth the fluctuation of renewable power as much as possible.Second,the load adaptation scheme is a greedy algorithm for delaying loads.Based on the renewable power smoothing scheme,the load adaptation scheme further maximizes the utilization of renewable power.The salient features of the scheme are to address the volatility of renewable power and maximize utilization of renewable power via smooth supply.Extensive trace-driven experimental results demonstrate that the renewable power smoothing scheme significantly reduces the energy switching times,and the load adaptation scheme increases the utilization of renewable power from 19% to 81%.