Optimal Workload and Energy Storage Management for Cloud Data Centers

Electricity expenditure comprises a significant fraction of the total operation cost in cloud data centers and cloud service providers are required to reduce their electricity cost as much as possible. In this paper, we consider electricity cost minimization problem in data centers with energy storage under wholesale electricity markets by considering both the delay-sensitive and delay-tolerant workloads. Considering the stochastic nature of workload arrivals and electricity price, we formulate the problem as a stochastic program and propose an online control algorithm based on the technique of Lyapunov optimization to solve it approximately. The analytical results show that our algorithm can achieve close-to-optimal performance with the increase of energy storage capacities while guaranteeing the quality of service of different workloads.