Because data centers are energy-intensive industries, data centers have made significant efforts to improve energy efficiency to save energy cost and utilize renewable energy to contribute to environmental sustainability. In addition, data centers have been considered as a potential application of demand response because they can possibly adjust energy consumption to process workload in response to time-of-use prices and intermittent renewable generation. Quality of service (QoS) should be guaranteed at the desired level to respond to users’ requests with their satisfaction in an effort to achieve energy-efficient and sustainable data center operations. Given this context, this study intends to develop a two-stage stochastic program that can be used to investigate optimal data center operations based on demand response that results in minimum energy cost with QoS guarantee against the stochastic workloads. Then, an additional chance constraint and risk constraint are considered as means of managing the level of QoS guarantee, and comprehensive numerical experiments are conducted with various parameter settings to evaluates the impact of the QoS guarantee with different levels on overall performance. The results show that the risk associated with the QoS guarantee for data centers needs to be properly managed to improve energy efficiency.

Motivated by applications in data centers, we consider a scenario where multiple classes of requests arrive at a dispatcher at time-varying rates which historically has daily or weekly patterns. We assume that the underlying environment is such that at all times the load from each class is very high and a large number of servers are necessary which, for example, is fairly common in many data centers. In addition, each server can host one or more classes. Design, control and performance analysis under such heterogeneous and transient conditions is extremely difficult. To address this shortcoming we have suggested a holistic approach that includes a combination of sizing, assignment, and routing in an integrated fashion. Our proposed approach decomposes a multi-dimensional and non-stationary problem into a one-dimensional, simpler and stationary one, and achieves time-stability by introducing an insignificant number of dummy requests. Based on time-stability, our suggested approach can provide performance bounds and guarantees for time-varying and transient system. Moreover, we can operate the data centers in an energy-efficient manner via suggested approach.