Demand Response in Data Centers: Integration of Server Provisioning and Power Procurement

Data centers are among the fastest growing industries in the United States economy in terms of IT service and usage of data. However, data centers consume an excessive amount of energy that contributes to huge energy costs and CO₂ emissions; thus, they face challenges in energy cost savings and making a positive impact on the environment. Therefore, data centers have been striving to: 1) improve energy efficiency in server operations by applying server provisioning techniques, and 2) utilize renewable energy and energy storage for power procurement. Although data centers have obtained financial and environmental benefits from the aforementioned efforts, there still exists significant room for improvement by facilitating demand response based on the co-optimization of server operations and power procurement, which has not been well implemented in the current practice. The main objective of this paper is a new holistic approach that considers server operations and power procurement simultaneously for implementing demand response. Also, this paper develops a decision-making model based on two-stage stochastic programming to successfully address the inherent stochastic nature and demonstrate the impact of the proposed model via numerical experiments. The proposed model is designed to support practitioners’ strategic and operational decision-making for demand response in data centers.