Demand Response of Data Centers: A Real-time Pricing Game Between Utilities in Smart Grid

We study the demand response (DR) of geodistributed data centers (DCs) using a dynamic pricing scheme. Our proposed pricing scheme is constructed based on a formulated two-stage Stackelberg game where each utility sets a real-time price to maximize its own profit in Stage I; and based on these prices, the DCs’ service provider minimizes its cost via workload shifting and dynamic server allocation in Stage II. First, we show that there exists a unique Stackelberg equilibrium. Then, we propose an iterative and distributed algorithm that converges to this equilibrium, where the "right prices" are set for the "right demand". Finally, we verify our proposal by traced-base simulation and results show that our pricing scheme outperforms other baseline schemes significantly.