Power Dispatching in Cloud Data Centers Using Smart Microgrids: A Game Theory Approach

Abstract

The proliferation of cloud-based applications in smart systems has made the cloud data center a vital and critical part for ensuring a connected world. Due to their energy-hungry servers and huge power facilities, cloud data centers tend to consume a lot of power. In fact, the data centers use about 1.4% of all the power generated on the planet. On one hand, this has a negative impact on the power grid and may cause blackouts. On the other hand, it has a negative impact on the environment and it is mainly involved in global warming. Thus, one of the most important challenges in cloud data centers is power consumption minimization. In this paper, we propose a microgrid-cloud based architecture and study the grid power dispatching problem to cloud data centers. At first, we model the power quantity demand between the data centers and smart microgrids as a non-cooperative game, due to their non-cooperative power demands behavior. Then, we try to allocate the optimal quantity of power to each data center according to its green power consumption, network bandwidth usage and its network equipment power usage. Second, we formulate the game payoff function as a non-linear optimization problem and solve it using Lagrange multipliers and KarushKuhnTucker (KKT) conditions. Finally, we compare the performance of our approach with two power allocation algorithms and showed that our game is more effective and reduces power load with a rate of 40.5%. Furthermore, our scheme incites data centers to use green energy and significantly reduces dioxide carbon emission.