Energy-Aware Resource and Revenue Management in Federated Cloud: A Game-Theoretic Approach

Reduction of energy expenditure is becoming an important issue for a cloud provider (CP) when providing cloud services over the Internet. Federation among CPs, whereby a set of CPs cooperating together to provide virtual machine (VM) instances requested by users, can be an effective solution to address this issue. In this paper, we present an effective capacity-sharing mechanism in a federated cloud environment that can lead to a global energy sustainability policy for the federation and encourages them to cooperate. A coalition game theory was utilized to model various interactions among providers. However, unlike the existing approaches, the proposed game model looks for a set of low-energy-cost CPs in a federation such that the social welfare is maximized and provides a fare and suitable revenue for them. In addition, we consider the demand variations of internal users of a CP when sharing VM resources. Moreover, a detailed analysis of various costs and revenue aspects is presented. Experiment results demonstrated that the proposed game model is able to maximize the social welfare of the CPs while satisfying the fairness and stability properties of the federation.