A New Adaptive Data Center Resource Provisioning Scheme Based on the Dual-Level Cooperative Game Approach

With the popularity of 5G network technology, cloud computing paradigm in data center (DC) networks has drawn increased attention. DC is a pool of computational and storage resources interconnected using a communication network. To handle the growing demands of cloud computing services, resources in DC should be used efficiently. In this paper, we design a new DC resource provisioning scheme based on the cooperative game theory. Using the matching game model and Mood value, we propose new resource allocation algorithms, which attempt to equalize users' satisfactions with respect to an effective distribution of the DC resource. Due to the competitive and coordinative DC environments, our proposed game approach can adaptively respond to the current cloud computing service requests. The main novelty possessed by our resource provisioning scheme is to capture the dynamics of DC network operations with the consideration of efficiency and fairness. Finally, we show the advantages of our proposed scheme through computer simulations. Comparing with other existing protocols, performance evaluations demonstrate that our game-based approach can outperform the existing state-of-the-art methods in terms of DC resource usability, service success ratio and fairness. In addition, we provide the guidance on the future research direction including other issues.