Game theory-based optimization of distributed idle computing resources in cloud environments

With the rapid cloud technological advancement and economic growth, more and more organizations have purchased cloud resources for daily business operations besides building their own IT infrastructures. Thus, it is very important to understand the economy of cloud computing. In this paper, we mainly focus on examining the private idle computing resources owned by various organizations who are willing to form a network of ad hoc cloud provider and sell the services to cloud users. In such a case, the organizations cannot only meet their own demands, but also sell their idle computing resources in the form of ad hoc cloud. Naturally, the organizations, as provider, aim at maximizing their own profit through adjusting business costs and sale prices. Due to the uncertainty of the amount of idle computing resources, dynamic pricing is challenging. We approach the problem from the perspective of game theory and formulate it as a non-cooperative game among multiple organizations, i.e., the game player. For each player, a utility function is used to represent its profits. The players choose request strategies and sales service strategies to maximize the utility function. This paper has proved that there exists Nash equilibrium for this game problem. We proposed an iterative proximal algorithm (IPA) for calculating the Nash equilibrium. After analyzing the convergence of the IPA, we found that the algorithm converges to the Nash equilibrium solution when reasonable conditions are satisfied and conforms to the theoretical proof. Experimental results show that our proposed algorithm can quickly converge to a stable state, and by calculating the appropriate service (resource) request strategies and selling service strategies for all organizations, organizations' profit are increased compared to without IPA algorithm. (C) 2019 Elsevier B.V. All rights reserved.