A game-based resource pricing and allocation mechanism for profit maximization in cloud computing

摘要

In cloud computing environment, Software as a Service (SaaS) providers offer diverse software services to customers and commonly host their applications and data on the infrastructures supplied by Infrastructure as a Service (IaaS) providers. From the perspective of economics, the basic challenges for both SaaS and IaaS providers are to design resource pricing and allocation policies to maximize their own final revenue. However, IaaS providers seek an optimal price policy of virtual machines to generate more revenue, while SaaS providers want to minimize the cost of using infrastructure resources, and comply with service-level agreement contracts with users at the same time. In this situation, there exists conflict in maximizing revenue of both IaaS and SaaS providers simultaneously. In this paper, we model this revenue maximization problem as the Stackelberg game and analyze the existence and uniqueness of the game equilibrium. Moreover, considering the impact of resource price on users' willing to access service, we propose a dynamic pricing mechanism to maximize the revenue of both SaaS and IaaS providers. The simulation results demonstrate that, compared to fixed pricing and auction-based pricing mechanisms, the proposed mechanism is superior in the revenue maximization and resource utilization.