An Uncompleted Information Game Based Resources Allocation Model for Cloud Computing

Considering the competing characteristics under multi-tenant environment in cloud computing and aiming to improve the profit of both the resource supply and demand sides, an uncompleted information game based cloud resource allocation model is proposed. Firstly, a hidden Markov model (HMM) is introduced to predict the current bid of service providers based on the historical resource demand. Then a game model for dynamical pricing is established base on the predicted bid value. It can motivate service providers to choose the optimal bidding strategy in accordance with overall interests and so as to achieve maximum benefits. Finally, a resource allocation model on the basis of unit prices of different types of resources is put forward to guarantee optimal gains for infrastructure provider (INP). The allocation model can support synchronous allocation for both multi service providers and various resources. Simulation results show that, in the pricing game model, the predicted price is close to the actual transaction price which is lower than the actual valuation, so it can guarantee the profit of service providers; the resource allocation model can simultaneously increase infrastructure provider’s revenue.