Optimal Tiered Pricing Strategy For Monopolistic Cloud Services Considering Heterogeneous Customers' Demand

Computing resources become necessary for fulfilling daily life routines which is the same as utility services such as water,electricity,gas and telephone.The increasingly popularity of cloud computing paradigm brings out the challenging proble m of optimal strategy on pricing cloud services which becomes the focus of cloud service industry and academic circle.However,most of nowadays studies concentrate o n pricing strategy which aim to maximize the profit of cloud service providers(CSPs)Moreover,tiered services are provided by CSPs in practice,which are designed to meet various demand of heterogeneous customers.However,only a few researches focus on tiered strategy of cloud services.On the aspect of a long-term and dynamic pricing mechanism,we build a bilevel programming model considering heterogeneous customers' demand and reach the optimal tiered pricing strategy with analytical method and numerical method for linear and nonlinear utility functions respectively.Furthermore,we investigate the impact of several factors and aim to provide effective suggestions for CSPs.The main research contents and results are listed as follows.(1)The design of pricing model.We develop a bilevel programming model which aims at maximizing CSP's profit while customer's utility is relatively maximized.In this model,the upper level is a monopoly CSP making decision on price and quality of tiered services to maximize profit considering operation cost and delay cost and lower level is all the customers deciding which level of service to buy to maximize their utility.(2)The optimal tiered pricing strategy under linear utility function.We prove the optimal strategy for CSP under linear utility function is to provide nontier service and the quality of service(Qo S)is highest while the price formula is deduced under certain circumstance.In this situation,delay cost no longer exists.Through the investigation of related factors,we illustrate the approprate price ranges in(0.5,0.6).The cost coefficient(c)and service utilization ratio(?)is positively correlated to optimal price amd negetively correlated to CSP's profit,while the price elasticity of demand shows the opposite result.To get more profits,it is more effective to control operational cost coefficient than improve quality factor.(3)The optimal tiered pricing strategy under nonlinear utility function.Since the model is more complicated if customers' utility is nonlinear,we solve the problem through numerical methods based on genetic algorithm.We indicate the optimal tiered strategy for CSP is 3-tier planning and the corresponding prices of each tier locate in [0.6,0.7],[0.4,0.5],[0.3,0.4] while the Qo S of each tier is 1,0.7,0.6.Through the investigation of related factors,we indicate that the existence of delay cost can raise the CSP's profit under 3-tier strategy and the price of each tier is positively correlated to operation cost coefficent.Different pricing discrimination strategy has great influence on the strategy of CSPs through affecting the choices of customers.Moreover,under different customer distribution,different pricing strategy should be applied.For example,in Gaussian distributed market,it's better for CSP to adopt high prices and high Qo S.In exponential distributed market,low prices and low Qo S will be better.The same tiered strategy results in the highest profit for Gaussian distributed market while highest market coverage for uniform distributed market.