面向能耗优化的数据中心服务定价策略  

Data center service pricing strategies for energy consumption optimization

作　　者：余晓[1] 吴剑章[2] 王巍[2] 

机构地区：[1]东南大学继续教育学院,南京210096 [2]东南大学计算机科学与工程学院,南京210096

出　　处：《东南大学学报（自然科学版）》2012年第6期1080-1084,共5页Journal of Southeast University：Natural Science Edition

摘　　要：针对降低云数据中心的能耗问题,从经济学角度出发,通过制定3种合理的价格策略来引导消费者正确消费,从而达到节能减排的目的.鉴于数据中心规模庞大、承载任务繁重的特点,采用基于重负载近似的大规模排队系统对数据中心进行建模.针对不同利益方的利益诉求,探讨了满足能耗最小化、服务方收益最大化、全局效益最大化这3种不同优化目标的定价和节能策略,给出了最优价格和能耗的近似解.实验结果表明,近似解能够较好地满足各方的利益需求,并在数据中心服务器数量趋于无穷的情况下,渐进地逼近精确解.To minimize the energy consumption in green data center,three optimal pricing strategies are presented,from the economic point of view,to guide consumers toward right choices and achieve the purpose of energy saving.In view of the fact that the data center has a large size and takes heavy traffic,a large-scale queuing system based on heavy traffic approximation is used to model the data center.According to the interests and needs of different stakeholders,the strategy of pricing and energy consumption,which meets three different optimization objectives,the minimization of energy consumption,the maximization of the interests of supplier and the maximization of global benefits,is discussed.The best price and energy consumption of the approximate solution are given.The experimental results show that the proposed approximate solution can meet the interests and needs of all parties.Furthermore,when the number of servers in data centers approach infinity,the approximation solutions gradually approximate the exact solutions.

关 键 词：数据中心 能耗 价格策略 排队系统 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]