云计算系统虚拟机内存资源预留方法  

Memory resources reservation method for virtual machine in cloud computing system

作　　者：阚运奇 刘宏伟[1] 左德承[1] 张展[1] KAN Yunqi;ZUO Decheng;ZHANG Zhan;LIU Hongwei(School of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China;School of Information Engineering, Northeast Dianli University, Jilin 132400 , China)

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,黑龙江哈尔滨150001 [2]东北电力大学信息工程学院,吉林吉林132400

出　　处：《国防科技大学学报》2016年第5期45-51,共7页Journal of National University of Defense Technology

基　　金：国家863计划资助项目(2013AA01A215)

摘　　要：为降低消费者租借云计算系统资源的开销,提出了成本约束的内存预留随机整数线性规划模型及方法。结合预留计划和按需计划的内存资源价格,设计包含成本及资源总量约束条件的随机开销函数,并以函数期望值最小化为目标,基于内存消耗量概率分布求出优化的内存预留量。试验表明,消费者利用该方法租借资源的开销比利用预留计划、按需计划及同类方法租借资源的开销更小。In order to reduce the consumers' costs of renting resources in cloud computing system, a random integer linear programming model and a method for memory reservation were presented on the basis of cost constraints. Combined with the memory resource price of the reservation plan and the on-demand p lan, the random cost function which consists of costs and total amount of resources constraints was designed. Aimed at minimizing the expected value of the cost function, the optimal amount of memory reserved was obtained on the basis of the probability distribution of memory consumption. The experiments show that the cost of renting recourses by utilizing this method is less than the cost of renting recourses by adopting reservation plan, on-demand plan and other similar methods.

关 键 词：云计算 虚拟机 内存资源预留 成本优化 

分 类 号：TN95[电子电信—信号与信息处理]