云计算系统虚拟机内存资源预留方法

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 plan, 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.