一种云计算环境下的虚拟机动态迁移策略

Dynamic Migration Strategy of Virtual Machine in Cloud computing environments

摘要:

提出一种云计算环境下的虚拟机动态迁移策略DMS-VM(Dynamic Migration Strategy for Virtual Machine) 。首先,假设了一种云计算环境下虚拟机迁移场景,在该场景下多种应用服务请求处于动态变化之中,并且提供的应用服务与虚拟机是一对一绑定的；其次,提出了该场景下的基于多约束的多目标规划模型,并设计遗传算法作为虚拟机的主要迁移策略；最后利用某个企业的大数据中心作为云端测试环境, 对比测试DMS-VM算法与已有的虚拟机迁移算法的性能。实验结果表明,DMS-VM迁移策略能够更好的减少物理主机的使用数量和虚拟机的迁移次数,同时降低数据中心能耗,性能优于已有的迁移策略。

Abstract:

A Dynamic Migration Strategy of Virtual Machine in Cloud computing environments called DMS-VM was proposed and discussed in this paper.First, a virtual machine migration scenario is assumed under the cloud computing environment. In this situation, multiple application service requests are dynamically changed, and the application services provided with the virtual machine are one to one. Secondly, a multi-objective programming model based on multiple constraints is proposed, and a genetic algorithm is designed as the main migration strategy of the virtual machine. In order to evaluate the performance of DMS-VM,a serial of test had been done.The experimental results show that DMS-VM can effectively reduce the number of physical machine usage and the number of migration virtual machines conpared with the common strategy. The energy consumption of data center had also obtained in our test.