云计算虚拟化平台的内存资源全局优化研究

Research on Memory Resources Global Optimization of Cloud Computing Virtualization Platform

借助虚拟化技术,云计算技术能将大规模计算资源统一管理,提高利用效率,但其物理服务器的内存资源边界限制了资源的全局优化能力。为此,对全局优化框架进行改进,增加虚拟机内部资源空闲时的最小内存边界值,基于改进框架,将虚拟机的内存资源分为利用率低和利用率高2种情况,并分别给出2种调节算法及其相互关系。实验结果表明,该方法能降低每次与全局空闲内存池交换的次数,又可降低虚拟机之间的内存交换次数,平均内存资源利用效率得到较大提高。

With the help of virtualization technology, cloud computing will be unified computing resource management to improve the use efficiency,but the combination of virtualization and cloud computing platform brings a new mode of use, especially the physical server resource boundary limits the resources optimization ability. This paper improves the global optimization framework, increases is idle the minimum memory boundary value when internal resources of virtual machine. Based on the framework, the memory resources are divided into virtual machine for the low utilization rate and high utilization rate of the situation, and gives two kinds of control algorithm, as well as the relationship between these two kinds of algorithms. The method not only reduces the number of the exchanging between virtual machines and platform, but also reduces the switching frequency, so average memory resource utilization efficiency can be greatly improved.