基于负载优化的虚拟机放置方法  

The method of virtual machine placement based on workload optimization

作　　者：闫健恩[1] 张宏莉[1] 许海燕[1] YAN Jianen;ZHANG Hongli;XU Haiyan(School of Computer Science and Technology,Harbin Institute of Technology,Harbin 150001,China)

机构地区：[1]哈尔滨工业大学计算机科学与技术学院

出　　处：《智能计算机与应用》2019年第6期178-180,183,共4页Intelligent Computer and Applications

基　　金：国家重点研究计划(2017YFB0803001);国家自然科学基金(6137021561370211)

摘　　要：随着云计算平台运营成本增加,对物力资源的利用率要求也越来越高。虚拟机放置方法是为虚拟机选择合适的物理宿主机方法,其优劣直接影响到宿主机资源利用的高低,进而影响云计算平台的运营成本。本文提出了一个基于负载优化的虚拟机放置方法,利用CPU、MEM和硬盘的占用率作为资源负载的指标,描述系统的负载情况,采用"称重"的方式,计算宿主机的资源利用情况,从而达到资源利用的最大化。实验表明,相较于OpenStack自带的虚拟机调度方法,该方法在资源利用率方面有明显提升。The popularization and application of cloud computing make virtualization technology develop rapidly. With the increase of operation cost of cloud computing platform,the needs improving the utilization rate of physical resources is becoming higher. The virtual machine placement is to select the appropriate physical host for the virtual machine. The advantages and disadvantages of the method of virtual machine placement directly affect the utilization of the host resources,and then affect the operation cost of the cloud computing platform. In this paper,a virtual machine placement method,named weight calculating algorithm,based on the workloads optimization is proposed,which uses the occupancy rate of CPU,MEM and hard disk as the index of the workloads,describes the workloads of the physical host to calculate the resource utilization of the host,so as to maximize the utilization of the resources. Experimental results show,compared with OpenStack ’s own scheduling method,the method has significantly improved for resource utilization.

关 键 词：云计算 负载优化 虚拟机放置 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]