开源云计算资源调度策略优化研究  

Research on Optimization of Resource Scheduling in Open Source Cloud Computing

作　　者：邱亚 颜金尧[2] 陈宇 耿俊杰 QIU Ya;YAN Jin-yao;CHEN Yu;GENG Jun-jie(Technology Bureau,China Media Group,Beijing 100020,China;State Key Laboratory of Media Convergence and Communication,Communication University of China,Beijing 100024,China)

机构地区：[1]中央广播电视总台技术局,北京100020 [2]中国传媒大学媒体融合与传播国家重点实验室,北京100024

出　　处：《计算机技术与发展》2023年第6期8-15,共8页Computer Technology and Development

基　　金：国家重点研发计划(2019YFB1804303);国家自然科学基金(61971382);中国传媒大学高校基本科研业务费专项资金(CUC22CGJ02)。

摘　　要：云计算技术是互联网服务模式的一种产物,具有大规模、可扩展、通用性强、可靠性高等特点。针对OpenStack云平台虚拟机调度管理机制进行研究,围绕虚拟机初始分配及运行阶段的资源调度策略存在的问题,对OpenStack云平台虚拟机调度管理策略进行了改进,以提高云计算平台资源利用率,改善物理主机负载均衡,进而保障用户云服务质量。该文优化了计算节点主机过滤机制,同时提出虚拟机运行阶段的动态迁移机制以及待迁移虚拟机的选择机制,保证了虚拟机初始分配的合理性以及运行阶段各计算节点物理主机的负载均衡。使用云计算仿真平台CloudSim对提出的优化策略进行了仿真。实验表明提出的虚拟机调度优化策略作用显著,实现了物理主机的负载均衡、提高了资源利用率,改善了平台服务质量。Cloud computing technology is a product of the Internet service model,which has the characteristics of large scale,scalability,strong versatility and high reliability.We study the virtual machine scheduling management mechanism of the OpenStack cloud platform.In view of the problems existing in the initial allocation of virtual machines and the resource scheduling strategy in the running phase,the virtual machine scheduling management strategy of the OpenStack cloud platform is improved to improve the resource utilization rate of the cloud computing platform,improve the load balancing of physical hosts,and thus ensure the quality of cloud services for users.We optimize the computing node host filtering mechanism and propose a dynamic migration mechanism during the running phase of the virtual machine and a selection mechanism for the virtual machine to be migrated,which ensures the rationality of the initial allocation of virtual machines and the load balance of the physical hosts of each computing node during the running phase.We use the cloud computing simulation platform(CloudSim)to simulate the proposed optimization strategy.Experiments show that the virtual machine scheduling optimization strategy proposed has a significant effect,realizes the load balancing of physical hosts,improves resource utilization,and improves platform service quality.

关 键 词：云计算 OPENSTACK 虚拟机调度 负载均衡 资源利用率 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]