Journal of Computer Applications ›› 2017, Vol. 37 ›› Issue (12): 3386-3390.DOI: 10.11772/j.issn.1001-9081.2017.12.3386

Previous Articles     Next Articles

Virtual machine placement algorithm for heterogeneous cloud environment based on resource demand distribution feature

XUE Hongye, ZHU Tianlei, LUO Xiangyu, FENG Jian   

  1. College of Computer Science and Technology, Xi'an University of Science and Technology, Xi'an Shaanxi 710054, China
  • Received:2017-06-22 Revised:2017-08-23 Online:2017-12-10 Published:2017-12-18
  • Supported by:
    This work is partially supported by the Natural Science Foundation of Shaanxi Province (2017JQ6053), the Scientific Research Program Funded by Shaanxi Provincial Education Department (15JK1468), the Ph. D. Start-up Fund of Xi'an University of Science and Technology (2015QDJ031).

基于资源需求分布特征的异构云环境虚拟机放置算法

薛弘晔, 朱天磊, 罗香玉, 冯健   

  1. 西安科技大学 计算机科学与技术学院, 西安 710054
  • 通讯作者: 罗香玉
  • 作者简介:薛弘晔(1960-),男,陕西扶风人,教授,博士,CCF会员,主要研究方向:网络与高性能处理、图像实时处理;朱天磊(1992-),男,河南商丘人,硕士研究生,主要研究方向:分布式与并行计算;罗香玉(1984-),女,河北宁晋人,讲师,博士,CCF会员,主要研究方向:分布式与并行计算、容错理论;冯健(1973-),女,陕西西安人,副教授,博士,CCF会员,主要研究方向:复杂网络、网络安全。
  • 基金资助:
    陕西省自然科学基础研究计划项目(2017JQ6053);陕西省教育厅专项科学研究计划项目(15JK1468);西安科技大学博士启动基金资助项目(2015QDJ031)。

Abstract: Focusing on the problem of Virtual Machine Placement (VMP) in heterogeneous cloud environment, a Resource Demand Distribution Feature based Placement Algorithm (RDDFPA) for virtual machines was proposed. Firstly, a method of describing virtual machine requirements and physical machine configuration based on scale factor of CPU resource and memory resource was established. Based on the scale factor, all the virtual machines were sorted. Secondly, by analyzing the proportion relationship of virtual machine requirements and physical machine configuration in the CPU resources and memory resources, the proportion demarcation point was determined, and the partition of virtual machine set was completed. The requirement proportion of matched physical machines with different configurations was reflected by the size of each virtual machine subset.Finally, by using the heuristic algorithm such as the First Fit algorithm, the virtual machine subset was placed on the subset of physical machines with matched configuration. Theoretical analysis and simulation experimental results show that, compared with the total number of physical machines with any single configuration, the total number of physical machines required by the proposed algorithm is reduced by 2%-17%.The proposed RDDFPA can determine the number of physical machines with various configurations according to the distribution of virtual machine resource requirements, and efficiently complete the placement of virtual machines, which can improve the resource utilization rate and reduce the system energy consumption.

Key words: cloud computing, data center, Virtual Machine Placement (VMP), heterogeneous cloud environment, energy efficiency

摘要: 针对异构云环境中的虚拟机放置(VMP)问题,提出一种基于虚拟机资源需求分布特征的放置算法(RDDFPA)。首先,建立基于CPU资源和内存资源比例系数的虚拟机需求和物理机配置描述方法,并根据该比例系数对所有虚拟机进行排序;其次,通过分析虚拟机需求与物理机配置各自在CPU资源和内存资源比例方面的关系,确定比例分界点,完成虚拟机集合的划分,每个虚拟机子集合的规模反映出对相匹配的不同配置物理机的需求比例;最后,利用启发式算法如首次适应(First Fit)算法完成虚拟机子集合在相匹配配置的物理机子集合上的放置。理论分析和仿真实验结果表明,与采用任意单一配置的物理机总数量相比,所提算法所需物理机的总台数减少了2%~17%。RDDFPA能够根据虚拟机资源需求分布的不同,确定各类配置物理机的数量,高效完成虚拟机的放置,在提高资源利用率的同时,降低了系统能耗。

关键词: 云计算, 数据中心, 虚拟机放置, 异构云环境, 能源效率

CLC Number: