云计算环境下网络感知的虚拟机迁移研究
发布时间：2019-01-09 15:47
【摘要】：随着信息技术的高速发展,云计算改变了计算机的服务种类和形式。无论是实体的计算机主机、网络设备,还是抽象的计算、存储、网络等资源,都能随时随地按需请求和付费。云计算开创了一种新的服务模式,它将成千上万的计算机整合在一起为用户提供全方位的服务,包括基础设施、平台和软件,用户只需请求所需服务,按需付费。然而,随着云数据中心规模不断扩大,网络资源负载成倍地增长,且网络资源负载不均。如何通过虚拟机迁移优化网络资源成为了目前研究的热点。本文研究的重点是通过虚拟机迁移降低网络资源负载并实现网络资源负载均衡,迁移包含虚拟机选择和放置两个过程。本文提出了网络流量感知的虚拟机迁移算法,根据各虚拟机之间通信流量的大小对虚拟机配对并排序,通信流量最高的虚拟机对作为待迁移虚拟机,再根据虚拟机迁移开销,可用资源的大小确定目标主机。该算法可以有效降低云数据中心的网络流量。本文还针对网络资源负载不均提出了网络资源分层均衡的虚拟机迁移算法,将云数据中心分为多个区域,利用虚拟机迁移均衡各区域内的网络资源负载,最终实现云数据中心的负载均衡,该算法每次选择带宽利用率最高或最低的虚拟机,同时选择网络资源负载最低或最高的目标主机。最后,利用OpenStack平台的云主机移植和资源管理功能设计了一套虚拟机迁移模块并成功实现了本文提出的网络资源分层均衡的虚拟机迁移算法。
[Abstract]:With the rapid development of information technology, cloud computing has changed the types and forms of computer services. Whether it is a computer host, a network device, or an abstract computing, storage, network, and other resources, can anytime, anywhere on demand and payment. Cloud computing creates a new service model that brings together thousands of computers to provide users with a full range of services, including infrastructure, platforms and software. However, with the expansion of cloud data center, the network resource load increases exponentially, and the network resource load is uneven. How to optimize network resources through virtual machine migration has become a hot topic. The emphasis of this paper is to reduce the network resource load and realize the network resource load balance through virtual machine migration, which includes two processes: virtual machine selection and placement. This paper proposes a virtual machine migration algorithm based on network traffic awareness. The virtual machine is paired and sorted according to the traffic of each virtual machine. The virtual machine pair with the highest traffic is used as the virtual machine to be migrated, and then according to the migration overhead of the virtual machine. The size of the available resources determines the target host. This algorithm can effectively reduce the network traffic of cloud data center. This paper also proposes a hierarchical migration algorithm for network resources, which divides the cloud data center into multiple regions and uses virtual machines to balance the load of network resources in each region. Finally the load balancing of cloud data center is realized. Each time the virtual machine with the highest bandwidth utilization or the lowest bandwidth utilization is selected and the target host with the lowest or highest network resource load is selected at the same time. Finally, a virtual machine migration module is designed by using the cloud host migration and resource management functions of OpenStack platform, and the virtual machine migration algorithm proposed in this paper is successfully implemented.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP302;TP393.09
本文编号：2405825
[Abstract]:With the rapid development of information technology, cloud computing has changed the types and forms of computer services. Whether it is a computer host, a network device, or an abstract computing, storage, network, and other resources, can anytime, anywhere on demand and payment. Cloud computing creates a new service model that brings together thousands of computers to provide users with a full range of services, including infrastructure, platforms and software. However, with the expansion of cloud data center, the network resource load increases exponentially, and the network resource load is uneven. How to optimize network resources through virtual machine migration has become a hot topic. The emphasis of this paper is to reduce the network resource load and realize the network resource load balance through virtual machine migration, which includes two processes: virtual machine selection and placement. This paper proposes a virtual machine migration algorithm based on network traffic awareness. The virtual machine is paired and sorted according to the traffic of each virtual machine. The virtual machine pair with the highest traffic is used as the virtual machine to be migrated, and then according to the migration overhead of the virtual machine. The size of the available resources determines the target host. This algorithm can effectively reduce the network traffic of cloud data center. This paper also proposes a hierarchical migration algorithm for network resources, which divides the cloud data center into multiple regions and uses virtual machines to balance the load of network resources in each region. Finally the load balancing of cloud data center is realized. Each time the virtual machine with the highest bandwidth utilization or the lowest bandwidth utilization is selected and the target host with the lowest or highest network resource load is selected at the same time. Finally, a virtual machine migration module is designed by using the cloud host migration and resource management functions of OpenStack platform, and the virtual machine migration algorithm proposed in this paper is successfully implemented.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP302;TP393.09
【参考文献】
相关期刊论文 前6条
1 朱百合;;云计算的特点与关键技术及其在物联网中的应用[J];电子制作;2013年10期
2 郎为民;杨德鹏;李虎生;;中国云计算发展现状研究[J];电信快报;2011年10期
3 龙文;;云时代数字图书馆的抉择——公有云、私有云与传统数据中心[J];中国教育信息化;2011年19期
4 罗军舟;金嘉晖;宋爱波;东方;;云计算:体系架构与关键技术[J];通信学报;2011年07期
5 孙昱;李小勇;管海兵;;虚拟机实时迁移技术研究[J];微型电脑应用;2008年07期
6 薛海峰;卿斯汉;张焕国;;XEN虚拟机分析[J];系统仿真学报;2007年23期
，本文编号：2405825
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2405825.html
