移动边缘计算中基于用户移动的虚拟机迁移策略研究

Research on virtual machines migration strategy based on mobile user mobility in mobile edge computing

随着移动边缘计算(mobile edge computing,MEC)技术的深入发展,其面临的挑战也越来越复杂。由于用户移动而造成用户与云服务器之间的通信延时增加,用户服务质量(quality of service,Qo S)下降是目前的研究热点,由此提出移动边缘计算中基于网络拥塞情况下的虚拟机迁移策略。将用户运行的带宽延迟敏感型应用卸载到用户周围代价最小的边缘服务器的虚拟机上,当出现网络拥塞情况时,通过将运行带宽延迟敏感型应用的虚拟机从拥塞的边缘服务器迁移到资源丰富的服务器上,达到提高链路的TCP(transmission control protocol)吞吐量的目的,同时改善用户的Qo S。提出了4种不同的网络与边缘节点拥塞情况,通过仿真结果分析,所提出的虚拟机迁移策略可以有效地改善网络的TCP吞吐量,提高用户服务质量。

With the further development of mobile edge computing(MEC)technology,the challenges are becoming increasingly complicated.The movement of mobile users increases the communication delay between users and cloud servers and hence decreases the users’quality of service(Qo S).It is the current research hotspot.In this paper,based on network congestion,a virtual machine migration technique in mobile edge computing is proposed.Firstly,the bandwidth and delay sensitive application running by the user is offloaded to a virtual machine on the least costly edge server around the user,and then when network congestion occurs,the virtual machine running bandwidth delay sensitive application is migrated from the congested edge server to the resource-rich server,the TCP throughput of the link can be enhanced and the Qo S of the user can be improved.We designed four different congestion cases of network and edge nodes.The simulation results show that the proposed virtual machine migration strategy can effectively improve the TCP throughput of the network as wellas the Qo S of users.