云计算、虚拟化等技术在计算机领域中发展迅速,虚拟化技术中的Docker容器技术是当前主要的研究热点。由于云数据中心的流量负载随时变化,网络可能频繁发生拥塞,造成网络设备资源紧张,导致应用的吞吐量下降、丢包增多和时延增大,影响通信性能和整个云平台服务器的质量。对此,研究在保持物理机不变的情况下通过迁移物理机上挂载的Docker容器来改善网络拥塞,提出一种网络感知的Docker容器重调度算法,通过迁移挂载在物理机上的Docker,来提高物理机通信的能力以达到提高全局的通信效率。算法通过迁移较少的虚拟机来提高物理机之间的通信能力,以提升整个数据中心的整体通信性能。

Cloud computing and virtualization technologies are developing rapidly in the computer field. Docker container technology has become the focus of the research direction. As the traffic load of cloud data center changes at any time, network congestion may frequently occur, resulting in network equipment resource tension, and the decline of application throughput, increase of packet loss and delay, affecting the communication performance and the quality of the entire cloud platform server. Aiming at these problems, this paper improves the network congestion by transferring the Docker container mounted on the physical machine, and a network-aware Docker container rescheduling algorithm is proposed to improve the communication capability of the physical machine to improve the global communication efficiency. The algorithm improves the communication capability between physical machines by migrating fewer virtual machines to improve the overall communication performance of the entire data center.