Docker容器弹性集群资源调度算法研究

Research on the Elastic Cluster Resource Scheduling Algorithm of Docker Container

对Docker弹性集群的设计和优化进行了研究,通过自顶向下的设计思路,完成了集群的整体设计过程,然后从容器访问的原理出发,联系传统方法中跨主机弹性集群中单个主机在小规模应用中的伸缩方法,提出弹性伸缩组在自动服务中的功能优化,并且为了改善集群的性能,提出资源均衡分配的调度思路,基于此思路建立了弹性调度的设定、计算和处理的详细流程,包括初步的筛选、进一步的优选,以及基于阈值判断与统计学分析的综合处理策略。仿真实验结果表明本文提出的Docker容器弹性集群资源调度算法具有优越的性能。

In this paper, the design and optimization of Docker elastic clusters are studied, through the top-down design ideas, complete the whole design process of the cluster, and then starting from the principle of container to access, contact the traditional method of elastic across the host in the cluster expansion method of a single host in small-scale applications, of elastic expansion set forward in the automatic service function optimization, and to improve the performance of the cluster, puts forward some ideas of balanced allocation of resources scheduling, based on this idea the establishment of a flexible scheduling setting, calculation and processing of the process in detail, including the preliminary screening, further optimization, and based on threshold judgment and statistical analysis of comprehensive treatment strategy. The simulation results show that the proposed Docker container elastic cluster resource scheduling algorithm has superior performance.