一种适用于Docker Swarm集群的调度策略和算法

摘要

Swarm is a cluster management tool based on the Docker container.Aiming at the problem of resource fragment and load balancing of cluster, container scheduling algorithm combining static balance arid dynamic prediction is proposed by analyzing the overall structure and scheduling strategy of Swarm.Through testing, the algorithm can reduce the static resource fragment of the cluster, improve the utilization rate of the cluster resource, and maintain the cluster load balance.%Swarm是基于Docker容器的集群管理工具.通过分析研究Swarm整体架构和调度策略,针对集群的资源碎片问题和负载均衡问题,提出一种静态平衡和动态预测相结合的容器调度算法.通过测试,该算法能够减少集群的静态资源碎片,提高集群资源的利用率,并能维持集群的负载均衡.