基于Kubernetes的气象私有云自动化部署与优化  

Automatic deployment and optimization of meteorological private cloud based on Kubernetes

作　　者：关兴民[1] 张兆伟 王祝先[1] 王子洋[1] GUAN Xing-min;ZHANG Zhao-wei;WANG Zhu-xian;WANG Zi-yang(Heilongjiang Provincial Meteorological Data Center, Harbin 150000,China)

机构地区：[1]黑龙江省气象数据中心,哈尔滨150000

出　　处：《信息技术》2019年第5期76-80,共5页Information Technology

基　　金：黑龙江省气象局重点项目(HQZD2018001)

摘　　要：容器技术日趋成熟,在应对微服务和系统集约化方面具有巨大优势。基于容器技术,采用谷歌开源的Kubernetes,对私有云平台架构中各个组件进行分析,通过开发一键式部署安装脚本,搭建了基于容器的气象私有云平台。这种自动化部署方式,简化了部署过程、提高了工作效率,降低了手动部署过程中误操作的几率。采用IPVS代替集群默认的Iptables,通过CPU和内存使用效率、新增规则延时、新增规则网络带宽消耗四个方面的对比情况,得出了IPVS作为负载均衡器优化了集群的网络性能。Container technology is mature and has great advantages in coping with micro service and system intensification. Based on container technology,this paper uses Google’s open source Kubernetes to analyze the components of the private cloud platform architecture,and builds a container-based private cloud platform by developing a one-click deployment and installation script. This automatic deployment method simplifies the deployment process,improves work efficiency,and reduces the probability of misoperation in manual deployment process. IPVS is used to replace the default iptables of the cluster.By comparing CPU with memory efficiency, new rule delay and new rule network bandwidth consumption,it is concluded that IPVS as load balancer optimizes the network performance of the cluster.

关 键 词：云计算 自动化部署 Kubernetes DOCKER 负载均衡 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]