Kubernetes HPA技术在气象通信系统中的应用  

Application of kubernetes hpa technology in meteological communication system

作　　者：张兆伟 冯鹏[1] 董洋[1] 杨英奎 ZHANG Zhao-wei;FENG Peng;DONG Yang;YANG Ying-kui(Heilongjiang Provincial Meteorological Data Center,Harbin 150000,China)

机构地区：[1]黑龙江省气象数据中心

出　　处：《信息技术》2020年第3期39-43,共5页Information Technology

基　　金：黑龙江省气象局重点项目(HQZD2018001)

摘　　要：为提高气象数据的传输时效,在气象通信系统中引入了Kubernetes水平自动伸缩技术(HPA),并分析了基于CPU和内存指标的应用自动伸缩流程。针对气象观测站周期性传输大量小尺寸数据的特点,提出了一种根据文件或者消息数量指标对Pod进行自动伸缩的改进算法。通过对FTP协议和消息模式的传输时效进行实验分析,得出了文中提出的HPA算法具有较高的通信时效和资源利用率,能够满足现代化气象业务的要求。In order to improve the transmission efficiency of meteorological data,Kubernetes Horizontal Auto-Scaling(HPA)technology is introduced into meteorological communication system,and the appli-cation auto-scaling process based on CPU and memory index is analyzed.According to the characteristics of periodic transmission of large amounts of small-scale data by meteorological observatories,an improved automatic scaling algorithm for Pod based on the number of documents or messages is proposed.Through the experimental analysis of the transmission timeliness of FTP protocol and message mode,it is concluded that the HPA algorithm proposed in this paper has high communication timeliness and resource utilization,and can meet the requirements of modern meteorological services.

关 键 词：气象通信 HPA Kubernetes 消息中间件 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]