基于对等监控网络的云监控系统设计与实现

发布时间：2018-01-21 02:40

本文关键词： 云监控对等监控网络云平台分布式存储集群　出处：《计算机应用》2015年12期 　论文类型：期刊论文

【摘要】：为解决传统方案中单节点带来的性能瓶颈和低可靠性问题,基于对等监控网络设计并实现了云监控系统解决方案。在硬件部署上,该解决方案将监控节点封装在应用容器中,分布式部署在不同机架上,组建对等的监控网络;监控节点间采用非关系型数据库构建分布式存储集群,实现了监控数据的异地访问和备份。在软件实现上,该解决方案进行了分层设计,采用推拉结合的方式收集数据,对采集的数据进行可信度评估和分布式存储,使用阈值控制和主机空闲评估相结合的策略对云中主机进行智能化管控。通过系统测试,发现该监控系统对计算资源的平均占用率仅有2.17%;而1 ms内响应读写请求的平均比率达到93%以上,表明该解决方案具有资源消耗率低、高频次读写效率高的性能优势。
[Abstract]:In order to solve the performance bottleneck and low reliability problems caused by single node in traditional solutions, a cloud monitoring system solution based on peer-to-peer monitoring network is designed and implemented. The solution encapsulates the monitoring node in the application container and distributes it on different frames to build a peer-to-peer monitoring network. Monitoring nodes use non-relational database to construct distributed storage cluster, which realizes remote access and backup of monitoring data. In software implementation, the solution is designed in a hierarchical way. The data is collected by the combination of push and pull, and the reliability of the collected data is evaluated and distributed storage is carried out. The strategy of combining threshold control and host idle evaluation is used to intelligently control the host in the cloud. Through the system test, it is found that the average utilization rate of the monitoring system for computing resources is only 2.17; The average rate of response to read / write request in 1 Ms is over 93%, which indicates that the solution has the advantages of low resource consumption and high frequency reading and writing efficiency.
【作者单位】：空军工程大学信息与导航学院;
【分类号】：TP393.06
【正文快照】： 0引言随着云计算技术的迅猛发展,针对大规模云集群进行及时有效的管理变得越来越重要。云监控系统,是针对云资源进行动态监控的有效工具。该系统主要由监测和管控两大部分构成。数据监测一般采用分别在监控节点和被监控节点中运行守护进程的方式加以实现[1]。监测的数据包括硬

【参考文献】