J4

• 计算机科学 • 上一篇    下一篇

一个综合性集群监测模型MCM的设计与实现

秦海波1, 魏晓辉1, 李博1, 袁曙涛2   

  1. 1. 吉林大学 计算机科学与技术学院, 长春 130012; 2. 平台计算公司, 多伦多L3R 3T7, 加拿大
  • 收稿日期:2007-04-29 修回日期:1900-01-01 出版日期:2008-01-26 发布日期:2008-01-26
  • 通讯作者: 魏晓辉

Design and Implementation of MCM a Comprehensive Modelof Cluster Monitoring

QIN Haibo1, WEI Xiaohui1, LI Bo1, YUAN Shutao2   

  1. 1. College of Computer Science and Technology, Jilin University, Changchun 130012, China; 2. Platform Computing Inc, Toronto L3R 3T7, Canada
  • Received:2007-04-29 Revised:1900-01-01 Online:2008-01-26 Published:2008-01-26
  • Contact: WEI Xiaohui

摘要: 根据已有的网络监测技术, 提出一个集群系统监测模型MCM. MCM将每个监测任务交给一个监测模块, 并可以灵活地加入和删除这些监测模块, 这种设计使得MCM可以有效地支持对分布式计算资源、 服务以及异常事件的监测. MCM为集群资源管理, 跨域并行作业, 网格资源协同分配和元调度算法提供了资源监测基础设施. 最后, 基于MCM和Platform公司的集群产品EGO, 实现了一个高效的综合性集群监测系统.

关键词: 集群, 分布式, 资源管理, 监测

Abstract: On the basis of existing network monitoring techniques, we presented a model of cluster monitoring called MCM. MCM resorts diverse monitoring jobs to a set of monitoring modules. MCM can add/remove the modules flexibly, and thus can monitor various resources, services and events efficiently. MCM constructs resources monitoring infrastructure for cluster resources management, cross domain parallel jobs, grid resourcescoallocation and metaschedule algorithm. Based on MCM and Platform Computing Corporation’s production EGO, we implemented an efficient comprehensive cluster monitoring system.

Key words: cluster, distributed, resources management, monitoring 

中图分类号: 

  • TP391