摘要
为保障云计算平台的可靠性,基于Ganglia,采用Eucalyptus开源云平台,提出了云计算平台资源监控系统架构.从总体结构、监控信息处理流程等多个方面对监控系统进行了详细阐述,提出了动态更换虚拟机簇头的算法,并对算法与原型系统进行了验证和性能评测.分析与评测结果表明,所提系统能够适用于云计算环境资源信息的实时监测与预警,具备系统开销较低,用户接口友好的特点,能准确反映系统、虚拟机的负载信息,有助于云计算系统服务可靠性的提升.簇头动态更换算法能及时发现失效的簇头并更换为更健壮的虚拟机,保证了监测系统的正常运行.
To guarantee the reliability of the cloud platform, by employing Eucalyptus cloud platform, this paper proposes a resource monitoring model based on Ganglia for loud plat- form. The overall structure, work flow, and the prototype system performance evaluation are expatiated in detail. An algorithm which dynamically replaces cluster head for Ganglia is also proposed. Analysis and evaluation results show that the proposed system can be ap- plied to real-time monitoring and warning of the resource information in a cloud environment with characteristics of low system overhead and user interface friendly, the system and vir- tual machine load information are accurately reflected, and it helps to prompt system service reliability. The proposed algorithm can discover the failing cluster head and replace it with robust virtual machine to guarantee monitoring system normally operating.
出处
《数学的实践与认识》
CSCD
北大核心
2014年第7期210-218,共9页
Mathematics in Practice and Theory
基金
中央高校基本科研业务费专项资金(2012-IV-043)
