摘要
为了解决现有云存储监测方法无法获得完整的系统特性,以确定最佳应用场景并定位性能瓶颈,根据云存储系统的分层架构,调查研究了云存储系统层上的性能监测和采集方法,并提出了一种针对云存储系统层进行分层性能监测和采集的框架。该框架可以获得云存储系统各个系统层次的性能数据,并做进一步的综合对比分析,确定系统的应用场景并定位系统瓶颈,从而对其进行进一步优化。最后在ceph云存储系统上进行了实验,验证了新方法的可用性。
In order to solve the problem that existing cloud storage monitoring methods can＇t obtain the whole system characters to find the best application scenario or perform failure analysis,this paper reviewed the models that used to monitor and gather performance data on system layers of cloud storage system,and proposed a frmework which can evaluate the whole system by gathering and analyzing performance information of main layers in cloud storage according to it＇s layered architecture. This framework can gather performance data of system layers to do further analysis,determine the best application scenario and locate system bottlenecks,then provide some optimized advises to improve the system. In the end,an experiment was conducted on the ceph cloud storage system using this method,the result verified the availability of proposed method.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2016年第3期529-535,共7页
Journal of Northwestern Polytechnical University
基金
国家"863"重大项目(2013AA01A215)
自然科学基金面上项目(61472323)
西北工业大学基础研究基金(3102015JSJ0009)
高效能服务器和存储技术国家重点实验室开放基金(2014HSSA11)资助
关键词
云存储
分层架构
性能监测
数据采集
cloud storage
performance evaluation
monitoring model
failure analysis
