###
DOI:
计算机系统应用英文版:2016,25(2):244-249
本文二维码信息
码上扫一扫!
面向云计算的集群文件系统性能对比测试
(1.解放军第401医院信息科, 青岛 266071;2.山东乾云启创信息科技有限公司, 济南 250101;3.石家庄学院 计算科学和工程系, 石家庄 050035;4.中国科学院 软件研究所, 北京 100190)
Performance Comparison Between Two Kinds of Clustered File System in Cloud Computing
(1.The 401 hosptial of PLA, Qingdao 266071, China;2.Shandong Massclouds CO., LTD, Jinan 250101, China;3.Computer Science Engineering, Shijiazhuang University, Shijiazhuang 050035, China;4.Institue of Software, Chinese Academy of Sciences, Beijing 100190, China)
摘要
图/表
参考文献
相似文献
本文已被:浏览 1840次   下载 3568
Received:May 26, 2015    Revised:July 02, 2015
中文摘要: 虚拟机使用集群文件系统是云计算主要架构模式之一,主要分为共享式集群文件系统和分布式集群文件系统两大类.其中,后者是近年来学术界关注的热点,在产业界也逐渐得到深入应用.采用测试法,选取fio作为测试工具,以IOPS作为度量指标,对比两类集群文件系统的性能优劣.实验结果显示,在单物理机节点配置下进行单台虚拟机IOPS写性能测试,共享式文件系统要好于分布式集群文件系统大约40%,但读性能要差1.5倍;对于多台虚拟机并发测试,则当支持虚拟机运行的共享式/分布式集群文件系统规模小于16台时,前者和后者性能基本一致.否则,后者会明显好于前者.
中文关键词: 集群文件系统  IOPS  性能
Abstract:Clustered file system is a key component of cloud computing. It can be divided into two major groups, which are shared-disk file system and distributed file system, in which the latter is the hot spot of academic research in recent years. This paper mainly focuses on IOPS, which is an important performance indicator for VM(virtual machine). We employ fio as the testing tool and experimental results show that,(1) when a VM runs on sharing-disk file system and distributed cluster file system both with 3 storage nodes, the write IOPS of the former is higher than the latter about 40%, but the read IOPS of latter is 1.5 times as much as the former;(2) when multiple VMs run simultaneously on sharing-disk file system and distributed cluster file system both with less than 16 storage nodes, the write IOPS of VMs for both clusters are almost the same; otherwise, the write IOPS performance of the latter is better than the former.
文章编号:     中图分类号:    文献标志码:
基金项目:
引用文本:
刘建平,张辉,于铁忠,吴恒.面向云计算的集群文件系统性能对比测试.计算机系统应用,2016,25(2):244-249
LIU Jian-Ping,ZHANG Hui,YU Tie-Zhong,WU Heng.Performance Comparison Between Two Kinds of Clustered File System in Cloud Computing.COMPUTER SYSTEMS APPLICATIONS,2016,25(2):244-249