期刊文献+

分布式存储系统中高并发客户端的设计与实现 被引量:2

Design and Implementation of a Highly Concurrent Client in Distributed Storage Systems
下载PDF
导出
摘要 BlueOcean是基于对象存储技术的大规模分布式存储系统,本文详细描述了该系统中高并发客户端的设计与实现.该高并发客户端基于FUSE(用户态文件系统)框架进行开发,既保证了通用性,也降低了软件开发和维护的难度;该客户端实现了大部分POSIX语义接口,支持现有的绝大部分应用软件在其上透明运行;同时该客户端采用事件驱动编程模型,相对于传统的同步客户端,提升了在多文件并发访问情况下的性能.测试表明,使用本文实现的高并发客户端输出BlueOcean存储系统时,随着并发访问文件数量的增多,系统总体的数据读写性能,元数据操作性能以及事务操作性能均有很显著的提升. BlueOcean system is a large distributed storage system based on object-oriented Storage technology. This paper describes the design and implementation of a highly concurrent client in BlueOcean system. This highly concurrent client is developed based upon the FUSE framework, which guarantees the generality and reduces the difficulty on developing and maintenance. The client implements most POSIX interfaces, and it can support most existing application. In addition, the client is implemented using the eventdriven model. When accessed by multi applications simultaneously, the client improves the performance significantly comparing with the traditional synchronization client. The results of the experiments show that, when the BlueOcean storage system is outputted by the highly concurrent client, the data read/write performance, metadata operations performance and transactions performance of the overall system has a significant performance improvement.
出处 《小型微型计算机系统》 CSCD 北大核心 2014年第1期24-29,共6页 Journal of Chinese Computer Systems
基金 上海市科学技术委员会基金重点项目(10DZ1500200)资助
关键词 分布式存储系统 事件驱动模型 高并发 FuseAgent FUSE distributed storage system event driven model high-concurrency FuseAgent FUSE
  • 相关文献

参考文献1

二级参考文献5

  • 1Sandberg,R.Goldberg,D.Kleiman,S.Walsh,D.andLyon.B.Design and implementation of the Sun networkfilesystem[C].In Proceedings of the Usenix Confer-ence,1985,119–130.
  • 2Sanjay Ghemawat,Howard Gobioff,and Shun-TakLeung.The Google File System[C].Proceedings of the19th ACM Symposium on Operating Systems Principles.New York,2003,29-43.
  • 3Mesnier,M.Ganger,and G.R.Riedel.E.Object-basedStorage[J].Communications Magazine,IEEE,2003,41(8):84–90.
  • 4John Howard,Michael Kazar,Sherri Menees,David Nich-ols,MahadevSatyanarayanan,Robert Sidebotham,andMichael West.Scale and Performance in a distributed fi-lesystem[J].ACM Transactions on Computer Systems(TOCS),1988,volume 6 Issue 1,51–81.
  • 5Aditya Rajgarhia,Ashish Gehani.Performance and ex-tension of user space file systems[C].Proceedings of the2010 ACM Symposium on Applied Computing.NewYork,2010,206-213.

共引文献6

同被引文献34

  • 1张忠平,刘欣媛.网格环境下的静态启发式任务调度算法[J].计算机研究与发展,2008,45(z1):21-25. 被引量:2
  • 2秦金磊,朱有产,李玉凯.基于网格计算的关键技术研究[J].计算机技术与发展,2006,16(11):103-105. 被引量:4
  • 3何琨,赵勇,陈阳.分布式环境下多任务调度问题的分析与求解[J].系统工程理论与实践,2007,27(5):119-125. 被引量:12
  • 4Marshall Kirk McKusick, Sean Quinlan. GFS; Evolution on fastforward [J]. ACMQueue, 2009, 7 (7).. 10-20.
  • 5Mohamad Sindi. Evaluating MPI implementations using HPL on an infiniband Nehalem Linux cluster [C] //Seventh Interna- tional Conference on Information Technology: New Genera- tions, 2010: 19-25.
  • 6Olson M. HAIXX)P: Scalable, flexible data storage and anal- ysis [J]. IQTQuarterly, 2010, 1 (3): 14-18.
  • 7Oracle. LustreTM 1.8 operations manuaL [EB/0L]. http:// wiki. lustre, org/images/0/09/821-0035_vl. 3. pdf, 2010.
  • 8Sun microsystems, lustre file system datasheet [R]. Santa Clara: Sun Microsystems, 2008.
  • 9Swapnil Patil, Garth Gibson. Scale and concurrency of GIGA +: File system directories with millions of files [C] //Pro- ceedings of the 9th USENIX Conference on File and Storage Technologies, 2011.
  • 10Cams P, Lang S, Ross R, et al. Small-file access in parallel file systems [C] //Processdings of the 23rd IEEE International Parallel and Distributed Processing Symposium, 2009.

引证文献2

二级引证文献9

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部