期刊文献+

分布式文件系统存储介质评测与分析 被引量:9

Performance Testing and Analysis to Storage Medium of Distributed File System
下载PDF
导出
摘要 文中分析了磁盘、SSD和DRAM的价格和能耗,测试了分布式文件系统HDFS、MooseFS和CarrierFS部署在不同存储介质上的数据吞吐率,接着提出了一种存储介质的性价比理论模型,并分析了当前、2015年和2020年不同存储介质的性价比和变化趋势.通过性价比分析可得,DRAM具有较大的性价比优势,适合于高并发随机访问应用.随着DRAM容量的快速提升,未来DRAM有可能成为可选的数据存储介质.当前,SSD虽然具有较好的访问性能,但其价格高、容量小、性价比低于磁盘.2015年SSD的性价比将超过磁盘,有可能替代磁盘成为大规模存储系统的基本存储设备. This paper investigates the price and power consumption of the disks, SSD, and DRAM, and also tested the read and write throughputs of HDFS, MooseFS, and CarrierFS based on these storage mediums. Then, it proposes an estimation model of performance-price ratio for these storage mediums and also predicted and the change trends of their performance-price ratio in 2015 and 2020. From the analysis results, we can see that DRAM has the advantage of high performance-price ratio and are suitable for highly parallel random access applications. As its ca- pacity grows, it will become an optional device for storage systems in the near future. Although SSD can provide better access performance than hard disks, it has lower performance-price ratio due to its higher price and smaller capacity. However, the authors predict that in 2015, its per- formance-price ratio will become higher than hard disks and thus may substitute for hard disks to become the major device for large-scale storage systems.
出处 《计算机学报》 EI CSCD 北大核心 2010年第10期1873-1880,共8页 Chinese Journal of Computers
基金 国家自然科学基金(60803121 60773145 90812001 60963005) 国家"八六三"高技术研究发展计划项目基金(2010AA012401)资助~~
关键词 分布式文件系统 吞吐率 存储介质 distributed file system throughput storage medium
  • 相关文献

参考文献12

  • 1Ghemawat S, Gobioff H, Leung S T. The google file system//Proceedings of the 19th ACM Symposium on Operating Systems Principles. Sagamore, 2003:29-43.
  • 2Caulfield A M, Grupp L M, Swanson S. Gordon: Using flash memory to build fast, power-efficient clusters for dataintensive applications//Proceedings of the International Con ference on Architectural Support for Programming Languages and Operating Systems. Wangington, 2008:217-228.
  • 3Andersen D G, Franklin J, Kaminsky M. FAWN: A fast ar ray of wimpy nodes//Proceedings of the 22nd ACM Symposi um on Operating Systems Principles. Big Sky, 2009:1-14.
  • 4Outerhout J, Agrawal P, Erickson D et al. The case for RAMClouds: Scalable high-performance storage entirely in DRAM. Operating Systems Review, 2009, 43(4) 92-105.
  • 5Schmidt K, Ou Y, Harder T. The promise of solid state disks: Increasing efficiency and reducing cost of DBMS processing//Proceedings of the Canadian Conference on Comput er Science & Software Engineering. Montreal, 2009. 35-41.
  • 6Polte M, Simsa J, Gibson G. Comparing performance of solid state devices and mechanical disks//Proceedings of the 3rd Petascale Data Storage Workshop Held in Conjunction with Supercomputing. Pittsburgh, 2008:1-7.
  • 7Narayanan D, Thereska Eno, Donnelly A et al. Migrating server storage to SSDs: Analysis of tradeoffs//Proceedings of the 4th ACM European Conference on Computer Systems (EuroSys'09). Nuremberg, 2009. 145-158.
  • 8Anderson E, Spence S, Swaminathan R et al. Quickly finding near-optimal storage designs. ACM Transactions on Computer System, 2005, 23(4) : 337- 374.
  • 9Strunk J, Thereska E, Faloutsos C, Ganger G. Using utility to provision storage systems//Proceedings of the USENIX Conference on File and Storage Technologies. San Jose, 2008, 313- 328.
  • 10Kryder M H, Kim C S. After hard drives-What comes next? IEEE Transactions on Magnetics, 2009, 45 (10): 3406-3413.

同被引文献66

引证文献9

二级引证文献19

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