期刊文献+

基于分组码的跳跃纠删码

Jump erasure code based on block code
在线阅读 下载PDF
导出
摘要 针对纠删码中分组码多组之间关联性差导致的容错率低等问题,本文基于分组码的思想提出一种跳跃纠删码—跳跃局部重构码。通过将初始数据分组并且在每组选取单块数据跳跃再生成校验块来加强组间联系,以提升容错率、容错能力以及降低重构开销等不同性能,并且可以权衡存储开销与其他性能以满足分布式系统的不同需求。此外,本文通过改变跳跃局部重构码自身参数来对比性能变化,并且与其他常用类型纠删码进行对比实验。结果表明:跳跃局部重构码能够在较小的存储开销下,达到较高的容错能力和较低的重构开销的效果,可跳跃生成校验块并且相对性能较优。 Aiming at the problem of low fault tolerance rate caused by poor correlation between multiple groups of block codes in the erasure code,this paper proposes a jump erasure code called jump local reconstruction code(JLRC)based on the idea of block code.Different performances such as fault tolerance rate,fault tolerance capability,and reconstruction overhead could be improved by grouping the initial data and selecting a single block of data in each group to generate parity blocks.Moreover,JLRC can also balance the storage overhead with other performance to meet different needs of distributed systems.In this study,by varying the parameters of the JLRC code,the performances at different parameters were compared,and the JLRC was compared with other commonly used types of erasure codes.The results show that the JLRC can achieve higher fault tolerance and lower reconstruction overhead with less storage overhead and can,by jumping,generate check blocks with relatively better performance.
作者 孟宇龙 关智允 徐东 张子迎 任龙 MENG Yulong;GUAN Zhiyun;XU Dong;ZHANG Ziying;REN Long(College of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China)
出处 《哈尔滨工程大学学报》 EI CAS CSCD 北大核心 2019年第11期1896-1902,共7页 Journal of Harbin Engineering University
基金 装发重大预研项目(31511030201)
关键词 纠删码 分组码 跳跃局部重构码 存储开销 容错能力 容错率 重构开销 erasure code block code JLRC code storage overhead fault tolerance ability fault tolerance rate reconstruction overhead
作者简介 孟宇龙,男,副教授;通信作者:徐东,男,教授,E-mail:xudong@hrbeu.edu.cn.
  • 相关文献

参考文献7

二级参考文献60

  • 1周松,王意洁.EXPyramid:一种灵活的基于阵列结构的高容错低修复成本编码方案[J].计算机研究与发展,2011,48(S1):30-36. 被引量:5
  • 2魏青松,卢显良,侯孟书.AdpReplica:自适应副本管理机制[J].计算机科学,2004,31(12):34-36. 被引量:4
  • 3KO B J. Scalable service differentiation in a shared storage cache[A].Washington,DC,USA,2003.184-194.
  • 4CHEN Yan. SCAN:a dynamic, scalable, and efficient con-tent distribution network[A].Zürich,Switzer-land,2002.282-286.
  • 5KORUPOLU M R. Placement algorithms for hierarchical co-operative caching[A].PA,USA,1999.586-595.
  • 6KARGER D,LEHMAN E,LEIGHTON T. Consistent hashing and random trees:distributed caching protocols for relieving hot spots on the world wide web[A].CA,USA,1997.654-663.
  • 7BORTHAKUR D. The hadoop distributed file system:archi-tecture and design[EB/OL].http://ha-doop.apache.org/core/docs,2012.
  • 8GUY L,LAURE E,STOCKINGER H. Replica man-agement in data grids,GGF5[R].Global Grid Information Document,2002.
  • 9STONEBRAKER M,ABADI D J,De WITT D J. Ma-pReduce and parallel DBMSs:friends or foes[J].Commu-nication of the ACM,2010,(01):64-71.
  • 10BORDAWEKAR R,LANDHERR S,CAPPS D. Ex-perimental evaluation of the Hewlett-Packard Exemplar file system[A].1997.

共引文献101

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