一种基于双容错RDP码的扩展RAID码  

English Title a Extension RAID Code Base on Double Failures RDP Code

作　　者：万武南[1] 杨威[1] 

机构地区：[1]成都信息工程学院信息安全工程学院,成都610226

出　　处：《小型微型计算机系统》2014年第11期2477-2481,共5页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60873216)资助;四川省教育厅重点项目(12ZA223)资助

摘　　要：随着分布式存储系统的发展和在复杂环境中应用,双容错RAID码已经无法满足现有系统的可靠性要求,存储系统的可靠性受到严重的挑战,需要可靠性更高的RAID码.对双容错RDP(Row Diagonal Parity)码进行扩展,提出一种新的3容错扩展RAID码,在有限域GF(2)上,采用类似范德蒙的二元生成矩阵描述阵列码的编码代数定义方法.并通过扩展码的生成矩阵,并提出一种基于矩阵变换的译码算法,与Zig-Zag译码算法相比,此译码算法简单易实现,通用性更高.与现有EEOD码、STAR码和RDDP码进行比较,分析结果表明扩展RAID码的存储效率、编译码复杂度、更新复杂度,以及平均数据失效时间达到了3容错RAID码综合性能达到最优.With the development of the distributed storage system and its application in complex environment, the RAID codes for tol- erating double failures can not meet the requirement of the reliability in current systems. So there is a big challenge and the higher reli- able RAID codes are demanded in the reliability of storage systems. On the basis of the RDP （ Row Diagonal Parity ） code for double toleration failures, a new class of Maximum Distance Separable（ MDS ） array codes for triple failures is presented. An algebraic enco- ding method for array codes is defined by the Vender-monde binary generator matrices over Galois field GF （2 ）. The algorithms of de- coding are directly depicted by matrix transformation, are more easily implemented by software and hardware for simple structure and have more good generality than the Zig-Zag decoding algorithm. The theoretical analysis shows that the comprehensive properties of the proposed code is optimal such as storage efficiency encoding and decoding complexity, update complexity and mean time to data loss by comparing with the EEOD code,the START code and the RDDP code.

关 键 词：编码 纠删码 RDP码 数据容错 分布式存储 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]