容许多个磁盘故障的RAID编码方法研究  

Research on RAID Coding Schemes for Tolerating Multiple Disk Failures

作　　者：刘军[1] 刘璟[2] 

机构地区：[1]天津财经大学理工学院,天津300222 [2]南开大学信息技术科学学院,天津300071

出　　处：《微电子学与计算机》2011年第8期9-11,共3页Microelectronics & Computer

摘　　要：随着磁盘阵列规模的增大,同时发生多个磁盘故障的概率将大大增加,单容错编码难以满足应用对高可靠性存储的需求.分析了主要的双容错RAID编码方法及其特点,对各种双容错编码方法的冗余性能进行了比较.给出了一种基于循环置换矩阵构建的能容许三个磁盘故障的MDS交换群阵列码,其编码和解码效率较高,是大规模RAID存储系统的应用方向.As the Redundant Arrays of Inexpensive Disks（RAID） scale up,multiple disk failures are likely to take place at the same time.The single erasure code is hard to meet the requirements of high reliable storage.We reviews the coding schemes and features of various MDS array codes for tolerating up to double disk failures and compares their redundancy performances.A class of MDS Abelian group array codes for tolerating up to three disk failures is presented based on circular permutation matrices.Since the encoding and decoding are very efficient,it becomes the trend of application in large-scale RAID storage systems.

关 键 词：RAID编码 MDS阵列码 多磁盘故障 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]