云存储系统的分区编码冗余方法研究  

A Redundancy Method of Partition Coding Applied to Cloud Storage System

作　　者：谢垂益[1] 鲁向前[1] 卿斯汉[2,3] 

机构地区：[1]韶关学院数学与信息科学学院,广东韶关512005 [2]中国科学院软件研究所,北京100190 [3]中国科学院信息工程研究所,北京100093

出　　处：《信息网络安全》2013年第6期2-6,共5页Netinfo Security

基　　金：国家自然科学基金[6097013561170282];韶关市创新资金[2012-10];韶关学院科研项目[2012-2]

摘　　要：文章根据云存储的结构特点,将数据存储区域划分本地副本区、本地编码区、远端编码区。本地编码区和远端编码区采用确定型网络编码来实现数据的冗余备份。将解码过程的所有系数矩阵进行三角分解并保存为查找表,采用查表方法辅助方程组求解,减少计算量。实验结果表明,当单个存储结点的可用性在0.7至0.99之间、数据可用性目标为99.99999%时,采取数据分片数为8的方案,该方法的平均每字符运算次数低于3次,比副本冗余方法节省50%以上的存储空间.分区编码冗余方法解码速度快,能够显著提高云存储系统的可用性和存储效率.According to the structural characteristics of cloud storage, the data storage area is divided into three parts： local copy zone, local coding zone and remote coding zone, the latter two using a deterministic networkcoding to achieve the purpose of data redundancy. Coefficient matrix of the decoding process triangular decomposition and save the results as a search table that can assist equation solving, thereby reduce the amount of calculation. The experiment results show when the availability of a single storage node in the range of 0.7 to 0.99, data availability target of 99.99999%, taking the data slice number 8, the average operations per character less than 3 times, the method can save more than 50% of storage space than copy redundancy method. The partition coding redundancy method has high decoding speed, improves the availability and storage efficiency of cloud storage system significantly.

关 键 词：云存储 冗余 网络编码 每字符运算次数 

分 类 号：TP333[自动化与计算机技术—计算机系统结构]