基于稀疏随机矩阵的再生码构造方法

发布时间：2018-05-06 18:35

本文选题：分布式存储可靠性 + 再生码　；参考：《计算机应用》2017年07期

【摘要】：针对已有的再生码编码方案的运算是基于有限域GF(q)、运算复杂度高、效率低的问题,提出了一种将GF(2)上的稀疏随机矩阵和乘积矩阵框架相结合的再生码构造方法。首先,将文件数据矩阵式排布后根据编码矩阵进行行异或运算;其次,节点失效后,参与帮助节点根据失效节点的编码向量编码本地数据并发送至修复节点;最后,修复节点根据接收到的数据译码出失效节点原有的数据。实验结果表明修复带宽至多只有传统纠删码修复方案的1/10,相比基于传统范德蒙编码矩阵的再生码,编码速率提升了70%,译码恢复速率提升了50%,方便了再生码在大规模存储系统中的应用。
[Abstract]:In order to solve the problem that the operation of existing regenerative code coding scheme is based on the finite field GFFQ, which has high computational complexity and low efficiency, a method of constructing reproducing codes by combining the framework of sparse random matrix and product matrix on GF2) is proposed. First, the file data matrix is arranged according to the encoding matrix to perform the XOR operation. Secondly, after the node fails, it helps the node encode local data according to the coding vector of the failed node and send it to the repaired node. The repair node decodes the original data of the invalid node according to the received data. The experimental results show that the restoration bandwidth is only 1 / 10 of that of the traditional erasure code repair scheme, compared with the traditional Van der Mon coding matrix. The coding rate is increased by 70 bit and the decoding recovery rate is increased by 50%, which facilitates the application of the regenerative code in the large-scale storage system.
【作者单位】：中国科学院成都计算机应用研究所;中国科学院大学计算机与控制学院;
【基金】：四川省科技厅支撑计划项目(2015GZ0088)~~
【分类号】：TP333

【参考文献】