基于二维汉明码结构的闪存纠错存储系统  

Flash Error Correcting Storage System Using Two Dimensional Structure Hamming Code

作　　者：余辉龙[1] 张健[1] 李清[2] 覃翠[1] 赵静[1] 花涛[1] 

机构地区：[1]南京工程学院,江苏南京211167 [2]中北大学,山西太原030051

出　　处：《微电子学与计算机》2018年第3期1-3,共3页Microelectronics & Computer

基　　金：国家自然科学基金(61501222);山西省自然科学基金(2015011060);南京工程学院校级科研基金(YKJ201322)

摘　　要：针对闪存存储单元会出现多位错误,提出了NAND型闪存存储器的二维汉明码算法,从行和列二维结构上对任意2Jbit字长的存储数据,用统一的公式化算法进行汉明编码,通过计算纠错偶对能够纠错多位数据误码的情况,译码时先纠错,再校验,保证存储系统数据的完整性,解决了MLC结构闪存的纠错问题.在此基础上分析了不能纠错的图样模式.该算法核心结构上只有异或操作,易于实现,适用于NAND型闪存存储器.实验表明,闪存系统中有效数据二维结构为1 024×128时,采用二维汉明码结构,编码效率为11%,实际编码效率为12.6%.其编码速度和译码速度均达到了102Gbps,适用于高速存储系统.NAND flash unit causes multi-bit data error, the two-dimensional Hamming code algorithm for NAND flash memory is proposed. Any two-dimensional length of 2J bit storage data From the row and column view, the uniform Hamming encoding algorithm formula is provided, by calculating error-correcting-pairs, multiple bit data error can be corrected. The error of the MLC structure and the integrity of data storage system is solved by first decoding, and then check decoding. On this basis, the error no-corrected pattern is analyzed. XOR operation which is the core structure of the algorithm is easy to implement and suitable for NAND flash memory. Experimental results show that the effective data in two-dimensional structure 1 024 ＊ 128. Coding efficiency is 11%, actual 12. 6%. Its encoding speed and decoding speed are up to 102 Gbps, suitable for high-speed storage system.

关 键 词：二维汉明码 纠错编码 NAND闪存 纠错偶对 

分 类 号：TP333.5[自动化与计算机技术—计算机系统结构]