Multi-level data redundancy method of large scale FLASH memory array

Abstract

The invention provides a multi-level data redundancy method of large scale FLASH memory array, which includes the steps of: dividing the storage cell of large scale FLASH memory array into three levels: page level, FLASH chip level and FLASH chip group level, according to divided levels as well as required fault-tolerant level, choosing different redundancy encodings for storing into reserved data redundancy spacing in the large scale FLASH memory array; when error is occurred, using redundancy encoding instant correction in the data redundancy spacing; compared with the prior art, the method provided by the invention has the advantages that: against large scale FLASH array application, different redundancy policies are taken for different levels for greatly improving reliability of data storage, and parallel arithmetic capacity of hardware are fully utilized for greatly reducing IO performance reduction induced by adding data redundancy check bit and shortening time for reconstructing damaged data, moreover systematical reliability and redundancy policy can be determined according to practical situation of user with flexible and convenient configuration for realizing purpose of the invention.

Description

technical field [0001] The invention relates to a data reliable storage method of a FLASH memory, in particular to a multi-level data redundancy method suitable for reliable data storage of a large-scale FLASH storage array. Background technique [0002] With the advancement of science and technology, FLASH memory has gradually become the solid-state non-volatile memory with the highest performance-price ratio. In some fields, such as aerospace engineering, there are many tasks that may generate a large amount of data in a short period of time, such as cameras, various real-time sensors, etc.; due to the limited interaction time window with the ground receiving station, these data must be At the same time, because the space environment is very harsh, many factors such as large temperature differences and cosmic rays may cause data damage or loss, which puts high demands on the capacity and reliability of the memory. Require. [0003] Generally, because the application scal...

AI Technical Summary

Problems solved by technology

Therefore, most of the existing research on FLASH memory and its controller focuses on the management of bad blocks in small-scale applications and the study of appropriate check codes, or the use of parallelism and pipeline technology to improve access rates in large-scale applications. , but not enough attention is paid to reliable data storage strategies for large-scale applications

[0004] The Chinese Patent Publication No. CN101320592A discloses a large-capacity FLASH solid storage controller, which solves the problem of bad block management and the verification and error correction of data in the FLASH chip page, but for block damage or FLASH chip damage The resulting data loss issue does not involve

[0005] For another example, in the article "Design of High-Speed ​​and Large-capacity FLASH Storage System" published in Volume 36 of "Fire Control Radar Technology" in March 2007 by Li Chao of Xi'an University of Electronic Science and Technology, he performed in-page cyclic check coding on each FLASH chip and Store the encoding result in the redundant area reserved in the page, and at the same time perform parity encoding on the 16 FLASH chips in the same group and store the encoding result in the 17th FLASH chip in the same group. This data redundancy method is To a certain extent, the reliability of data storage is enhanced, but due to the parallel operation, if two or more FLASH blocks in the same group are damaged or two or more FLASH chips are damaged, this The plan is useless

Images