Data classified storage method, device and equipment and storage medium

Abstract

The invention discloses a data classified storage method, device and equipment and a storage medium, and the method comprises the steps: determining the user data needing to be written at present as target data if the user data needs to be written into a flash memory, and determining a classifier needing to be used at present as a target classifier, wherein the classifier is obtained by training data characteristics and categories of the user data written into the flash memory; extracting data features of the target data, and inputting the data features of the target data into a target classifier to obtain a category to which the target data output by the target classifier belongs; and determining a storage area corresponding to the category of the target data in the flash memory, and writing the target data into the determined storage area. When garbage collection is carried out on the user data, the possibility of selecting blocks which are invalid data or most of the blocks are invalid data is relatively high, so that the moving amount of the user data is reduced, the write amplification coefficient is reduced, and the garbage collection efficiency is improved.

Description

technical field [0001] The present invention relates to the technical field of flash memory, and more specifically, to a method, device, device and storage medium for classifying and storing data. Background technique [0002] Flash memory (NAND) technology is constantly improving. Regardless of 2D or 3D flash memory, there is a feature that it can be written according to Page (page, the smallest writing unit in flash memory, and there are multiple Pages in a Block), but it can only be written according to Block (block, the smallest erasing unit in flash memory, there are multiple Blocks in a flash memory) erase. Therefore, when the flash memory is nearly full, some of the multiple pages in a block may be invalid data, and the other part may be valid data. In order to free up more space, it is necessary to move the pages that store valid data in the block to a new block. In , after all valid pages are moved, the Block can be erased; after erasing, the Block can be reused. ...

AI Technical Summary

Problems solved by technology

Therefore, there is a problem of write amplification, that is, the amount of data written by the user is N, but because the data has to be moved internally, the actual amount of data written in the device exceeds N, assuming M, this ratio M / N is the write amplification factor

The inventor found that if the write amplification factor is high, it means that the amount of data to be moved is large, and garbage collection at this time will lead to low efficiency of garbage collection; therefore, it is very necessary to reduce the write amplification factor

Images