A method for judging ssd hot and cold data and a garbage collection method

Abstract

The invention discloses a method for judging hot and cold data of an SSD and a garbage collection method, which are characterized in that two levels of data hot and cold judgment of page temperature and block temperature are added at the same time, and a logical page temperature table is added to record the pages of each logical page The temperature value TemLP, the page temperature value is determined by the number of recent write accesses; the block temperature value TemBlock is determined by the page temperature value TemLP of the pages contained in each block, and TemLP is specifically determined by the sum of the page temperature values ​​of valid pages and invalid pages . By introducing this solution, using the write access frequency as the criterion for judging the degree of hotness and coldness of data, the invalid migration of cold data pages during subsequent garbage collection is reduced, the efficiency of garbage collection is improved, write amplification is reduced, and the life of NAND flash is extended.

Description

technical field [0001] The present invention relates to the field of storage technology, in particular to a method for judging hot and cold data of an SSD and a garbage collection method. Background technique [0002] Due to the inherent characteristics of NAND flash, the number of rewritable times of NAND flash is limited, the number of rewritable times of SLC particles is the largest, and the number of rewritable times of TLC particles is the least. The service life of the SSD depends on the service life of the NAND flash. In order to improve the service life of the SSD, a wear leveling strategy is required. The key to wear leveling is the separation of cold and hot data. The current separation of cold and hot data is based on the amount of data written by HOST Judging by the size, this method reduces the system overhead, but it cannot effectively separate the cold and hot data, which will cause a large number of invalid garbage collection GC operations. [0003] Using th...

AI Technical Summary

Problems solved by technology

The disadvantage is that after the data is written, since the number of erasures of the block storing the cold data is relatively small, it is used as the data migration target block during the GC process, and the cold data needs to be re-migrated to increase the write amplification

[0004] figure 1 It is the flow chart of the existing hot and cold data judgment method, separate the hot and cold data, set the hot and cold degree of the data according to the size of the data, put the data with similar temperature together, and reduce the frequency of garbage collection; but it will increase the follow-up garbage collection process Invalid migration of intermediate cold data pages

Images