A method and device for separating cold and hot data of a SATA solid-state hard drive

Abstract

The present invention relates to a method and device for separating cold and hot data of a SATA solid-state hard drive; wherein, the method includes: S1, defining data streams, numbering data streams, and binding data stream IDs; S2, applying data Bind with different data stream IDs; S3, the application data generates multiple commands and sends them to the solid state disk; S4, caches the data in the corresponding data pool according to the data stream ID of the command; S5, when the SSD cache is full, the Brush the cold data to the first physical block of the flash memory, and write the hot data to the second physical block of the flash memory; S6, the host application layer rewrites the hot data, and writes the hot data in the second physical block to the third physical block of the flash memory , and erase the second physical block. The invention separates the hot and cold data in the cache, and writes the hot and cold data into different physical blocks, thereby reducing the trigger probability of garbage collection tasks, thereby improving the read-write performance of the host and the service life of the SSD.

Description

technical field [0001] The invention relates to the technical field of cold and hot data separation of solid-state hard disks, and more specifically refers to a method and device for separating cold and hot data of SATA solid-state hard disks. Background technique [0002] SSD (solid state drive) with SATA interface and NAND flash memory still occupies a mainstream position in the server and data center business. In the above business model, there may be multiple applications on the host side that write data to the SSD, and the data streams of different applications have different hot and cold properties (refresh frequency), and the data of each application may be split by the host into Multiple commands are sent to the SSD, and the commands of different applications are interleaved and sent to the SSD; the existing firmware technology cannot effectively distinguish the degree of hotness and coldness of these commands (data), and can only be sent to the SSD in the order in w...

AI Technical Summary

Problems solved by technology

In the above business model, there may be multiple applications on the host side that write data to the SSD, and the data streams of different applications have different hot and cold properties (refresh frequency), and the data of each application may be split by the host into Multiple commands are sent to the SSD, and the commands of different applications are interleaved and sent to the SSD; the existing firmware technology cannot effectively distinguish the degree of hotness and coldness of these commands (data), and can only be sent to the SSD in the order in which the host sends commands. The data is first cached in the cache, and then flashed to the physical block of the flash memory, so the hot and cold data will be written to the same physical block with a high probability, and many garbage blocks will be generated as the hot data is frequently refreshed (physical blocks at the same time) Contains valid data and invalid data), frequently triggers garbage collection tasks, which will lead to a sharp decline in SSD performance and life; therefore, cannot meet the demand

Images