Method for implementing parallel-flash translation layer and parallel-flash translation layer system

Abstract

The invention discloses a parallel-flash layer translating method, which comprises the following steps: establishing a mixed address mapping mechanism which makes a page group and a block as an unit, wherein the page group can be divided into a logic page group and a physical page group, the block can be divided into a logic block and a physical block; when the quantity of I/O requests sent by a file system is larger than a predetermined request threshold value, establishing an I/O request queue and putting the I/O requests which are beyond the predetermined request threshold value in the I/O request queue; dividing the I/O requests of the I/O request queue into a plurality of groups of I/O request sub-queues according to the mixed address mapping mechanism; and carrying out parallel reading-writing operation to each group of I/O request sub-queues, wherein each group of the I/O request sub-queues comprise a plurality of I/O requests and the plurality of I/O requests can operate parallelly. The method utilizes excellent parallel-flash feature completely through the parallel reading-writing operation; writing-in and erasing time is reduced; thus, complexity is reduced; and high performance is obtained. The invention also discloses a parallel-flash translation layer system.

Description

technical field [0001] The invention relates to the technical field of computer storage, in particular to a parallel flash conversion layer method and system. Background technique [0002] At present, NAND flash memory is widely used in mobile electronic products, such as mobile phones, MP3 / MP4 players, digital cameras, handheld computers, tablet computers, notebook computers, etc. Flash memory is often used to replace disk drives as a non-volatile secondary storage medium. Compared with disk, flash memory allows random access and has the advantages of low energy consumption, high throughput, small size, shock resistance and portability. The main reason why manufacturers and users of mobile electronic devices pay more and more attention to flash memory is that the capacity of NAND flash memory continues to increase, but its price continues to decrease. However, the long time consumed by NAND flash write and erase operations is the main performance bottleneck of the storage...

AI Technical Summary

Problems solved by technology

Block-mapped flash translation layers can save a lot of memory space for storing mapping information, but their performance is relatively low

Hybrid mapped flash translation layers provide a balance of high performance and low memory requirements, however are often too complex to implement

[0005] The traditional flash translation layer rarely takes into account the parallel operation properties provided by flash memory, and the time consumed by NAND flash write and erase operations restricts the flash memory storage performance of electronic products

Images