Fast Searching of Log Area in FAST FTL Using Log Area Indexing

Abstract

As flash memory is widely used for embedded devices, research on flash translation layer (FTL), a system software that reflects the pros and cons of flash memory, is steadily studied. FAST FTL, which stores the latest data in the form of a log, searches the log blocks to determine whether the data is updated. However, due to the low percentage of the latest data, it is inefficient to determine whether data is updated by exploring log blocks for all data. Therefore, this paper suggests a way to index information about log blocks to determine if data is up-to-date. The proposed method can improve read performance by indexing information about log blocks into a Binary Search Tree (BST) to quickly search the location of data that has been updated and not having to search log blocks.