FTL Translation Layer with RFFS

21 May

Reliable flash file system
Reliable flash memory and NAND flash memory file system (RFFS) for the majority of embedded systems development. FTL purpose is to reduce the installation time and recovery overhead, which occurred in a power failure or unpredictable obstacles. RFFS the flash memory is divided into two areas: the position information area (LIA), maintenance of the sub-region (i.e., meta-data, logs, etc.) and all types of data storage area (DA) of the sub-region, the latest position information.

The location information area is located in the first two sets of flash block, and the first to be scanned during the installation process.
Therefore, RFFS can reduce the installation time by scanning this position area. In In addition, RFFS can build a reliable file system, and by providing transaction management and log management, make a quick recovery in the event of a system failure. Their experimental results show that, RFFS can provide fast installation and recovery. 3.3 hot and cold data identification purposes, the separation of hot and cold data in response to valid data in the clean-up activities, in order to reduce overhead. The thermal data is usually that the access time than others, or the data is cold data. Because hot data is written frequently, thermal data will soon become outdated and invalid. Accordingly, the a mixed hot data and cold data block may contain valid data and invalid data at the same time.
If a block is recovered, also contains valid data, the need for effective data replication to other flash memory space and clean-up activities. This action will result in additional overhead. If a block is full of hot data, these hot data will soon be updated, and thus become obsolete and invalid together. Clean block will produce the lowest overhead of garbage collection. Therefore, there are many studies of hot and cold data separation, in order to reduce system overhead. In this section, we introduce three methods of heat and cold data separation: dynamic data clustering, two LRU list, and register-based data to identify hot and cold.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: