A log buffer-based flash translation layer using fully-associative sector translation

Abstract

the overall write performance of flash memory. In order to solve this "erase-before-write" problem, the flash memory controller can be integrated with a software module, called "flash translation layer (FTL)." Among many FTL schemes available, the log block buffer scheme is considered to be optimum. With this scheme, a small number of log blocks, a kind of write buffer, can improve the performance of write operations by reducing the number of erase operations. However, this scheme can suffer from low space utilization of log blocks. In this paper, we show that there is much room for performance improvement in the log buffer block scheme, and propose an enhanced log block buffer scheme, called FAST (full associative sector translation). Our FAST scheme improves the space utilization of log blocks using fully-associative sector translations for the log block sectors. We also show empirically that our FAST scheme outperforms the pure log block buffer scheme. . 2007. A log bufferbased flash translation layer using fully-associative sector translation. ACM Trans. Embedd.

Analysis of Performance for NAND Flash Based SSDs via Using Host Semantic Information

“…A number of studies have been conducted at the FTL level. Various FTL schemes improve random write performance by mapping logical addresses to physical addresses [11]- [13]. In addition, performance improvement through hot and cold data classification is also a typical performance improvement technique [13], [14].…”

Crash Recovery in FAST FTL

“…By the granularity of address mapping, FTLs can be largely classified into three types: page mapping FTLs [2], block mapping FTLs [1], and hybrid mapping FTLs including BAST [4] and FAST [5]. In a block mapping FTL, address mapping is done coarsely at the level of block.…”

“…This type of merge operation is called a full merge, and is usually very expensive as it involves performing merge operations as many as valid pages in the victim log block. Please see [5] for details. Besides its performance overhead, we should note that one victim log block in FAST may incur many merges, and a crash can occur at any point while handling them.…”

Power Modeling of Solid State Disk for Dynamic Power Management Policy Design in Embedded Systems

“…In order to overcome the low performance due to "erase before write", log buffers (often called update blocks) are utilized and the new write data is written to the log buffers. The Flash translation layer (FTL) on the SSD controller maintains the address mapping between the logical data and the physical data [15][16] [17]. In reality, the controller, to be specific, FTL, determines the performance and power consumption, especially, of random reads/writes.…”

“…The figure shows that the total power consumption drops again down to the level of the entered low power state. At t 15 , a read command for 8 pages arrives. However, since the SSD is in a low power state, it takes a wakeup time, T wakeup to make a state transition to the active state.…”