Fast responsive flash translation layer for smart devices

Abstract

The flash translation layer design of smart devices differs from that of traditional storages, because smart devices have frequent power-offs and limited resources. These unique features of smart devices generate inconsistent response times and performance degradation. This paper proposes a new FTL algorithm called the Fast Responsive Flash Translation Layer (FFTL) for smart devices. FFTL avoids frequent merge operations by first executing a cleaning policy and provides a faster response time than superblocklevel associativity algorithms by balancing the associativity of each log block. Furthermore, FFTL utilizes all the sectors within each page, thereby reducing the number of write operations. According to the experimental results, it outperforms MCSplit, Superblock, and MAST in terms of the average response time and number of operations.