h-Buffer: An Adaptive Buffer Management Scheme for Flash-Based Storage Devices

Due to the limitations of flash memory, such as asymmetric I/O latencies and not-inplace update, two kinds of buffer replacement algorithms exist, namely page-clustered policy and group-clustered policy. The former organizes pages in the page-level, which makes it easy to deal with hot pages, but shows a bad performance when the buffer size is big enough. The latter organizes pages in the group-level, as the RAM size inside SSDs(Solid State Disks) is limited, most of them ignore the read request from the host. However, as the read/write latency for flash memory is about 1:10, and most of desk and server application programs are readincentive, applying a small portion of buffer space for some hot clean pages will benefit most. In this paper, we propose such a buffer management scheme with three lists, which is called hBuffer. Applying less than 7.125% of the buffer size for clean pages, h-Buffer considers both the write and read request. h-Buffer adopts a replacement policy, a write-back policy and a HL(hot list) compensating policy. Unlike some existing algorithms, it does not only consider the recency and frequency of page references, but also interacts to the buffer capacities and FTL timely. Experiment results show that the erase count, read count and run time of h-Buffer decreases 50% over traditional algorithms on average.