January 11, 2010
Garth Goodson and Rahul Iyer
In this paper, we use trace-based simulation to analyze the trade-offs and designs of two popular classes of flash translation layers.
Although flash has better random access characteristics than disk, it has a number of undesirable complexities, including the inability to overwrite a block in place. A layer, called the flash translation layer is used to vitalize flash and presents it as a block device. Although many such designs have been proposed in the past, they often do not focus on the requirements of large-scale enterprise storage, nor do they examine the trade-offs between different approaches. In this paper, we use trace-based simulation to analyze the trade-offs and designs of two popular classes of flash translation layers.
In Proceedings of the HPCA Workshop on the Use of Emerging Storage and Memory Technologies 2010 (WEST ’10)
A copy of the paper is attached to this posting. tradeoffs-west10.pdf