Improving MLC Flash Performance with Workload-Aware Differentiated ECC

Abstract

The adoption of small geometries and multi-level cell (MLC) technologies significantly expands the capacity and drops the price of flash memory, which, at the same time, noticeably degrades the performance and reliability of the devices. As incremental-step pulse programming (ISPP) scheme is used to increase the programming accuracy for MLC cells, there is a trade-off between the SSD write performance and raw storage reliability. What's more, ECC is widely used in SSDs to provide error-tolerance ability. Therefore, if we could use stronger ECC to increase error correction strength, a low-cost write with coarser step sizes could be applied in the ISPP scheme to promote the write performance. However, stronger ECC scheme may hurt the read performance due to the increased decoding complexity and latency. In this paper, we propose a workload-aware differentiated ECC scheme to improve the SSD write performance without sacrificing the read performance. The main idea is to dynamically classify the logical pages into three categories: write-only, readonly, and overlapped part. For write-only logical pages, low-cost write with strong ECC scheme will be applied to increase the write performance. For write logical pages in the overlapped part, the low-cost writes with strong ECC will be selectively used based on their relative write and read hotness. While for any read logical pages encoded with a stronger ECC, we will rewrite them with the normal-cost write and ECC scheme if their hotness exceed a pre-defined threshold. The evaluation results show that our workload-aware differentiated ECC scheme could reduce the write and read response times by 48% and 11% on average, respectively. Even compared with the latest previous work, our workload-aware design can still gain about 4% write performance and 11% read performance improvements.

DOI: 10.1109/ICPADS.2016.0078

12 Figures and Tables

Cite this paper

@article{Xia2016ImprovingMF, title={Improving MLC Flash Performance with Workload-Aware Differentiated ECC}, author={Qianbin Xia and Weijun Xiao}, journal={2016 IEEE 22nd International Conference on Parallel and Distributed Systems (ICPADS)}, year={2016}, pages={545-552} }