Previous [ 1] [ 2] [ 3] [ 4] [ 5] [ 6] [ 7] [ 8] [ 9] [ 10]


Journal of Information Science and Engineering, Vol. 28 No. 6, pp. 999-1018 (November 2012)

ReSSD: A Software Layer for Improving the Small Random Write Performance of SSDs*

1Computer Science Department
Korea Advanced Institute of Science and Technology
Daejeon, 305-701 Korea
2School of Information and Communication Engineering
Sungkyunkwan University
Suwon, 440-760 Korea

Recently, NAND flash-based solid state drives (SSDs) have emerged as revolutionary storage media. Numerous studies have been carried out to employ SSDs in database systems and storage systems, motivated by SSDs attractive features such as decreased drive weight, increased shock resistance, low power consumption, and no seek latency. However, low-end SSDs targeting desktop and notebook environments show problematic random write performance which is only comparable to or lower than that of HDDs This paper proposes a novel software layer called ReSSD whose purpose is to improve the small random write performance of low-end SSDs with low memory usage. ReSSD works as a virtual block device on top of SSD which requires no modification neither in the operating systems nor in the applications. By inspecting all incoming requests, ReSSD identifies small random writes which have potential to degrade SSDs performance significantly and transforms them into sequential and ordered-sequential writes which are more favorable to SSDs. Our evaluation results through Postmark and OLTP benchmarks show that the proposed approach accomplishes noticeable performance improvement on low-end SSDs under all workloads.

Keywords: solid state drives (SSDs), NAND flash memory, small random write, virtual block device

Full Text () Retrieve PDF document (201211_01.pdf)

Received May 31, 2011; accepted March 31, 2012.
Communicated by Junyoung Heo and Tei-Wei Kuo.
* This work was supported by Next-Generation Information Computing Development Program (No. 2011- 0020520) and by Mid-career Researcher Program (No. 2011-0027613) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology. This work was also partly supported by the IT R&D program of MKE/KEIT (KI10041244, SmartTV 2.0 Software Platform).