Browse > Article
http://dx.doi.org/10.5573/JSTS.2015.15.5.570

Optimizing Fsync Performance with Dynamic Queue Depth Adaptation  

Park, Daejun (College of Information and Communication Engineering, Sungkyunkwan University)
Kim, Min Ji (College of Information and Communication Engineering, Sungkyunkwan University)
Shin, Dongkun (College of Information and Communication Engineering, Sungkyunkwan University)
Publication Information
JSTS:Journal of Semiconductor Technology and Science / v.15, no.5, 2015 , pp. 570-576 More about this Journal
Abstract
Existing flash storage devices such as universal flash storage and solid state disk support command queuing to improve storage I/O bandwidth. Command queuing allows multiple read/write requests to be pending in a device queue. Because multi-channel and multi-way architecture of flash storage devices can handle multiple requests simultaneously, command queuing is an indispensable technique for utilizing parallel architecture. However, command queuing can be harmful to the latency of fsync system call, which is critical to application responsiveness. We propose a dynamic queue depth adaptation technique, which reduces the queue depth if user application is expected to send fsync calls. Experiments show that the proposed technique reduces the fsync latency by 79% on average compared to the original scheme.
Keywords
fsync; NCQ; IO scheduler; SSD; latency; dynamic queue depth adaptation;
Citations & Related Records
연도 인용수 순위
  • Reference
1 Mathur, Avantika, et al. "The new ext4 filesystem: current status and future plans." The Linux Symposium. Vol. 2. 2007.
2 Lu, Lanyue, et al. "Physical disentanglement in a container-based file system." The 11th USENIX conference on Operating Systems Design and Implementation. USENIX Association, 2014.
3 Kang, Yunji, and Dongkun Shin. "Per-block-group journaling for improving fsync response time." Consumer Electronics (ISCE 2014), The 18th IEEE International Symposium on. IEEE, 2014.
4 J. Axboe, "Linux Block IO - Present and Future," The Ottawa Linux Symposium 2004, pp. 51-61, Jul., 2004.
5 Jeong, Daeho, Youngjae Lee, and Jin-Soo Kim. "Boosting Quasi-Asynchronous I/O for Better Responsiveness in Mobile Devices." The Conference on File and Storage Technology (FAST'15), pp. 191-202, Feb., 2003.
6 Dees, Brian. "Native command queuing-advanced performance in desktop storage." Potentials, IEEE 24.4, pp 4-7. 2005.   DOI   ScienceOn
7 Tweedie, Stephen C. "Journaling the Linux ext2fs filesystem." The Fourth Annual Linux Expo. 1998.
8 Chidambaram, Vijay, et al. "Optimistic crash consistency." The Twenty-Fourth ACM Symposium on Operating Systems Principles. ACM, 2013.
9 Axboe, J. FIO - flexible IO tester. http://freshmeat.net/projects/fio/.
10 Jeong, S., Lee, K., Hwang, J., Lee, S., and Won, Y. AndroStep: Android storage performance analysis tool. In Software Engineering (Workshops)'13, pp. 327-340. 2013.