Improving the Small File Synchronous Write Performance for Parallel Filesystem

Article Preview

Abstract:

In parallel filesystem, data for a single file is striped across multiple storage nodes to provide high performance. But the synchronous write performance of the small file is always poor because it is low efficiency to strip the data of small file to multiple storage nodes. So, many parallel filesystem use the uniform distribution method to save whole the small file to on storage node. This storage method of small file leads to the performance bottle of small file synchronous write in parallel file system. This paper presents a high performance automatically cooperative synchronous write mechanism for the small file in parallel filesystem: C-SWM. C-SWM is a set of pure software mechanisms which provide the reliability of synchronous write operations with the performance of asynchronous write operations. The data of asynchronous write in C-SWM are saved on both two cooperative nodes in parallel filesystm before being written to disk. The experiments in this paper prove that the C-SWM can great improve the small file synchronous write performance in parallel filesystem.

You might also be interested in these eBooks

Info:

Periodical:

Advanced Materials Research (Volumes 403-408)

Pages:

2432-2436

Citation:

Online since:

November 2011

Authors:

Export:

Price:

Permissions CCC:

Permissions PLS:

Сopyright:

© 2012 Trans Tech Publications Ltd. All Rights Reserved

Share:

Citation:

[1] M. G. Baker, J. H. Hartman, M. D. Kupfer, K. W. Shirriff, and J. K. Ousterhout, Measurements of a Distributed File System, In Proc. 13-th Symposium on Operating Systems Principles, pages 198-212, October (1991).

DOI: 10.1145/121132.121164

Google Scholar

[2] Juszczak, C., Improving the Write Performance of an NFS Server, USENIX Conference Proceedings, Winter 1994, San Francisco, CA, 247-259.

Google Scholar

[3] M. Baker, S. Asami, E. Deprit, J. Ousterhout, and M. Seltzer, Non-volatile Memory for Fast, Reliable File Systems. In Proc. of the 5-th International Conference onArchitectural Support for Programming Languages and Operating Systems, Boston, MA, October 1992, 10-22.

DOI: 10.1145/143365.143380

Google Scholar

[4] Avila, R.B. Navaux, P.O.A. Lombard, Proceedings of the 16th Symposium on Computer Architecture and High Performance Computing, 2004, 100-105.

Google Scholar