Performance Evaluation of Data Transfer Protocol GridFTP for Grid Computing

In Grid computing, a data transfer protocol called GridFTP has been widely used for efficiently transferring a large volume of data. Currently, two versions of GridFTP protocols, GridFTP version 1 (GridFTP v1) and GridFTP version 2 (GridFTP v2), have been proposed in the GGF. GridFTP v2 supports several advanced features such as data streaming, dynamic resource allocation, and checksum transfer, by defining a transfer mode called X-block mode. However, in the literature, effectiveness of GridFTP v2 has not been fully investigated. In this paper, we therefore quantitatively evaluate performance of GridFTP v1 and GridFTP v2 using mathematical analysis and simulation experiments. We reveal the performance limitation of GridFTP v1, and quantitatively show effectiveness of GridFTP v2. Through several numerical examples, we show that by utilizing the data streaming feature, the average file transfer time of GridFTP v2 is significantly smaller than that of GridFTP v1.

Authors:

• Hiroyuki Ohsaki
• Makoto Imase

Keywords:

• Grid Computing
• GridFTP
• Performance Evaluation
• Queuing Theory.

References:

[1] H. B. Newman, "Data intensive Grids and networks for high
energy and nuclear physics," InterAct Magazine, Sept. 2002,
also available as http://netlab.caltech.edu/FAST/references/InterAct
NetworksGridsforHEP %hbn091502.doc.
[2] V. Sander et al., "Networking issues for Grid infrastructure," OGF
Document Series GFD.37, Nov. 2004, also available as http://www.ggf.
org/documents/documents/GFD.37.pdf.
[3] J. Postel and J. Reynolds, "File transfer protocol (FTP)," Request for
Comments (RFC) 959, Oct. 1985.
[4] R. Fielding et al., "Hypertext transfer protocol - HTTP/1.1," Request
for Comments (RFC) 2616, June 1999.
[5] W. Allcock et al., "GridFTP: Protocol extensions to FTP for the Grid,"
OGF Document Series GFD.20, Apr. 2003, also available as http://www.
ggf.org/documents/GFD.20.pdf.
[6] I. Mandrichenko, W. Allcock, and T. Perelmutov, "GridFTP v2 protocol
description," OGF Document Series GFD.47, May 2005, also available
as http://www.ggf.org/documents/GFD.47.pdf.
[7] T. Ito, H. Ohsaki, and M. Imase, "On parameter tuning of data transfer
protocol GridFTP in wide-area Grid computing," in Proceedings of
Second International Workshop on Networks for Grid Applications
(GridNets 2005), Oct. 2005, pp. 415-421.
[8] ÔÇöÔÇö, "Automatic parameter configuration mechanism for data transfer
protocol GridFTP," in Proceedings of the 2006 International Symposium
on Applications and the Internet (SAINT 2006), Jan. 2006, pp. 32-38.
[9] ÔÇöÔÇö, "GridFTP-APT: Automatic parallelism tuning mechanism for data
transfer protocol GridFTP," in Proceedings of 6th IEEE International
Symposium on Cluster Computing and the Grid (CCGrid2006), May
2006, pp. 454-461.
[10] L. Kleinrock, Queueing systems, volume II: computer applications.
John Wiley & Sons, Inc., 1976.
[11] I. Mandrichenko et al., "GridFTP protocol improvements," OGF Document
Series GFD.21, July 2003, also available as http://www.ggf.org/
documents/GFD.21.pdf.