ABSTRACT
We describe our investigation of the effect of persistent connections, pipelining and link level document compression on our client and server HTTP implementations. A simple test setup is used to verify HTTP/1.1's design and understand HTTP/1.1 implementation strategies. We present TCP and real time performance data between the libwww robot [27] and both the W3C's Jigsaw [28] and Apache [29] HTTP servers using HTTP/1.0, HTTP/1.1 with persistent connections, HTTP/1.1 with pipelined requests, and HTTP/1.1 with pipelined requests and deflate data compression [22]. We also investigate whether the TCP Nagle algorithm has an effect on HTTP/1.1 performance. While somewhat artificial and possibly overstating the benefits of HTTP/1.1, we believe the tests and results approximate some common behavior seen in browsers. The results confirm that HTTP/1.1 is meeting its major design goals. Our experience has been that implementation details are very important to achieve all of the benefits of HTTP/1.1.For all our tests, a pipelined HTTP/1.1 implementation outperformed HTTP/1.0, even when the HTTP/1.0 implementation used multiple connections in parallel, under all network environments tested. The savings were at least a factor of two, and sometimes as much as a factor of ten, in terms of packets transmitted. Elapsed time improvement is less dramatic, and strongly depends on your network connection.Some data is presented showing further savings possible by changes in Web content, specifically by the use of CSS style sheets [10], and the more compact PNG [20] image representation, both recent recommendations of W3C. Time did not allow full end to end data collection on these cases. The results show that HTTP/1.1 and changes in Web content will have dramatic results in Internet and Web performance as HTTP/1.1 and related technologies deploy over the near future. Universal use of style sheets, even without deployment of HTTP/1.1, would cause a very significant reduction in network traffic.This paper does not investigate further performance and network savings enabled by the improved caching facilities provided by the HTTP/1.1 protocol, or by sophisticated use of range requests.
- 1.Padmanabhan, V. N. and J. Mogul, "Improving HTTP Latency," Computer Networks and ISDN Systems, v.28, pp. 25- 35, Dec. 1995. Slightly Revised Version in Proceedings of the 2nd International WWW Conference '94: Mosaic and the Web, Oct. 1994.]] Google ScholarDigital Library
- 2.Nagle, J., "Congestion Control in IPFfCP Intemetworks," RFC 896, Ford Aerospace and Communications Corporation, January 1984.]] Google ScholarDigital Library
- 3.Bemers-Lee, Tim, R. Fielding, H. Frystyk. "Informational RFC 1945- Hypertext Transfer Protocol-- HTrPI1.0," MIT/LCS, UC Irvine, May 1996.]] Google ScholarDigital Library
- 4.Fielding, R., J. Gettys, J.C. Mogul, H. Frystyk, T. Bemers- Lee, "RFC 2068 - Hypertext Transfer Protocol -- HTrP/1.I," UC Irvine, Digital Equipment Corporation, MIT.]] Google ScholarDigital Library
- 5.Touch, J., J. Heidemann, K. Obraezka, "Analysis of HTTP Performance," USC/Information Sciences Institute, June, 1996.]]Google Scholar
- 6.Spero, S., "Analysis of HTrP Performance Problems," http://www, w3.or~dl:'rotocol~/1.O/HTrPPerformance.html July 1994.]]Google Scholar
- 7.Heidemann, J., "Performance Interactions Between P-HTrP and TCP Implementation," A CM Computer Communication Review, 27 2, 65-73, April 1997.]] Google ScholarDigital Library
- 8.Shepard, T., Source for this very useful program is available at ftp://mercury.lcs.mit.edu/pub/shep. S.M. thesis '~TCP Packet Trace Analysis." The thesis can be ordered from MIT/LCS Publications. Ordering information can be obtained from + 1 617 253 5851 or send mail to [email protected]. Ask for MIT/LCS/TR-494.]] Google ScholarDigital Library
- 9.Mogul, J, 'The Case for Persistent.Connection HTrP", Western Research Laboratory Research Report 95/4, http://www,research.digital.com/wrl/publicationsgabstracts195.4. lltml, Digital Equipment Corporation, May 1995.]]Google Scholar
- 10.Lie, H., B. Bos, "Cascading Style Sheets, level 1," W3C Recommendation, World Wide Web Consortium, 17 Dec 1996.]]Google Scholar
- 11.Jacobson, Van, "Congestion Avoidance and Control." Proceedings of ACM SIGCOMM '88, page 314-329. Stanford, CA, August 1988.]] Google ScholarDigital Library
- 12.Postel, Jon B., "Transmission Control Protocol," RFC 793, Network Information Center, SRI International, September 1981.]]Google Scholar
- 13.Paxson, V., "Growth Trends in Wide-Area TCP Connections," IEEE Network, Vol. 8 No. 4, pp. 8-17, July 1994.]]Google ScholarDigital Library
- 14.Jacobson, V., C. Leres, and S. McCanne, tcpdump, available at flp://ftp.ee.lbl.gov/tcpdump.tar. Z.]]Google Scholar
- 15.Scheifler, R.W., J. Gettys, "The X Window System," ACM Transactions on Graphics # 63, Special Issue on User Interface Software,]] Google ScholarDigital Library
- 16.Manasse, Mark S., and Greg Nelson, "Trestle Reference Manual," Digital Systems Research Center Research Report # 68, December 1991.]]Google Scholar
- 17.Braden, R., "Extending TCP for Transactions -- Concepts," RFC-1379, USC/ISI, November 1992.]] Google ScholarDigital Library
- 18.Braden, R., "T/TCP -- TCP Extensions for Transactions: Functional Specification," RFC-1644, USC~SI, July 1994.]] Google ScholarDigital Library
- 19.Touch, J., "TCP Control Block Interdependence," RFC 2140, USC/ISI, April 1997.]] Google ScholarDigital Library
- 20.Boutell, 7'., T. Lane et. al. "PNG (Portable Network Graphics) Specification," W3C Recommendation, October 1996, RFC 2083, Boutell. Com Inc., January 1997. http://www, w3.org/pub/WW /Graphics/PNG has extensive PNG information.]] Google ScholarDigital Library
- 21.Ryan, M., tcpshow, I.T. NetworX Ltd., 67 Men'ion Square, Dublin 2, ireland, June 1996.]]Google Scholar
- 22.Deutsch, P., "DEFLATE Compressed Data Format Specification version 1.3," RFC 1951, Aladdin Enterprises, May 1996,]] Google ScholarDigital Library
- 23.Deutsch, L. Peter, Jean-Loup Gailly, "ZLIB Compressed Data Format Specification version 3.3," RFC 1950, Aladdin Enterprises, Info-ZIP, May 1996.]] Google ScholarDigital Library
- 24."Recommendation V.42bis (01/90- Data Compression procedures for data circuit terminating equipment (DCE) using error correction procedures," ITU, Geneva, Switzerland, January 1990,]]Google Scholar
- 25.Online summary of results and complete data can be found at http:liwww, w3.orgli:'rotocols/HTrPlPefformancel.]]Google Scholar
- 26.Mogul, Jeffery, Fred Douglis, Anja Feldmann, Balachander Kxishnamurthy, "Potential benefits of delta-encoding and data compression for HTrP," Proceedings of ACM SIGCOMM '97, Cannes France, September 1997.]] Google ScholarDigital Library
- 27.Nielsen, Henrik Frystyk, "Libwww- the W3C Sample Code Library," World Wide Web Consortium, April 1997. Source code is available at http://www.w3.org/Library.]]Google Scholar
- 28.Baird-Smith, Anselm, "Jigsaw: An object oriented server," World Wide Web Consortium, February 1997. Source and other information are available at http:/Avww.w3.org/Jigsaw.]]Google Scholar
- 29.The Apache Group, "The Apache Web Server Project." The Apache Web server is the most common Web server on the Internet at the time of this paper's publication. Full source is available at http://www.apache, org.]]Google Scholar
- 30.A Web page pointing to style sheet information in general can be found at http://www.w3.org/StyleL]]Google Scholar
- 31.Multiple-image Network Graphics Format (MNG), version 19970427. ftp:llswrinde.nde.swri.edulpublmngldocumentsldraftmng-19970427.html.]]Google Scholar
Index Terms
- Network performance effects of HTTP/1.1, CSS1, and PNG
Recommendations
Network performance effects of HTTP/1.1, CSS1, and PNG
We describe our investigation of the effect of persistent connections, pipelining and link level document compression on our client and server HTTP implementations. A simple test setup is used to verify HTTP/1.1's design and understand HTTP/1.1 ...
Improving HTTP performance using "stateless" TCP
NOSSDAV '11: Proceedings of the 21st international workshop on Network and operating systems support for digital audio and videoTCP is quite a heavyweight protocol when serving very small web pages. We introduce a server-side kernel modification which enables a web server to perform HTTP over a UDP socket while the kernel provides a regular TCP interface 'on the wire' to remote ...
Improving HTTP performance in a mobile satellite-terrestrial network
Satellite Network ProtocolsIn recent years, demand for broadband wireless IP networks which support mobility has increased tremendously. Satellite Internet could play a key role, provided that performance problems TCP protocol encounters on mobile satellite channels can be ...
Comments