Improving web performance by client characterization driven server adaptation

ABSTRACT

We categorize the set of clients communicating with a server on the Web based on information that can be determined by the server. The Web server uses the information to direct tailored actions. Users with poor connectivity may choose not to stay at a Web site if it takes a long time to receive a page, even if the Web server at the site is not the bottleneck. Retaining such clients may be of interest to a Web site. Better connected clients can receive enhanced representations of Web pages, such as with higher quality images.We explore a variety of considerations that could be used by a Web server in characterizing a client. Once a client is characterized as poor or rich, the server can deliver altered content, alter how content is delivered, alter policy and caching decisions, or decide when to redirect the client to a mirror site. We also use network-aware client clustering techniques to provide a coarser level of client categorization and use it to categorize subsequent clients from that cluster for which a client-specific categorization is not available.Our results for client characterization and applicable server actions are derived from real, recent, and diverse set of Web server logs. Our experiments demonstrate that a relatively simple characterization policy can classify poor clients such that these clients subsequently make the majority of badly performing requests to a Web server. This policy is also stable in terms of clients staying in the same class for a large portion of the analysis period. Client clustering can significantly help in initially classifying clients for which no previous information about the client is known. We also show that different server actions can be applied to a significant number of request sequences with poor performance.

REFERENCES

• 1: Tarek F. Abdelzaher and Nina Bhatti. Web Server QoS Management by Adaptive Content Delivery. In Proceedings of the International Workshop on Quality of Service, London, England, June 1999. http://www.eecs.umich.edu/~zaher/iwqos99.ps.
• 2: Mark Allman, Measuring end-to-end bulk transfer capacity, Proceedings of the 1st ACM SIGCOMM Workshop on Internet Measurement, November 01-02, 2001, San Francisco, California, USA [doi>10.1145/505202.505220]
• 3: V. Almeida, D. Menasce, R. Riedi, F. Peligrinelli, R. Fonseca, and W. Meira Jr. Analyzing the impact of robots on performance of Web caching systems. In Proceedings of the 6th International Web Caching Workshop and Content Delivery Workshop, Boston, MA, June 2001.
• 4: O. Ardaiz, F. Freitag, and L. Navarro. Estimating the time of service in a Web client starting from the server logs. In Proceedings of the ACM SIGCOMM America Latina Conference, San Jose, Costa Rica, April 2001. ACM.
• 5: Pierre Beyssac. BING: Bandwidth pING, March 1998. http://www.cnam.fr/reseau/bing.html.
• 6: Peter Bickford. Worth the wait? View Source, Human Interface Online, 1999. http://devedge.netscape.com/viewsource/bickford_wait.htm.
• 7: Ramon Caceres, Fred Douglis, Anja Feldmann, Gideon Glass, and Michael Rabinovich. Web proxy caching: the devil is in the details. In Workshop on Internet Server Performance, Madison, Wisconsin USA, June 1998.
• 8: CC/PP Working Group. http://www.w3.org/Mobile/CCPP/.
• 9: Willy Chiu. Best practices for high volume web sites, February 2001. http://www.worldinternetcenter.com/Other_Events/Challenge-The-Expert/Feb28megawebsite/SVWIC_Chiu_2_28_01.pdf.
• 10: Edith Cohen , Balachander Krishnamurthy , Jennifer Rexford, Improving end-to-end performance of the Web using server volumes and proxy filters, Proceedings of the ACM SIGCOMM '98 conference on Applications, technologies, architectures, and protocols for computer communication, p.241-253, August 31-September 04, 1998, Vancouver, British Columbia, Canada
• 11: Allen B. Downey, Using pathchar to estimate Internet link characteristics, Proceedings of the 1999 ACM SIGMETRICS international conference on Measurement and modeling of computer systems, p.222-223, May 01-04, 1999, Atlanta, Georgia, United States
• 12: Balachander Krishnamurphy , Jeffrey C. Mogul , David M. Kristol, Key differences between HTTP/1.0 and HTTP/1.1, Proceeding of the eighth international conference on World Wide Web, p.1737-1751, May 1999, Toronto, Canada
• 13: Balachander Krishnamurthy and Jennifer Rexford. Web Protocols and Practice: HTTP/1.1, Networking Protocols, Caching, and Traffic Measurement. Addison-Wesley, May 2001. ISBN 0-201-710889-0.
• 14: Balachander Krishnamurthy , Jia Wang, On network-aware clustering of Web clients, Proceedings of the conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, p.97-110, August 28-September 01, 2000, Stockholm, Sweden
• 15: Balachander Krishnamurthy and Craig Wills. Study of Piggyback Cache Validation for Proxy Caches in the World Wide Web. In USENIX Symposium on Internet Technology and Systems, pages 1--12, December 1997. http://www.research.att.com/~bala/papers/pcv-usits97.ps.g.
• 16: Balachander Krishnamurthy , Craig E. Wills, Piggyback server invalidation for proxy cache coherency, Proceedings of the seventh international conference on World Wide Web 7, p.185-193, April 1998, Brisbane, Australia
• 17: Robert B. Miller. Response time in man-computer conversational transactions. In Proc. Sprint Joint Computer Conference, Montvale, NJ, 1968. AFIPS Press.
• 18: J. Mogul, B. Krishnamurthy, F. Douglis, A. Feldmann, Y. Goland, A. van Hoff, and D. Hellerstein. Delta encoding in HTTP. RFC 3229, IETF, January 2002. Proposed Standard.
• 19: Jeffrey C. Mogul and Lawrence S. Brakmo. Method for dynamically adjusting multimedia content of a Web page by a server in accordance to network path characteristics between client and server, June 2001. United States Patent 6,243,761.
• 20: Jeffrey C. Mogul , Fred Douglis , Anja Feldmann , Balachander Krishnamurthy, Potential benefits of delta encoding and data compression for HTTP, Proceedings of the ACM SIGCOMM '97 conference on Applications, technologies, architectures, and protocols for computer communication, p.181-194, September 14-18, 1997, Cannes, France
• 21: Jakob Nielsen. Designing Web Usability. New Riders, 2000.
• 22: Ram Rajamony and Mootaz Elnozahy. Measuring client-perceived response time on the WWW. In USENIX Symposium on Internet Technology and Systems, San Francisco, California, USA, March 2001.
• 23: Srinivasan Seshan, Mark Stemm, and Randy H. Katz. SPAND: shared passive network performance discovery. In USENIX Symposium on Internet Technologies and Systems, Monterey, California, USA, December 1997.
• 24: Apache::Throttle - Apache/Perl module for speed-based content negotiation. http://www.dmi.usherb.ca/laboratoires/documentations-logiciels/Perl/lib/Apache/Throttle.html$.
• 25: Craig E. Wills , Mikhail Mikhailov , Hao Shang, N for the price of 1: bundling web objects for more efficient content delivery, Proceedings of the 10th international conference on World Wide Web, p.257-265, May 01-05, 2001, Hong Kong, Hong Kong [doi>10.1145/371920.372061]

INDEX TERMS

Primary Classification:
  C. Computer Systems Organization
  C.2 COMPUTER-COMMUNICATION NETWORKS
      C.2.2 Network Protocols

Additional Classification:
  H. Information Systems
  H.5 INFORMATION INTERFACES AND PRESENTATION (I.7)
      H.5.3 Group and Organization Interfaces
          Subjects: Web-based interaction

General Terms:
Measurement, Performance

Keywords:
client characterization, client connectivity, server adaptation