The World Wide Web (WWW) generates a significant and growing portion of traffic on the Internet. With the click of a mouse button, a person browsing on the WWW can generate megabytes of multimedia network traffic. WWW's growth and possible network impact merit a study of its traffic patterns, problems, and possible changes. This paper attempts to characterize World Wide Web traffic patterns. First, the Web's HyperText Transfer Protocol (HTTP) is reviewed, with particular attention to latency factors. User access patterns and file size distribution are then described. Next, the HTTP design issues are discussed, followed by a section on proposed revisions. Benefits and drawbacks to each of the proposals are covered. The paper ends with pointers toward more information on this area.
[1]
V. Jacobson,et al.
Congestion avoidance and control
,
1988,
SIGCOMM '88.
[2]
Ari Luotonen,et al.
World-Wide Web Proxies
,
1994,
Comput. Networks ISDN Syst..
[3]
Tim Berners-Lee,et al.
Uniform Resource Locators (URL)
,
1994,
RFC.
[4]
Tim Berners-Lee,et al.
The World-Wide Web
,
1994,
CACM.
[5]
Karen R. Sollins,et al.
Functional Requirements for Uniform Resource Names
,
1994,
RFC.
[6]
V. Jacobson.
Congestion avoidance and control
,
1988,
CCRV.
[7]
Jeffrey C. Mogul,et al.
Improving HTTP Latency
,
1995,
Comput. Networks ISDN Syst..