At 10:40 AM 23/01/02 -0500, Mike Champion wrote:
>Stealing a page from Tim Bray's hymnal <grin>, I feel compelled to
>ask: Is there solid, empirical, profiling evidence that the TCP
>substrate of HTTP is a significant bottleneck in real web
>applications? Optimizing non-bottlenecks is a well-known
>temptation...
Everyone agrees that in theory at least HTTP makes lousy use of the
underlying TCP/IP infrastructure. This is one of the reasons, as
I remarked earlier, it's so hard to understand why HTTP seems
to work so well in practice and degrades gracefully under load.
Having said that, there seems to be strong quantitative evidence
that HTTP 1.1 ought to really help.
A ton of research is available at
http://www.w3.org/Protocols/HTTP/Performance/
Also see Jim Gettys' talk at
http://www.w3.org/Talks/970210HTTP/all.htm
-Tim
