- From: Alex Rousskov <rousskov@measurement-factory.com>
- Date: Tue, 10 Dec 2002 09:10:27 -0700 (MST)
- To: Scott Lawrence <lawrence@world.std.com>
- cc: Yogesh Bang <Y.Bang@zensar.com>, <ietf-http-wg@w3.org>
On 10 Dec 2002, Scott Lawrence wrote:
> B) Express the limitation using HTTP Range (your choice 1). I have
> seen this done in exactly this application. See Alexs comments on
> the possible problems with proxies that don't honor Range (is there
> any emperical data about how deployed proxies respond to this?).
IIRC, a popular open-source Squid proxy would do one of the following,
depending on the Squid version, hit/miss status, and complexity of the
Range expression:
- tunnel the Range transaction through
- respond with entire [cached] entity
- convert Range request to a regular GET,
possibly cache the entire response, and
extract requested range(s) to be sent the
the client
- respond with requested [cached] range(s)
I bet that other deployed proxies use a combination of the above
techniques as well. Keep in mind that some proxies are "transparent"
and some are "reversed", making a naive client unaware of the proxy
presence. This makes it difficult, if not impractical, to auto-adjust
client behavior depending on the environment.
Alex.
--
| HTTP performance - Web Polygraph benchmark
www.measurement-factory.com | HTTP compliance+ - Co-Advisor test suite
| all of the above - PolyBox appliance
Received on Tuesday, 10 December 2002 11:10:35 UTC