Re: Protocol Benchmarking (HTTP protocol)

• Mail folder: WWW Talk Jan 94-present
• Next message: sg04%kesser@gte.com: "re: Mosaic Accessories"
• Previous message: neuss@igd.fhg.de: "Re: Configurable log proposal"
• Maybe in reply to: frans van hoesel: "Re: Protocol Benchmarking (HTTP protocol)"
• Reply: Dave_Raggett: "Re: Protocol Benchmarking (HTTP protocol)"

Message-id: <9402021252.AA23525@dxmint.cern.ch>
Date: Wed, 2 Feb 94 07:52:22 EST
From: hgs@research.att.com (Henning G. Schulzrinne)
To: www-talk@www0.cern.ch
Subject: Re: Protocol Benchmarking (HTTP protocol)
Content-Length: 629

"Chained"/multiple GETs would also really help with TCP-level firewalls,
where connection set-up cost are substantial, as well as with international
links where the three-way TCP handshake can easily add seconds to the
retrieval of a short document with lots of small images/icons.

The MGET method seems preferable; timeouts are always tricky to get
right. One possible difficulty is error reporting, unless an all-or-
nothing error indication is sufficient.

---
Henning Schulzrinne (hgs@research.att.com)
AT&T Bell Laboratories  (MH 2A-244)
600 Mountain Ave; Murray Hill, NJ 07974
phone: +1 908 582-2262; fax: +1 908 582-5809