- From: Simon E Spero <ses@tipper.oit.unc.edu>
- Date: Fri, 16 Dec 94 16:31:30 -0500
- To: Dave Kristol <dmk@allegra.att.com>
- Cc: connolly@hal.com, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
>>>>> "Dave" == Dave Kristol <dmk@allegra.att.com> writes: Dave> My tastes (obviously) run to a more evolutionary approach Dave> for HTTP. I'm unconvinced that the performance problems Dave> require a flash cut to a binary protocol. Spero has shown Dave> that doing multiple transactions over one connection Dave> achieves signficant performance improvements. His response Dave> is to change HTTP drastically. Mine is to do so within the Dave> current overall design. Well, there's always my other response :-) One of the things I talked about with Alex at the IETF was making a slight change to the SESSION proposal. The idea is to use the SESSION method to switch the connection over to running SCP, and then use SCP to manage multiple sequential HTTP 1.0 transactions. It turns out that this technique allows even more code to be reused than using MIME multipart, and provides a very obvious transition path to full HTTP-NG. One other thing that was discussed was the relative advantages of using a session method vs. an ignorable header. It turns out that there is a problem with using ignorable headers when proxies are used - if a proxy which doesn't interpret the header is used to talk to a server which does handle the header, the connection can become deadlocked (the end server things that the proxy doesn't want it to drop the connection, whilst the proxy is sitting there waiting for the connection to drop). Simon
Received on Friday, 16 December 1994 13:33:00 UTC