- From: Willy Tarreau <w@1wt.eu>
- Date: Wed, 8 Jul 2015 20:11:13 +0200
- To: Mike Bishop <Michael.Bishop@microsoft.com>
- Cc: Mark Nottingham <mnot@mnot.net>, Fedor Indutny <fedor@indutny.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
On Wed, Jul 08, 2015 at 05:59:23PM +0000, Mike Bishop wrote: > But if the terminator is speaking HTTP/2 on both sides, is it muxing requests > into fewer connections? A extension that adds "present on all requests" > semantics would work if every client connection has a corresponding server > connection, but that's not necessarily the case. I think there's no need for such a thing in the protocol, it can easily be implementation-specific as we all do with various header additions on the gateway such as x-ssl or whatever. The simple idea of making people *think* there's a correlation between both sides worries me about possible outcomes such as broken products which consider that their mode is the correct one and who will expect all agents around them to maintain the 1:1 relation between both sides connections. That sounds a bit scary to me to be honnest. Regards, Willy
Received on Wednesday, 8 July 2015 18:11:45 UTC