- From: Thomas Fossati <tho@koanlogic.com>
- Date: Tue, 24 Jan 2012 21:59:37 +0100
- To: Henrik Nordström <henrik@henriknordstrom.net>
- Cc: Poul-Henning Kamp <phk@phk.freebsd.dk>, Mark Nottingham <mnot@mnot.net>, HTTP Working Group <ietf-http-wg@w3.org>
On Jan 24, 2012, at 8:37 PM, Henrik Nordström wrote: > tis 2012-01-24 klockan 08:13 +0000 skrev Poul-Henning Kamp: > >>> * Ability to be deployed on today's Internet, using IPv4 and IPv6, in the presence of NATs >> >> Apart from "no reverse connections" I have no idea what this means. > > As worded it basically means it must run over port tcp/80. SCTP is not > an option if the above is taken literally as most NATs do not know what > to make out of SCTP packets. I think that this NAT-friendliness requisite is excessively limiting and, what's more, could be happily outdated by the time this new HTTP is going to be really deployed - which would be really regrettable. I hope we don't waste this occasion to rethink HTTP deeply, by which I mean that we should not hold us back in experimenting with other transports other than TCP, or else our final outcome will be unnecessarily weakened. Let this be a real 2.0, not just a buzzword :-)
Received on Tuesday, 24 January 2012 21:00:55 UTC