W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2015

Re: HTTP/2.0 independent of TCP/IP

From: Martin Thomson <martin.thomson@gmail.com>
Date: Fri, 18 Dec 2015 09:46:32 +1100
Message-ID: <CABkgnnXM1YE1svCLiMBmQE5z6p4HqmipEVS1hXRxyzEEk0J55A@mail.gmail.com>
To: Daniel Stenberg <daniel@haxx.se>
Cc: Dave Wain <dave.wain@ntlworld.com>, HTTP Working Group <ietf-http-wg@w3.org>
While it's never really been written down (Mike wrote a good piece on
it here: https://tools.ietf.org/html/draft-bishop-decomposing-http-01),
the assumption is that whatever transports HTTP requests has the
ability to correlate requests and responses.  HTTP/1.1 does that by
using TCP and by matching the order of requests and responses.  HTTP/2
does that with stream identifiers.  Anything else would be a different
protocol again.

On 17 December 2015 at 21:01, Daniel Stenberg <daniel@haxx.se> wrote:
> On Thu, 17 Dec 2015, Dave Wain wrote:
>
>> Guys, a brief reminder for making HTTP/2.0 support low bandwidth/offline
>> applications.
>
>
> Can you take this a step further and state what you actually think we or
> anyone should do? HTTP/2 (without .0!) is already done and shipped.
>
> As it isn't TCP/IP, what is the transport you're considering? Why do you
> think we need to modify HTTP/2 for everyone instead of you fixing your
> transport to have the necessary mechanisms to transfer HTTP/2 properly?
>
> --
>
>  / daniel.haxx.se
>
Received on Thursday, 17 December 2015 22:47:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 11:11:40 UTC