- From: Mike Belshe <mike@belshe.com>
- Date: Wed, 2 Jul 2014 04:56:01 -0700
- To: Greg Wilkins <gregw@intalio.com>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
- Message-ID: <CABaLYCs4a=zidQac0+gVHdLqoXfgxeRNV2NhakFQSMySK_XxKg@mail.gmail.com>
On Wed, Jul 2, 2014 at 4:36 AM, Greg Wilkins <gregw@intalio.com> wrote: > > On 2 July 2014 21:30, Poul-Henning Kamp <phk@phk.freebsd.dk> wrote: > >> In message <CAH_y2NHdOt0nmxZkXcoHyWrv-Map7cCqJh= >> O3JRVow-1sDhB5g@mail.gmail.com>, Greg Wilkins writes: >> >> >c) This WG will realise the error of conflating framing with application >> >layer semantics and the protocol is significantly rewritten. >> >> Isn't that called SCTP ? >> > > Indeed. And the inability to actually deploy SCTP in a world with hard > baked support for TCP should serve as a lesson against hard baking support > for HTTP into the h2 framing. > No no no - we absolutely *did* learn the SCTP lesson. We intentionally yet begrudgingly incorporated some of the features out of the transport (multiplexed streams) because we learned from SCTP that it couldn't effectively be done at the transport layer. We learned that transport protocols had an upgrade problem, but application protocols did not. So we should all feel great about this - multiplexed streams were trapped for over a decade inside SCTP, and h2 is finally unlocking it for http. This is a huge learn-from-history victory. Mike > cheers > > > -- > Greg Wilkins <gregw@intalio.com> > http://eclipse.org/jetty HTTP, SPDY, Websocket server and client that > scales > http://www.webtide.com advice and support for jetty and cometd. >
Received on Wednesday, 2 July 2014 11:56:28 UTC