W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2014

Re: HTTP/2 vs. proxies ?

From: Eric J. Bowman <eric@bisonsystems.net>
Date: Sat, 21 Jun 2014 18:21:20 -0600
To: "Poul-Henning Kamp" <phk@phk.freebsd.dk>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <20140621182120.86e7817cc2247f36cc40197b@bisonsystems.net>
"Poul-Henning Kamp" wrote:
> 
> I'm increasingly getting the feeling that we have people who like
> the HTTP/2.0 draft and people who work with proxies, and that those
> two sets are almost exclusive ?
> 

Yes, from the outside looking in, it's quite clear that this is the
first major Internet protocol being driven by corporate concerns rather
than architecture. Focusing on the requirements of the next product-
release cycle, to the exclusion of any concerns about how the Web would
have collapsed under its own weight without intermediaries, quite
naturally leads to the aversion we're seeing to T-E or anything else
that slows down and fattens up the protocol in the name of universal
interoperability.

The corporate concern is to go faster, my concern as a freelancer is
free scaling, which we don't see with lesser corporate-driven protocols
like WebSocket -- obviously, to me, a result of being on port 80 in a
way which excludes intermediary participation. WS costs far more to
deploy than HTTP/1.1, and if this winds up being the case with HTTP/2
then it'll be just as "corporate" a solution as WS and HTTP/1.1 will be
with us for a good while to come.

I'd say this one comes down to money. The big players have a financial
stake in being able to use their size to dominate the market; the rest
of us have loved HTTP caching for 15+ years now, because it levels the
playing field. This issue manifests itself in "working with proxies",
but I'd rather not kid myself about the real motivation for this break
with the time-proven Taylor-school approach to Web architecture. I
can't stress enough, if it ain't broke, don't fix it; intermediaries
were the fix to what was broken before, why would anyone want to do
away with it in this day and age for architectural reasons?

> 
> A)  I think HTTP/2.0 is ready for last call		YES/NO
> 

No, the only thing HTTP/2 is ready for is T-E, which will only happen
if, not when, the corporate interests jump on the interoperability
bandwagon. So I'm sure HTTP/2 will go to last call without it, this is
simply not the consensus view (although I question whether this would
be the case if HTTP/2 were being developed by those who care about
architecture over the corporate bottom line for the next quarter).

>
> B)  My primary HTTP/2.0 interest is proxy technology	YES/NO
> 

I'd state that my primary concern with HTTP/2 is, and always has been,
that the ability of intermediaries to participate in client-server
communication be preserved. Otherwise we regress architecturally, back
into known problems with scaling which can otherwise only be overcome
by those with deep pockets -- the very corporations driving development
of this protocol, who don't share this concern with those marginalized
here.

-Eric
Received on Sunday, 22 June 2014 00:21:18 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:31 UTC