Re: h2 use of Upgrade

• From: Poul-Henning Kamp <phk@phk.freebsd.dk>
• Date: Mon, 01 Sep 2014 05:52:27 +0000
• To: "Roy T. Fielding" <fielding@gbiv.com>
• cc: HTTP Working Group <ietf-http-wg@w3.org>
• Message-ID: <4435.1409550747@critter.freebsd.dk>

--------
In message <91C83850-A992-4655-8AD5-CDC59794156B@gbiv.com>, "Roy T. Fielding" w
rites:
>This comment is in reference to section 3.2 of

>>    A server MUST ignore a "h2" token in an Upgrade header field.
>>    Presence of a token with "h2" implies HTTP/2 over TLS, which is
>>    instead negotiated as described in Section 3.3.
>
>There is no such token.  ALPN tokens have nothing to do with Upgrade
>and this is an HTTP/1.1 request, so this spec has no ability to make
>such an (unnecessary) requirement.

Roy,

You lack the background here:

Of the major browsers, only Microsoft has indicated that they are
going to support HTTP/2 without TLS.  Three or maybe four others
have said rather categorically that they are only going to implement
HTTP/2 with TLS.

This has made non-TLS HTTP/2 a stepchild in the WG, and a lot of
assumptions and decisions, stated or unstated, are based on this
"TLS-only" world-view.

One particular place where this has been evident, is that it has
been impossible to get a frame-layer and frame-layout optimized
for high-performance load-balancing and hardware implementations.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Monday, 1 September 2014 05:52:52 UTC