Re: alpn identifiers (not again!)

new identifiers are far from cheap.

Any proxy that needs to understand them needs to be re-deployed.

I think since we're hell-bent on breaking layered networking basic 
tenets, we will run into these sorts of problems.

------ Original Message ------
From: "Martin Thomson" <martin.thomson@gmail.com>
To: "Stefan Eissing" <stefan.eissing@greenbytes.de>
Cc: "HTTP Working Group" <ietf-http-wg@w3.org>
Sent: 11/06/2015 4:06:40 a.m.
Subject: Re: alpn identifiers (not again!)

>On 10 June 2015 at 02:40, Stefan Eissing <stefan.eissing@greenbytes.de> 
>wrote:
>>  The server may offer ALPN with DTLS and what will the client send? 
>>"h2".
>
>It would send h2dtls: the version of HTTP/2 that runs over DTLS.  It
>would be different from the one that runs over TLS in several crucial
>ways, so it does need a new identifier (I mean, you could use other
>information to disambiguate, but why do that when new identifiers are
>cheap).
>

Received on Wednesday, 10 June 2015 22:00:52 UTC