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

Re: ext#7 / ext#8: multiple alt-svc

From: Kinkie <gkinkie@gmail.com>
Date: Fri, 4 Jul 2014 11:36:23 +0200
Message-ID: <CA+Y8hcOOsyVFe4SpkDhpoccbUUBvhOQYLfPrDTfa+Mm1oayh8Q@mail.gmail.com>
To: Martin Thomson <martin.thomson@gmail.com>
Cc: HTTP Working Group <ietf-http-wg@w3.org>
On Thu, Jul 3, 2014 at 9:25 PM, Martin Thomson <martin.thomson@gmail.com> wrote:
> On 2 July 2014 22:02, Kinkie <gkinkie@gmail.com> wrote:
>> To me this sounds like a step back compared to DNS SRV records, which offer
>> priority, weight and ttl.
>
>
> I'll note that if (and it's a pretty big if) we support DNS SRV, then
> adding priorization and weight to the alternative service declaration
> would be *in addition to* SRV.  I think that keeping this simple is
> best, noting that the server can apply things like prioritization and
> weighting *before* sending the alternative service indication out.
> This is not an option in the DNS scenarios due to the DNS
> architecture, i.e., caching, but it's totally feasible here.

DNS SRV has been out for a long time, and so far it hasn't gained
traction except in very specific cases for some reason.
Using ALT-SVC at the HTTP level has advantages over it however: it
allows to have easy, dynamic, server-driven load redirection and
redistribution policies.

> There are some arguments for having the client be able to choose, but
> isomorphism with SRV is not one of those.

I agree. However feature-parity IMO is one such argument. Especially
if the needed change is - at least apparently - trivial.


-- 
    Francesco
Received on Friday, 4 July 2014 09:36:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:09 UTC