Re: [dnsext] Re: I-D ACTION:draft-vandergaast-edns-client-ip-00.txt

[Colm MacCárthaigh <[e-mail]>]

On Thu, Jan 28, 2010 at 5:18 PM, Paul Vixie <[e-mail]> wrote:
> if we're going to add something to a global protocol and a global system,
> we have a burden to do good engineering, not just offer point solutions
> that help a single industry.  there's a reason why web browsers can state
> their preferred language, which is customized content.  if we're going to
> customize dns responses at all (which is controversial) then we have to
> allow for the possibility that some dns content producers will want to
> customize based on things beyond just client ip.

Customising responses is already a de-facto part of shared internet
infrastructure. It seems like a standard that mitigates the
sub-optimalities of this approach and improves service for users and
providers alike is desireable. The alternative is non-interoperable
and behind-the-scenes ad-hoc mechanisms.

Restricting the technique to Client IP seems very reasonable, mainly
because it's the most prevalent such factor today. But also because
the aim here should be to improve DNS as a naming mechanism for the
internet, not to make it arbitrary client-identification mechanism.
That doesn't seem like poor engineering.

> (also, many client-ip addresses will be nat'd, and in that sense, ip is no
> longer end to end, and so including a client-ip in the query is a layering
> violation beyond just its fealty to dns incoherency.)

I think the draft takes explicit account of this issue.

-- 
Colm