Re: [dnsext] Fwd: New Version Notification for draft-sury-dnsext-cname-dname-00

[Andrew Sullivan <[e-mail]>]

On Fri, Apr 23, 2010 at 06:41:15PM +0200, Ondřej Surý wrote:
> I think that this part of SMTP standard is wrong.  Users (applications  
> or end users) of DNS should not resolve CNAMEs on their own.  

Woah, nelly.  Let's not put ourselves in the position of telling other
applications what they should or should not be doing with data in the
DNS, if for no other reason than that it runs exactly opposite to
years of other advice.

> a) MTA implements a query cache
> b) the query cache is very strict on CNAMEs
> c) mail is sent to target.dname.cz and CNAME for target.dname.cz is  
> stored in the cache
> d) mail is sent to www.target.dname.cz and the query cache tries to  
> store everything received (including the DNAME) into the query cache.
>
> => error is reported when storing the DNAME record into the query cache
>
> But I dare to say that such implementation of dns client is broken.

Why is it broken?  The CNAME rules say nothing else should be at the
CNAME.  Your proposal violates that rule.  You can't then say that
someone checking that the CNAME doesn't have anything else at that
name is actually broken: it isn't, at least today.

I'm not saying we shouldn't make this change to CNAME and DNAME, but
let us not dismiss the objection that we're making a change by begging
the question.

A

-- 
Andrew Sullivan
[e-mail]
Shinkuro, Inc.