Re: [DNSOP] Some distinctions and a request - Have some class?

[Andrew Sullivan <[e-mail]>]

On Sun, Jul 05, 2015 at 08:17:03AM +0100, Ray Bellis wrote:
> 
> Sure, CNAME is *defined* for all classes, but AFAIK there's no way to "jump"
> out of one class into another using a CNAME.

No, that's correct.  But if the point of using a class is to create a
separate namespace, then the fact of class-independent RRTYPEs means
you can't do that.  As Paul Vixie notes, there appears to be some
ambiguity with CNAMEs on this front, but as nearly as I can tell RFC
6672 makes this plain for DNAME.

Imagine the alternative-resolution class FAKE.  In the IN class,
example.com has a DNAME entry pointing to example.net.  What should
happen when someone performs a query for QNAME localentry.example.com,
TYPE AAAA, and CLASS FAKE?

RFC 6672's description of the algorithm does not use CLASS as a
distinguishing criterion.  So, I think the answer is the DNAME
processing should return the results for localentry.example.net,
regardless of the class.  As a consequence, CLASS does not work to
provide different completely independent namespaces, and therefore
co-ordination across the class registrations will be necessary.  In
effect, CLASS doesn't work.

At least, that's my reading of the RFC.  I'd be pleased to be wrong.

Best regards,

A

-- 
Andrew Sullivan
[e-mail]