wildcard cnames

[Barry Margolin]

In article <a8cm5k$b86 at pub3.rc.vix.com>, bert hubert  <ahu at ds9a.nl> wrote:
>I was wondering though, what is the opinion on wildcard CNAME records? A
>strict reading of 1034 means that they should NOT work, and in fact, in
>PowerDNS they don't.
>
>Now, Bind does support wildcard CNAMEs and a potential customer of ours has
>a slave zone with a wildcard CNAME in it.
>
>So we're wondering, is this something we should support. It requires
>modifying the RFC 1034 algorithm. Note how 'a.' mentions CNAME indirection
>and c. doesn't. 

My reading of the algorithm concurs with yours.

In fact, BIND has another wildcard "feature" that doesn't correspond to the
algorithm: it allows wildcards as non-terminal nodes, e.g.

foo.*.domain.com. <type> <value>

>Does anybody know of a legitimate use of wildcard CNAME records?

One of our customers has a bunch of domains that are all of the form:

@  A      x.x.x.x
*  CNAME  @

Another customer has a bunch of domains which all contain just:

*  CNAME  <name of their main webserver>

I guess this simplifies dealing with DNS.  They can create virtual hosts on
the web server using names within these domains, and they don't need to
contact us to add DNS entries all the time.  They only need to interact
with us to add new domains.

-- 
Barry Margolin, barmar at genuity.net
Genuity, Woburn, MA
*** DON'T SEND TECHNICAL QUESTIONS DIRECTLY TO ME, post them to newsgroups.
Please DON'T copy followups to me -- I'll assume it wasn't posted to the group.