SOA and NS are canonical records?

[phn at icke-reklam.ipsec.nu]

June <nfbz2003 at yahoo.com> wrote:
> In DNS and BIND 4th ed. p499, there is a sentence seems confusing to me:

> "...Remember that fx.movie.edu has an SOA record and NS records, so
> attaching a CNAME record to it violates the rule that a domain name be
> either an alias or a canonical name, not both."

> My understanding is that SOA and NS are neither alias or canonical names,
> and I don't understand this statement here.

Both are canonical names. The issue is what the "key" of this line
is, the leftmost part. In addition both SOA and NS records should
have a "cononical name" on their right side too.

> On the other hand, you may say by nature, SOA and NS point to canonical
> names, so they are kind of alias (CNAME), ...and then, still doesn't make
> sense as then will be all aliases (allowable based on the above statement?).

CNAME defines "aliases". Basically a cname tells the querier :
"restart your query but ask for <RHS of CNAME> instead.

So any instance of a record that has both a CNAME and anything else
creates an ambiguity ( is the answer what one of the returned answers
or should i follow the new name instead).

Maybe it's easier to say that a CNAME may never be combined with
any other RR ( not even another cname)

> Actually the sentence before this one seems already made the point:

> "you can't have a CNAME record attached to an interior node like
> fx.movie.edu if it owns other records."

> So, basically "the other records" include A, SOA and NS, and CNAME record
> (alias) cannot exist with any of them at the same time.

Frankly there is occations where cnames coexists with other data
the most common is SIG ( used in DNSSEC). As dnssec is not
widley spread you can ignore this for now.


-- 
Peter Håkanson         
        IPSec  Sverige      ( At Gothenburg Riverside )
           Sorry about my e-mail address, but i'm trying to keep spam out,
	   remove "icke-reklam" if you feel for mailing me. Thanx.