SOA and NS are canonical records?

June nfbz2003 at yahoo.com
Wed May 5 02:37:07 UTC 2004


thanks for jumping in....

SOA and NS as canonical names are different from the other type of canonical
name (A RR), for one thing, they can not be used as rdata, right?

In addition, these, and others like MX record, can also be understood as
"alias", they redirect a particular type of record (MX, NS, SOA, etc) to the
canonical names (an A record, as RDATA or MNAME in case of SOA).  My
questions is why they are called canonical names, not alias, as they are not
really that different from CNAME RRs.  I don't see them being referenced as
canonical names in docs (I haven't read a lot reference), and I'd like to
know these references if someone has seen them.  If there is a RFC defined
them this way, I'd also like to know; or if people just refer to them this
way, and have been doing this years, that's fine too, but I'd like to get
some confirmations from the veterans here.

"Christian Smith" <none at i.am.invalid> wrote in message
news:c79dlp$2488$1 at sf1.isc.org...
> In article <c79ab4$20ur$1 at sf1.isc.org>, "June" <nfbz2003 at yahoo.com>
> wrote:
>
> > thanks for the reply.
> >
> > Is there any reference says SOA and NS RRs both are canonical names? Or
all
> > RRs, except CNAME (alias), are canonical names, by definition?
>
> In a CNAME record, the left side is the "alias" and the right side is
> the canonical name.
>
> Grab you local Websters/Oxford's and look up "canonical".
>
> With an SOA and an NS record, the left side is always a canonical name.
>
> In an NS record the right side is also a canonical name.
>
> Obviously the right hand side of an SOA is neither a canonical name nor
> an alias.
>




More information about the bind-users mailing list