[dns-operations] Why are cnames called "canonical" names?

[Edward Lewis]

On 4/10/15, 12:28, "Fred Morris" <m3047 at m3047.net> wrote:

>But gee, people keep going to example.com instead of www.example.com. I'd
>better create a CNAME which says "if you're visiting example.com, you
>really should be going to www.example.com".

Another way to look at this is, www.example.com is not a host name
(server) but a service name.  The same way ftp.example.com and
smtp.example.com and gopher.example.com have been in use (but for some
reason not telnet.example.com or ssh.example.com).  All of those could own
CNAME RR (singleton) sets with example.com in the RDATA (except for
ns.example.com) - or some other "canonical" host.

I used to have address records at the apex, with CNAME RRs at the service
names (except for NS) referring to the apex.  It seemed to work, but
sometimes folks (managers) just didn't like the idea of using CNAME RRs.

The desire for CNAME-at-apex is unrelated, it is an artifact of how cloud
service providers manage IP addresses for their customers.

Going back to the original question - my off-list recommendation was to
just ask the authors of the RFC and/or other people involved back then.  I
know a lot of rationalizations were just never written down.  The list can
haggle over what they think the meaning is - or try to feed this to Paul
Hoffman who is trying to write the "terminology" document.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4604 bytes
Desc: not available
URL: <https://lists.dns-oarc.net/pipermail/dns-operations/attachments/20150410/3e5873fa/attachment.bin>