--On Monday, December 28, 2009 01:16 +0000 John Levine
<johnl(_at_)iecc(_dot_)com> wrote:
It seems to me that if we think it's a good idea to specify a
domain name that doesn't exist, we're better off clarifying
the status of the ones already specified rather than inventing
new ones. Since the people who manage .ARPA are the exact
same people who manage the root (IANA, operated by ICANN, in
both cases), one is as likely to flake as the other.
While you are presumably right about "flake", .ARPA is rather
clearly under IAB supervision, while the root is under the
supervision of the ICANN Board and ICANN policy-making
processes. That distinction is fairly significant for some
purposes. It may or may not be for this one. FWIW, I'm not
aware of there being an established procedure for adding new
names to the ICANN reserved list: while several of the names are
"ours", I don't recall their being requested in any systematic
way. Others of the names are of entities who have leverage
within ICANN. One could make a strong case for other bodies
having their names blocked or reserved on a par with those
listed: IEEE, ISO, and ITU come to mind as examples.
In fact, ICANN is quite aware of the reserved names list. In
the current draft of the application process, one of the steps
is to check to see if a proposed name is one of the Reserved
ones, in which case the application fails immediately. Here's
their reserved list:
...
Nonetheless, it occurs to me that the set of DNS names that are
reserved or that have special meanings in some protocols are
scattered over a lot of different RFCs. So I wrote a strawman
to collect them all in one place and make a registry of them:
draft-levine-reserved-names-registry-00.txt
I think I got all the names, I did some greps over all of the
text RFCs looking for things that resembled domain names, and
I looked to see what's actually in .ARPA and the root.
Two quick observations on that draft: (1) It would be good to
have "domain" explicitly in the title. We have reserved names
in other places. (2) While there are reasonable arguments for
gathering a list in one place, having multiple lists under
parallel development in multiple groups presents a
synchronization problem. Since the RFC-writing process is not
the only way that names can be added to the list and this is a
list of names to not be allocated (i.e., IANA may not know), it
is reasonably likely to diverge from reality even if it doesn't
start that way. For example, the "new names" allocation
process in ICANN might put another name on the list without
identifying that fact clearly enough to IANA to get the name on
the list. A clearly non-normative glossary of reserved names
might be a more practical idea.
If other people agree that it's a good idea to have a place
that IANA can point to for the reserved names, I'd be happy to
move this ahead. Or if we think the situation is OK as it is,
we can forget about it. But I see little wisdom in adding
another does-not-exist name with semantics not meaningfully
different from .INVALID or FOO.INVALID.
I see relatively little harm in reserving one subdomain of .ARPA
for protocol use, since doing so would isolate the string from
any ICANN issues. As I've said before, I think the notion of a
registry process allocating more of those names at the second
level would be a bad idea. And I'm not sure that the case has
been made for the single reservation: your last sentence above
is an argument that it has not been.
best,
john
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf