Re: DNS: defining "official" domains

Re: DNS: defining "official" domains

From: Simon Hackett <simon§>
Date: Fri, 19 Jun 1998 11:11:06 +0900
With respect to the response from Leni (below) to Rick's original
assertions about "" and "" etc being "bad" and being
"good" for some definitions of those words, here's my $0.02 worth:

 Actually, if you think it over, there is not a huge difference between and  (say), in that in the end, there is an individual who
can decide that it's all too hard and there won't be any more names
registered any more, if they really really want to.

In the former case, the name of that person is Robert Elz for (NB
Not Internet Names Australia). In the latter case (,  the name of
that person is Simon Hackett. And in both cases, the likeihood of that
happening is very low. And in both cases, the business risks inherent in
obtaining a delegation in either are, correspondingly, also very low.

I think that Rick's notion of criticising people for allowing subdomains of
domain spaces that they control to be delegated to others is rather a
strange one, given that the entire internet DNS works that way! 

[NB I don't delegate into for a living - the operator of,
however, does - which isn't a reason not to use - since Larry's
committed to keeping working precisely because he _does_ do it for a
living, I'd say the chances of the name space surviving in that case are
(also) extremely high]

The only thing that *is* really silly is the notion of using a completely
alternative domain name space that the existing, core root servers that
service 99%+ of the internet DNS resolvers cannot see, formed out of a
dislike for the existing processes of operating root servers.

Please stop here if you already know why that notion is silly. The rest of
this message is simply an explanation of why it is, in some detail....

(still here? ok, read on...)

The big problem with this isn't that nearly everyone can't see them - the
problems would really start if a statistically significant number of DNS
resolvers *could* see them - because there can be more than one of them!
THere is no limit to the number of people who can create alternative root
structures, and nothing stopping them. Indeed, when run in private, between
consenting adults operating internet connected organisations, it is simply
another (private) use of the internet to run applications across it -
doesn't affect the rest of us, positively *or* negatively.

However, they would be an issue for the rest of the planet if they started
actually being *used* by significant numbers of users in the world instead
of the 'real' root servers.

When it comes to alternative universe #1 and alternative universe #2  both
deciding that, within their universes, that they will both create a
(different) top level domain but with the same name (hmmm, lets say
".UNSTABLE" just for instance), then you get an unresolvable conflict of

Now , where multiple disparate confederations do this, the answer to a DNS
query in the .UNSTABLE domain goes from being either the correct answer or
"not found" (if you're not using that confederation's root servers), which
is already bad enough, and instead it becomes "alledgedly correct answer
#1", or "alledgedly correct answer #2" or "not found" depending on your
choice of root DNS server. Clearly untenable!

"Two men say they're jesus, one of them must be wrong".

In other words, in case it's not obvious, the only stable number of
authoratative root name server farms is *one*. Not any number greater than
one. The notion that a new root name server confederation is somehow more
worthy than the existing one is a non sequitur, and I suggest we
concentrate on issues related to the globally accessible namespace and to
changing it _in place_ in rational manners if and as needed.

The other aspect, of course, in which the 'real' root servers are distinct
to other confederations is that the 'real' ones are subject to huge, long
winded, time consuming, public discussions prior to the creation of
critical things like new top level domains.Those processes are,
hollistically, quite a good idea, because they make it very hard for
whimsical or stupid changes to be made at a critical point in the internet
(one of the few single points of failure in the whole game) 

But new top level domains in other universes can get created at the whim of
a single individual. This *can* actually work if the individual is seen by
the community to be a responsible, careful, and caring person in the realm
we're dealing with. Indeed, it is largely due to the care exercised by Jon
Postel to date that the whole shooting match works as well as it does - and
now the stability he helped architect is being gradually migrated to a new,
controlled, structure to help it remain stable.  

But ... imagine if some other individual was not nearly as caring, or
consistent, and as the operator of an alternative confederation, decided to
actually create new top level domains in response to a whim, or a joke?
Such a person might, clearly, decide to revoke them with just as little
consideration, or due process, and that would literally ransom the world to
the moods of that person on a given day. Now that IS scary, isn't it?...


At 10:23 19/06/98 +1000, Leni Mayo wrote:
>Rick Welykochy wrote:
>> (*) *.au domains will go on forever ... they are THE sanctified
>>     TLD's in Australia
>Like many others I have a gut feel that some domains are "official"
>while others aren't.  It might be interesting to elaborate on what
>this means.
>Rick points out that no fees need to be paid to renew the domain.
>That seems to be a necessary condition, but not sufficient.
>There needs to be a chain of delegation authority from IANA.  Again,
>necessary, but not sufficient.
>Continuing to mull...
Simon Hackett, Technical Director, Internode Systems Pty Ltd
31 York St [PO Box 284, Rundle Mall], Adelaide, SA 5000 Australia
Email: simon&#167;  Web:
Phone: +61-8-8223-2999          Fax: +61-8-8223-1777
Received on Fri Jun 19 1998 - 12:44:38 UTC

This archive was generated by hypermail 2.3.0 : Sat Sep 09 2017 - 22:00:03 UTC