On Fri, 02 Mar 2001 15:35:02 +0700, "Rahmat M. Samik-Ibrahim"
<rms46(_at_)vlsm(_dot_)org> said:
Rahmat> Mohsen BANAN-Public wrote:
>> Did you follow the discussions that I initiated on
>> a similar set of topics on the ietf(_at_)ietf(_dot_)org
>> mailing lists about two years ago?
Rahmat> Nope, but what was the conclusion? Where (URL) is it archived?
That thread is part of the ietf(_at_)ietf(_dot_)org archives.
You can use the date and subject of my previously
included message to refer to the rest of that
discussion.
Separately, I'll email you a segment of my own
archives of that thread.
Rahmat> Basically, my question was because of IAB's assertion
Rahmat> in RFC-2826 "IAB Technical Comment on the Unique DNS Root"
Rahmat> ( http://www.faqs.org/rfcs/rfc2826.html )
That document does not say much other than
- DNS has these limitations
- We should live with it
- IAB can't (or does not want to) move forward
What I am proposing keeps DNS names unique at all
times, and allows for existance of multiple root
server clusters.
The solution relies on viewing the problem not as just
a DNS problem but as a DOMAIN NOTATION limitation and
problem as well.
We can easily extend the Domain Notation and solve the problem.
Time is not on the side of those who want to live with
the DNS limitations.
Legitimacy of the root server cluster comes from the
users who point to them. Large communities of users
(AOL, Chinese government, Microsoft, ...) can easily
extend the name space and they can even extend the
Domain Notation. If such efforts diverge, we'll end up
with a real mess.
Remember Halloween, remember "Embrace and Extend",
remember "Instant Messaging", ...
I propose that instead of waiting, we do something
about it. See my previous message.
In this case, scarcity in the DNS implementation gene
pool (i.e., bind) can be viewed as an asset. I would
love to hear from the bind team why what I am
proposing can not be workable.
>> Bob Allisat <bob(_at_)fcn(_dot_)net> followed up on that idea...
Rahmat> I am not aware that he is interested in the technical aspect
Rahmat> of DNS.
Correct.
Participation of Bob Allisat in this mailing list was
often as a "User Advocate". As a tough user advocate,
he looked at what I had proposed and supported
it. That was of significance to me.
The DNS-ng solution should revolve around allowing the
user to choose amongst competing root server clusters
and naming authorities. Once that concept has been
recognized, the rest of the technology is real simple.
Rahmat> Perhaps, we should discuss this in private. What I have in
Rahmat> mind is somewhat of an "address book" that is publicly
Rahmat> accessible, perhaps through an ordinary DNS. Since it is
Rahmat> publicly accessible, it can be shared/adopted by others.
Okay.
I'll follow up privately soon.
...Mohsen.