Re: A long strange trip

[thomst@netcom.com (Thom Stark)]

> At 11:40 PM -0600 12/24/96, Thom Stark wrote:
> >Y'know, one thing the IAB, ISOC and ISOC's Board ought ALL to commit to
> >RIGHT NOW is that, when the NSI contract expires, NSI's ability to
> >update the DNS root servers should IMMEDIATELY be revoked.
> >
> >This would not only end NSI's monopoly on the .com domain, but forever
> >eliminate NSI as a registrar--something that, in my view, they richly
> >deserve for being:

To which Dave Crocker responded:

> 	How do you propose to implement this suggestion in a fashion which
> does not also penalize all the .com/.net/org users?
> 
> 	As we keep saying, operational impact needs careful consideration.

Absolutely.

As at least one of the lawyers on this list has observed, the records
which NSI has generated regarding "its" customer base are actually the
property of the US government, specifically of the NSF.  Upon the 
expiration of its contract with the NSF, NSI will be obligated to give
up those records to the NSF as part of the process of concluding the
contractual relationship.

No penalty would apply to .com/etc. users who were registered and whose
domains were being serviced as of the date NSI goes out of business as
a registrar.  Their WHOIS records, as well as the paper trail would be
transferred to NSF, which, in turn, could simply pass them on to the
new registrar's group via CORE or any other intermediate which might be
created for the purpose.  No service interruption need occur for the
users.

Assuming that the current group of gTLDs will be shared among the newer
registries in the same fashion that the IAHC draft has envisioned the 
newer gTLDs be shared, (i.e.--.com/etc. registrations will be made
available through any registrar), then the only question that really 
needs to be answered about NSI's existing customer list is how it will
be parceled out fairly among the newer registrars.  Any equitable
scheme could work.  A possible model would be to simply sort the records
by alphabetical order, then "deal" them out to the registrars much like
a deck of cards, in sequential, round-robin fashion, until the entire
"deck" has been dealt out.  Every registrar would then have "inherited"
an almost exactly equal number of existing .com/etc. registrations in
a demonstrably unbiased manner.

If any existing .com/etc. domain holder didn't like the registrar to 
which it had been randomly assigned, it would be free to change registrars
at will.  There would never be a time during the entire process when any
existing domain holder's name would fail to be routed.

Some things would inevitably fall through the cracks.  Two that spring to
mind are those folks who submitted new .com/etc. requests in the hours
leading up to the closure of NSI as a registrar and those who wished, for
whatever reason, to update their records during the transition period
between the paperwork leaving NSI and arriving at the new, randomly-
assigned registrar.  Mechanisms would have to be developed to minimize
the impact of the transition on these folks, but, with sufficient front-
end planning and with sufficient beofre-the-fact advertising of the
handoff, the number of users affected could be kept pretty small.

All the above is predicated on a whole bunch of assumptions on my part,
of course.  Some of those include:

  1. The shared gTLD model will be a part of the final draft

  2. IAHC and, more importantly, NSF are persuaded that NSI has
     demonstrated its unfitness to continue as a registrar.

  3. All of the newer registrars will actually want their equal
     share of the older gTLD customers

This is all off the top of my head.  If you want greater detail, RFC-
compliant section numbering, etc., just ask.

Regards,

Thom Stark

Email:  thomst@netcom.com              URL:  http://www.dnai.com/~thomst
             finger thomst@netcom.com for my PGP Public Key

(510) 526-9600 voice       STARK  REALITIES           fax (510) 526-9063
POB 457                     El Cerrito, CA                ZIP 94530-0457