ICANN ICANN Email List Archives


<<< Chronological Index >>>    <<< Thread Index >>>

Single-Registrant TLDs and Root Scaling

  • To: 3gtld-guide@xxxxxxxxx
  • Subject: Single-Registrant TLDs and Root Scaling
  • From: Werner Staub <werner@xxxxxxxx>
  • Date: Sun, 22 Nov 2009 22:53:50 +0100

The root scaling study has shown that there are operational concerns if
several thousand TLDs are introduced in over a short time frame.

It is evident that the DNS root cannot be managed in the same fashion
with 300, 3000 and 30,000 TLDs. If it moves to different orders of
magnitude, its management process must have time to evolve.

Single-Registrant TLDs are the only trend that can cause such an
explosion (high rate of change) of the root zone size.

All currently existing TLD registries are required to accept
registrations from arms-length registrants. None of them is allowed to
reserve its TLD for owners, affiliates or employees of the group
controlling the registry.

The underlying rule was stated by Jon Postel in RFC 1591 with the
following words:

 'Concerns about "rights" and "ownership" of domains are
  inappropriate. It is appropriate to be concerned about
  "responsibilities" and "service" to the community.'

In others words, the concept of "ownership" has always been ruled out
in the DNS root zone. Suddenly allowing single-registrant TLDs is
tantamount to declaring that anybody may grab a piece of the DNS root.

Requiring registries to accept unaffiliated registrants prevents an
avalanche of brand-based TLDs. Very few brands owners would use their
brand for a TLD where they are required to allow registrations by
arms-length third parties. The reason is that a brand would be weakened
if unaffiliated third parties could use it.

If on the other hand single-registrant TLDs are allowed, it is evident
that major brands will not be the only ones vying for them. Many
applications will be of a purely speculative nature. Given the priority
that existing TLDs enjoy with respect to TLD string confusability, it
is likely that many brands will apply defensively. Brand owners will
also be concerned about actions by their competitors. Not being in the
root would become a loss of customer visibility if their competitors
were to be in the root.

The result of the GNSO Policy Development Process for new gTLDs and for
Contractual Conditions of New gTLDs was unclear with respect to the
status of single-registrant TLDs. As a matter of fact, the GNSO policy
document requires in "Implementation Guideline I" that a gTLD must be
"used" within a specified timeframe, without however defining what
"use" means. If "TLD use" is defined in the current generally accepted
sense – namely to allow registration by arms-length third parties,
then single-registrant TLDs are not allowed.

If single-registrant TLDs are allowed, we should expect 3000 new gTLDs
in no time. If they are not allowed, there will probably not be more
than 200 new gTLDs.

(The figures could of course be affected by the troubles of the ICANN
process itself. There could be a lower number of single-registrant TLDs
in the first round because many potential applicants were not ready or
had not understood the process. But in the subsequent round, there
would easily be 5000, or 10,000 applications. If ICANN belatedly
intervened at that point, it would cause further unfairness, as it
could not take back the existing single-registrant TLDs.)

The responsible thing to do is to defer the decision on
single-registrant TLDs. This means that they should not be allowed in
the coming round. One way to do this is to reflect the GNSO PDP’s
"Implementation Guideline I" and to add the missing definition of "TLD

Here is a proposed wording:

 "Registry Operator commits to making the TLD available for second
  -level domain registrations based on an objective and transparent
  process. Registry Operator’s failure to do so within the 18 months
  of TLD Agreement signature will result in automatic termination of
  the Agreement and the removal of the TLD from the DNS root zone.
  Eligible registrants must include third parties unaffiliated with
  the Registry Operator."

Many TLD experts expect to sell consultancy services for
single-registrant TLDs. At first sight, consultants or technical
operators might mistakenly believe that it is in their interest to
advocate the immediate opening of the DNS to single-registrant TLD use.
This is a fallacy because the specter of single-registrant use is
likely to delay the New gTLD process for years. Moreover, the most
likely result after these delays would be a severe resource bottleneck
followed by a decision to prohibit or strongly discourage further
single-registrant TLDs.

Werner Staub

<<< Chronological Index >>>    <<< Thread Index >>>

Privacy Policy | Terms of Service | Cookies Policy