Comments and a detailed policy proposal on vertical separation

[Vittorio Bertola <vb@xxxxxxxxxx>]

Hello,

the report has several good points, but only considers a few possible
TLD registry business models, basically two:
1) the classical "selling registrations for profit" model, ie .com;
2) a single corporate registration, either for internal use or for
defensive purposes.

It fails to consider other models for running a TLD, for example those
aimed at small communities or not for profit.

In these cases, the burden of having to cope with ICANN-accredited
registrars might simply kill the registry, while registrars themselves
(as noted by the report when discussing smaller registries) might have
little interest in reselling or promoting the TLD.

Also, for TLDs aimed at specific geographically-defined communities,
especially if the TLD is an IDN, there might be a serious scarcity of
ICANN-accredited registrars offering services in that part of the globe,
or there might even be none.

So, by keeping the vertical separation requirement intact for
non-corporate TLDs, the risk is that smaller TLDs, especially IDNs and
non-profit, will not be able to function due to lack of suitable and
committed registrars.

Even if registrars could be found, the registry will need additional
expenses to cover the costs of integrating registrars in the operations
and in the technical platform; these expenses will actually result in
higher prices for the registrants, rather than in lower ones (which is
the objective of requiring vertical separation). Furthermore, registrars
will add a margin and increase the final price; when no competition
exists due to lack of registrar and registrant interest, this increase
will not be compensated by the effect of competition. In these cases, as
recognized by the report in appendix A, adding a registrar to the value
chain is likely to be counterproductive to competition and
inexpensiveness of the service, and actually increase the final price.

So I would encourage the report to be integrated with the study of other
business models, and ICANN to consider a policy like the following one:


1) In case the TLD
   a) has less than 20'000 registrations, or
   b) gives away domain names for free and without forcing registrants
to buy other services with them,
the registry can choose to act as sole registrar for the TLD, and manage
the registrations directly.

2) In case the TLD does not fall under 1), but
   a) has less than 100'000 registrations, or
   b) there are less than four ICANN-accredited registrars willing to
sell it,
the registry can still choose to act also as registrar, but it is
required to accept and serve any ICANN-accredited registrar willing to
sell the TLD.

3) In case the TLD does not give away domain names for free, has more
than 100'000 registrations, and there are at least four ICANN-accredited
registrars willing to sell it, the present policy of registry-registrar
separation would apply, and the registry would not be allowed to act as
registrar any more.


This proposal would satisfy the need of several business models: the
effort in vertical separation would be proportionate to the size and
revenues of the TLD, thus avoiding to impose the use of registrars when
they would be economically ineffective and just add unnecessary costs to
the process, but continuing to promote competition for all TLDs of
medium and big size. Requirements would grow harmoniously with the
growth in size of the TLD, thus avoiding to overload newborn registries
with heavy operational burdens. The proposal would also keep into
account some extreme cases where the current and proposed policies would
get stuck, including big TLDs for which no suitable registrars existed.

This proposal would also address the need to differentiate "single-owner
TLDs" without entering into the hardships of defining what they are; in
fact, any corporate registration for defensive purposes or for internal
use would likely fall under both 1.a) and 1.b) above, or at least under
2). The only case in which a "single-owner TLD" would be required to
adopt full vertical separation is when it gives away domains for a price
and has more than 100'000 registrants - but in that case there would be
clear merit to introduce competition.

I hope that my proposal can be given due consideration. In any case, it
is hard to tell the difference between a single corporation giving away
domains for free to its employees, and a community entity running a
non-profit operation to give away domains for free to members of the
community. If the first one is exempted from using registrars, then the
second should be as well.

Regards,
-- 
vb.                   Vittorio Bertola - vb [a] bertola.eu   <--------
-------->  finally with a new website at http://bertola.eu/  <--------