Another Transition Plan Proposal

Eric Brunner-Williams ebw at
Wed Dec 16 19:25:06 CET 2009

[Apologies in advance for length, and policy, and even meta-policy 
content, however, as both Cary and Vint observe, the contributors, 
even the rare contributors, to the idna-update list, include policy 

My concern starts with "in theory", and then degenerates into "in 

In theory, the ICANN structure, with contracts between DNS registry 
sponsors, such as PIR and PuntCAT, and ICANN, should provide a full 
and productive mechanism for operators to inform contract managers on 
operational issues.

The difference between Verisign, NeuStar and Afilias, all integrated 
ICANN contracted parties and DNS registry operators, and CORE, the 
operator of a similar number of DNS registries as each of Verisign, 
NeuStar and Afilias, albeit smaller, and differently policied, though 
not an ICANN contracted party (save only as a registrar, a status not 
relevant here) should not give rise to problems, though CORE is 
excluded from all contracted-parties-only meetings, including those 
with ICANN, which inform, in part, the ICANN DNS IDN Guidelines 
drafting process, as well as other policy development processes.

CORE receives no notice of ICANN's Registry Consitituency, now 
Stakeholder Group, meetings, no agendas of those meetings, no minutes 
of those meetings, nor any opportunity to participate in that ICANN 
created entity, distinct from that offered "the public". CORE is 
merely "an observer".

The group of representatives of TLDs which ICANN has gathered for 
devising and implementing policies and procedures for DNS IDN 
registration does not directly include CORE, though it may include 
other operators. The same result is obtained for registry continuity, 
and other registry implementation, operations, and policy groups 
formed by ICANN from a subset of the gTLD registry operators.

And of course, in theory, no harm should arise from this situation. 
Save perhaps the possibility that a party seeking ICANN contractual 
status will see utility in selecting an operator from those which are 
also party to a registry contract, and either not select CORE, all 
other factors being reasonably equal, or, transfer from CORE to a 
competitor possessed of this "in theory" non-distinguishing attribute, 
as SITA did with .aero, after its successful launch by CORE.

In practice however, there are issues of greater substance.

Operators of contracted registries have an interest in IDN that 
transcends the operational, and the marketing interests of a registry 
contract holder. The operator's technical capability is not limited by 
marketing constraints, or contract. A party seeking to apply for a 
string in Arabic script may have no interest in CJKV IDN issues, and 
visa versa, and therefore the "contracted parties" channel of 
communication is composed of, with the exception of Verisign, NeuStar 
and Afilias, which have pluri-script interests, and ICANN, which also 
has pluri-script interests, the union of several limited script interests.

Restated, the fora for global contractual DNS IDN issue scope is 
limited to a fixed set of global-scope interested parties, and an open 
set of local-scope interested parties.

Still, this should, in theory, result in no competitive advantage, and 
no institutional non-function arising only from partitioned and 
restricted information about DNS IDN issues.

However, the universe of technically competent operators is much, much 
larger than the universe of operators which are directly, or 
indirectly, party to a contract with ICANN. The .uk operator may be 
selected by an applicant for "鸭汤" (yā tāng) (trans: "duck soup", an 
example used in the IDNG list, which for the moment CORE is allowed to 
participate in, but only through its standing in the ICANN created 
Registrar Constituency or Stakeholder Group, not by virtue of its 
being one of the four, multi-registry operators of the contracted 
registries), and so on.

In the near future, the universe of technically responsible operators 
of contracted registries will be much larger than at present, even 
when anti-competitive practice is assumed, and assumed to be 
effective. The unique situation of the present, where only one gTLD 
operator is systematically excluded from participating in DNS IDN 
policy formation which has a direct impact on operations, and on 
security and stability, will be the general situation.

Intentionally placing the locus of responsibility for IDN transition 
in a venue restricted to contract managers, of whom three have 
unrestricted access to operational and economic data, is an accident 
waiting to happen. It works at present, it may not in the future.

My purpose this morning is to restate the case for operational equity 
in a regime which is contractual. If DNS IDN, and also DNS security 
and stability, and similar DNS issues, are issues to be addressed by 
contracted parties only, then the preconceptions of lawyers, not 
engineers, will take precedent. We are fortunate that Cary and Tina 
have the responsibilities they have, but ICANN is technically thin 
with a much greater emphasis on legal form and unilateral contract 
than on technical substance and cooperation as the basis for policy 
development and interpretation.

The means by which any DNS IDN guidelines are produced should be 
intentionally, not accidentally, technically well-informed.

This is not an objection, but a caveat that the transition question, 
moved as proposed, could, in theory, silently fail to produce a 
competition neutral outcome. My hope for the actual practice is that 
it will succeed in producing a competition neutral outcome.


More information about the Idna-update mailing list