Approval process for gtld service changes

[Marc Schneiders]

On Thu, 8 Jan 2004, at 13:19 [=GMT-0500], Milton Mueller wrote:

> Feel free to suggest modifications --MM

Later than I wishes some remarks in between the lines.

> =====
> Proposed NCUC statement on
> "Procedure for use by ICANN in
> considering requests for consent and related contractual amendments
> to allow changes in the architecture or operation of a gTLD registry."
>
> NCUC observes that by initiating this procedure, ICANN is assuming that
> its contracts alone are not sufficient to provide a predictable and stable
> basis for the industry. It is assuming that it needs an ongoing form of
> regulatory oversight to supplement its contracts. ICANN should face the
> fact that it is expanding further into areas of industry regulation, although
> no one wants to admit that.

I am not completely sure, this follows. The contracts can obviously
not deal with every development possible in the future. Some changes
may require a modification of the contracts, other could be dealt with
in a more light weight manner. The PDP tries to develop such a light
weight method.

> In formulating its response, NCUC begins by asking: How are
> noncommercial domain name users specifically affected by this change?
> The answer, NCUC believes, is that there is no commercial/noncommercial
> angle to this issue. It is more a question of:
>
> a) how users/consumers relate to suppliers and what kind of regulatory
> procedures are needed to protect consumers given the high switching
> costs associated with changing registry suppliers after a domain name
> is well-established.

So 'consumer protection'.

> b) what kind of technical regulation or specifications are needed to protect
> third parties using domain names on the Internet from harmful changes
> made in registry operation, while preserving as much as possible the
> freedom of suppliers to respond to the market and innovate.

So 'stability'.

> NCUC notes that whether consumers or users are commercial or
> noncommercial has little bearing on these issues.
>
> We also note that a) and b) are distinct policy issues - a) involves
> protection of the parties buying service from a gTLD registry, who may
> have options, while b) involves protection of third party users of a domain
> name, who probably do not have any options if they want to connect to the
> party using the affected registry. We also note that a) involves economic
> forms of regulation which also involved competition policy concerns, while b)
> is more a matter of technical coordination.
>
> NCUC strongly recommends that the PDP distinguish clearly between
> a) and b) in its consideration of the new process. Is the object
> of the process economic regulation or technical coordination?
>
> The document we are asked to comment on proposes no policies,
> so our comments can only suggest questions or problems for the
> PDP process to consider.
>
> 1. One question the PDP should consider is whether all issues
> related to a) above should be handled by national regulatory
> authorities instead of ICANN. We support ICANN's need for
> technical coordination related to matters under b). We are less
> confident of ICANN's ability and right to engage in a). We are also
> not convinced of its ability to engage in competition policy-related
> forms of regulation.

I am not that convinced either. But I do not see how national
regulatory authorities can take care of consumer protection. How can a
Dutch registrant handle an issue with a registry located in the US?
Dutch bodies will not be able to help him. So he has to find one in
the US that helps him? If so, this brings up the topic of the US
centredness of ICANN again. This is a problem we need to acknowledge
in this context.

Plainly: We may not like ICANN getting more power. But who else is
going to protect us against cases like PIR deleting IDN domain names
without even so much as telling the registrants? To me this looks like
a jungle. And if I want to do something about it, I have to hire a US
lawyer. I think this is just bad.

> 2. The PDP document refers to a "quick look" process followed by
> a more involved process if a change fails the "quick look."
> A question the PDP needs to face squarely is: What is a subject to
> a "quick look" and what is not? What is a "new registry service"?
> How is that defined? Who will make that determination initially?
> What happens when the registry and ICANN disagree on that issue?

Plus there should be guidelines as to when an issue is important
enough to put it before ICANN bodies, invite public comment etc. Some
issues will be too important to leave them to ICANN staff.

> 3. The NCUC recognizes the danger that a registry can make
> damaging changes, such as in the Sitefinder case. We support
> clear, well-defined specifications for registry operation that make
> DNS a neutral platform for Internet functions.

This is hard to do in a water tight way. Who knows the future?

> We also recognize
> a threat that innovative changes, such as multilingual domain
> names, will be stifled by a central organization such as ICANN
> which may have incentives to prevent useful changes in order to
> maintain its control over the industry.

Yes, sure, but this is in fact NOT what happened. On the contrary.
ICANN did not regulate anything as regards IDN. And now we are left
with the mess:

1. PIR takes over IDNs, lets them work for some months and says it
will not accept new ones. After 3 months it secretly drops them from
the DNS. It refuses answers as to what happened and why. And after 13
months (in 3 weeks) it will secretly again drop them alltogether.
Again without informing us. I am sure PIRs lawyers have studied the
contracts. The contracts will not give us a handle to fight this.
The contracts do not protect users. They protect PIR and ICANN.

Even though it is dangerous I'd like ICANN to protect users. Nobody
else does.

> 4. The PDP should consider whether there should be a distinction between
> policies applied to sponsored and unsponsored TLDs. NCUC believes
> the answer is no: if the justification for regulation is economic; i.e,
> that users are locked in to a supplier and cannot switch service providers
> without incurring damaging costs, then the same fundamental economic
> problem applies regardless of whether the registry is sponsored or not.
> If the justification for the process is technical, the answer is the same:
> there is no relevant technical distinction between sponsored and un-
> sponsored registries.
>
> 5. The PDP should consider whether there should be a distinction
> between the treatment of dominant and non-dominant TLDs? In this
> case NCUC believes there is a stronger case for a distinction. A major
> dominant registry may have the power to move the entire industry and
> technology, whereas smaller ones would not. However, the lock-in
> problem of consumers applies regardless of whether the registry is dominant
> or not.

I assume behind 4 and 5 is the fact that Verisign was told to stop
Sitefinder, while other TLDs use a similar DNS trick? I am afraid we
cannot formulate in advance (since we are not clairvoyant) what
distinctions to make. It depends on each registry change in context.
So I would not make any distinctions except to explain that all
registries are not equal, and should therefore not be treated equally,
except when fair competition is involved.

> Thus, the policy must identify carefully what problem it is trying
> to solve.

Yes, I think that is the most important thing: What are we trying to
accomplish? Not: formulating a bureuacratic process, but dealing with
issues like SiteFinder and IDNs. Any procedure that comes out of this,
should work for these two issues. It should be clear how it would have
dealt with these in a better way, actually how it would have made sure
these incidents could not have happened.

Some additional remarks for our statement:

1. The new procedure will only work, if the registries will cooperate
in a friendly way. They have to submit requests. And do so in time,
not at the last moment. Example SiteFinder: Registry can ask ICANN
long before it wants to introduce it, while it is looking into it
itself.
This will then not mean extra delay for the introduction. Since there
is a possibility for confidentiality, there is no reason not to ask in
advance.
2. There is an advantage for registries. Costs would be much lower for
Verisign, if it had asked about SiteFinder.
3. We should emphasize that public consultation, and consultation of
constituencies, must be a regular part of dealing with more important
issues. We do not want to be excluded and let ICANN staff handle
substantial issues completely.

>
>
>
>
>