Nominalia's Comments on Accreditaion Guidelines

[Amadeu Abril i Abril <Amadeu@nominalia.com>]

RESPONSES TO THE QUESTIONS CONCERNING GUIDELINES FOR ACCREDITATION SUBMITTED
BY NOMINALIA

NOMINALIA respectfully submits its responses to the questions raised by the
Internet Corporation for Assigned Names and Numbers in its draft of Guidelines
for Accreditation of Internet Domain Name Registrars and for the Selection of
Registrars for the Shared Registry System Testbed for .com, .net and .org
domains, dated February 8, 1999

Q1.  Is the segmented model for the provision of domain name registration
services described above appropriate?  How could or should it evolve over time?

Yes it is. But We are nevertheless seriously concerned that in the current
situation the registrar segment is being somehow overregulated while the
registry segment is not. The registry function being a sort of natural
monopoly, it would be desirable to make it as benevolent a monopoly as we
could afford. 

We only see two possible ways of achieving such goal: carefully regulating and
closely overseeing it, with the huge administrative costs we can imagine and
the real risk that this will not be enough to control the rational behaviour
of a rational monopolist: exploit its dominance, and hence make the pretended
competition among its customers, the registrars, completely moot as they would
lack the degree of freedom that would allow passing the advantages of a
competitive model on to their customers, the domain name registrars.
Alternatively, the registry function, given its intrinsic monopolistic nature,
can be designed as a service to its users, the registrars, and therefore be
channelled through an open membership organisation where all accredited
registrars could participate if they met some objective criteria applied in a
non-discriminatory way. Registrars' selfish interest would keep the monopoly
as benevolent regarding themselves as they could afford while their large
number would make collusion at registrar level near impossible.

It is clear that our preference is for the second model, iv nothing else given
the limited resources ICANN would have to oversee registries and our concerns
about guiding large-scale overexpensive bureaucracies.

Q2.  Are there any reasons why resellers should be accredited?

No. Registries should be regulated and adequately overseen. Registrars should
meet objective accreditation principles. If carefully drafted this set of
principles should adequately protect both Net-stability concerns and customer
protection issues. Regulating further down the chain would not incurring any
added benefit and should therefore be avoided.

Q3.  In the context of accrediting registrars for the DNS system, is
preservation of universal and durable resolution of domain names an important
requirement for preservation of the Internet's stability?  Are there other
stability issues that are particularly relevant in that context?

We agree that this is the most relevant requirement.     

Q4.  What conditions, other than the availability world-wide of a reasonable
number of alternative registrars and easy portability of domain names, are
important to the introduction and maintenance of robust competition in the
provision of registrar services?

There is one critical aspect besides those mentioned above. Registrars might
easily become captive of their unavoidable monopolist provider: the registry.
It should be seriously considered that in the future registrars could jointly
contract the operation of the (mostly technical) functions of the registry.
Through open-membership associations, cooperatives or directly and jointly
contracting with the registry.

Furthermore allowing the registry to simultaneously act as registrar puts the
competitive structure of the registrar market at risk, and this option should
also be reviewed in the future

Q5.  Are there significant principles in addition to those stated above that
domain name registrar accreditation should seek to promote?

No.

Q6.  At what stage should requirements for accreditation of registry
administrators be introduced?

As soon as legally and technically possible. The registry function is at the
very heart of the whole system. ICANN develop principles for accreditation and
apply them, also for the addition of new gTLDs. As already stated, we support
that the guiding structural principle of such accreditation be based on
open-membership; non-discriminatory; cost-recovery criteria.

Q7.  Should the requirement (Section III.A.1) that applicants demonstrate
general present or future business capabilities be revised to incorporate more
specifically stated thresholds or characteristics (e.g.,  levels of employment
or backup procedures)?  For which aspects of Section III.A.1?  
What should the more specific requirements be?  If more specific thresholds or
characteristics should be incorporated for some aspects, should those
thresholds or characteristics be absolute minimum requirements, or should they
be "safe-harbours" that would automatically satisfy certain aspects of Section
III.A.1?  (As an example of a "safe harbour," a showing that the applicant has
the equivalent of five full-time employees could be deemed to satisfy the
employee requirement stated in Section III.A.1.j.)

The principles seem adequate, but its operationalization will prove really
difficult. The desirability of having capable, stable, skilful registrars
might lead to financial constraints that would keep entities form large areas
of the world outside the system.

Again we do have the feeling that registrars are being overregulated while the
registry function is being clearly underregulated.

Q8.  Is it appropriate to have a specified threshold of required commercial
general liability insurance coverage?  A requirement of US$500,000 is being
considered.  Is that level about right, too high, or too low?  Should the
required level vary by the location of the registrar to account for different
liability risks in different parts of the world?

Nominalia went through a similar process when joining CORE. The 88 members
that joined CORE proves that this and the remaining criteria were reasonably
well fit for a large number of companies and organisations, the absence of
registrars from less-developed countries also showed that this kind of
thresholds might be prohibitive in certain parts of the world. 

Q9.  Is it appropriate to require that a specified level of liquid capital be
available to the applicant, or is it sufficient to have a more general
requirement such as that in Section III.A.1.i?  A specified level of
US$100,000 in liquid capital is being considered.  Is that level about right,
too high, or too low?  What effect would a requirement at that level have on
the degree of competition?  Should the required level vary based on the
country in which the applicant is located?  By the type of business model the
applicant follows or proposes?

A liability insurance policy should be required in order to give users more
security in the system, and therefore more credibility to the registration
market. But see also comments above on the risk of excluding registrars form a
large number of countries.

Q10.  Are the grounds for ineligibility set forth in Section III.B adequate to
protect consumers and the registry administrator from unscrupulous registrars?

Yes, in general. This and the obligation to maintain certain data in escrow
for cases where the registrar vanishes, or simply abuses customers combined
with  healthy competition should be enough.

Q11.  Are any of the grounds for ineligibility set forth in Section III.B
likely to significantly diminish the number of applicants eligible for accreditation?

No.

Q12.  ICANN presently has not concluded whether its accreditation
responsibilities will be performed by its own staff or by a qualified outside
agency.  Does a decision on this issue have a potentially significant impact
on the proposed accreditation process?

We ignore whether ICANN has the staff and resources to carry this process
internally. In case an outside qualified agency is chosen, on a cost-recovery
basis if possible, reporting and review mechanisms have to be carefully
established beforehand.

Q13. Where the registry administrator is also a registrar, is direct
accreditation of registrars an appropriate role for ICANN, or should
accreditation flow through the registry administrator? 

Nominalia, nor any other responsible registrar, would participate in an
accreditation process carried directly by a direct competitor (a
registry-cum-registrar entity). We encourage ICANN to run all necessary
accreditation processes within the gTLDs.

Q14. If direct accreditation is necessary to address competitive concerns
raised by the registry administrator also acting as a registrar, do the draft
guidelines provide adequate protection?

Certainly not. The history of antitrust enforcement both in the US and in the
EC clearly shows that there are NO adequate protections in such cases (where
the upstream monopolist is also a "player", inevitably the dominant player in
the downstream captive market). Even intensive, expensive, public oversight
has failed. To say the least, this is a situation where the public interest is
not well served. 

>From a competition point of view, we have already stated that the best
solution is having the registry acting as an open-membership; cost-recovery
function for all its ·customers" (the registrars). Short of that, it is
inconceivable that a for-profit company being in an unavoidable monopolistic
situation is allowed to compete downstream. The problems created by such
situation cannot be solved by these accreditation's guidelines or ICANN for
that matter. It is a decision  taken by the USGov when drafting the so-called
Amendment 11, unfortunately forgetting all the recent lessons in similar
situations. We hope this position will be revised in the near future.

Q16. What special measures, if any, should be taken to monitor the registry
administrator's activities to ensure that all registrars get equal access to
the registry?

This will undoubtedly involve a large amount of oversight by ICANN: data
provided to registrars and data denied; quality of services provided to
different registrars. As sad before, antitrust history clearly shows that this
model is unmanageable per se, independently of the relative good faith of the
parties. 

Q17.  Are the data elements to be submitted by registrars to the registry
administrator appropriately specified?

Yes. And as the registration market develops and renewals become an integral
part of it, such data as the date the record was created will become of
critical importance and an issue ICANN will need to closely monitor.

Q18.  Is monthly an appropriate period for data escrow by accredited
registrars?  Should more frequent or incremental escrow be required?

It could be as a start, but we suspect that a per-volume measure will prove
more adequate as the system evolves.

Q19.  Should all registrar electronic records concerning dealings with
customers be subject to a data escrow requirement?

It seems an adequate solution if it is carried in a manner that does not
conflict with the diverse applicable privacy laws.

Q20.  Should registrars be prohibited from claiming rights in the registration
data they generate?

The key concept in a competing domain registration market is "portability".
Short of making this concept fully operational there is no such competition
possible. In this context, allowing the registrars to claim intellectual
property rights on such data will kill any shadow of competition. Same thing
applies, but for different reasons, to registries.

Q21.  Should registrars' claims to rights in registration data be subject to a
Whois license for ICANN-designated Whois servers only, or should registration
data be freely available to all operators of Whois servers?  What other
restrictions, if any, should be placed on the data?

Registration data should be freely available to all WhoIs users, with the only
restriction of privacy (data protections) legislation.

Q22.  Are the proposed requirements in Section IV.7 appropriate to promote
fair access to the registry?  Should other similar implementing requirements
be set forth?

Creating a level playing field in a situation in which a registry is also a
registrar is near impossible. But the Section IV.7 addresses the right
principles. But "equal treatment" should be extended form access to data or
SLDs to economic (financial) conditions..  

Q23.  Does the uniform application of the fair information practices
concerning personal data impose undue burdens or restrictions on registrars or
the registry administrator?  Should greater protections be required for
personal data provided by domain name holders?

We welcome the approach taken by the accreditaion guidelines on this issue.
Not all the personal data that is need in order to make the registration
workable needs to be available to the general public, but it is of critical
importance that sufficient contact data, personal or not, be available in case
of technical or legal conflict.

Q24.  Does the proposed mechanism for anonymous holding of SLDs adequately
ensure that accurate contact information is available as required for proper
operation of the Internet?

Yes.

Q25.  Does the proposed mechanism for anonymous holding of SLDs adequately
accommodate the legitimate interests of law enforcement authorities and those
seeking to enforce trademark and other rights?

Yes..

Q26.  Is it advisable to have a Code of Conduct that establishes basic
standards of fair business practices to which accredited registrars must adhere?

Within CORE has favoured this approach, but we have always found such a
mechanisms difficult to implement and often dangerous form a competition point
of view. But we are committed to participate in its development, if needed.

Q27.  Should a registrar/registry administrator organisation have the
principal responsibility for formulation of the Code of Conduct?  Should that
responsibility instead be assigned to ICANN or some other body?

This is, as stated above, a difficult issue. For which we don't have a
conclusive answer.

Q28.  What standards of conduct should be in the Code of Conduct?  A proposed
Interim Code of Conduct for DNS Registrars, to be used until the organisation
of registrars and registry administrators becomes operational, is currently
being prepared and will be posted before the Singapore meeting.  Early written
comment on this issue would be especially appreciated. 

Unfortunately we are unable to provide detailed comments at this stage, but we
reiterate our commitment in helping ICANN in exploring this possibility.

Q29.  How, and by what entity, should the Code of Conduct be enforced?  Should
enforcement decisions be subject to review and, if so, by what mechanism?

See above.

Q30.  Should the guidelines limit the duration (e.g., two years) for which
SLDs can be registered?

There are some good reasons for doing so, and two years seems a good start.   
 

Q31.  Would the prepayment requirement hinder any otherwise realistic and
desirable registrar business models?

We believe that there is large consensus that prepayment is a need. Even if it
might make registrar's life more difficult....     

Q32.  Should the durable connectivity assurance (against insolvency of
registrars) be limited to a specified length of registration?  

The assurance could be slightly longer than the registration  length.

Q33.  Should other WIPO recommendations be implemented in these proposed guidelines?

This issue will be better addressed once WIPO delivers the final report.

Q34.  Because there has not yet been any experience in processing
applications, the proposed application fee levels are necessarily rough
estimates.  Comments on the appropriate fee level to defray processing costs
are invited.

Application fees should be based on cost-recovery principles. Adequate or not,
USD$2500 seems rather high, and possibly prohibitive for entities in a large
number of countries..

Q35.  Are there any practical problems presented by collection of
per-registration fees at the registrar level?

The only practical question is : why is this per-registration fee collected at
the register and not the registry level?. More than practical, the problem is
one of principles: does this amount to an Internet tax? Why should registrars
pay by registration, instead of contributing to ICANN's expenses based on
ICANN's needs as set in the budget?

Q36.  Is it appropriate to have both fixed and variable components of the
accreditation fee?

See answer below.

Q37.  At what level should the fixed component be set?  Should the level vary
based on the country in which the registrar is located?

At the level of the need for contribution that ICANN's activities might
require form registrars. Differentiated fees (positive discrimination) has a
strong appeal in  this context., And a strong drawback: the extreme facility
that net activities provide for "offshore flagging". 

 Q38.  Is there some measure other than registration-years on which the
variable component should be based?

If there has to be such mechanism this measure seems the more appropriate. But
see comments above.

Q39.  Is it beneficial to state a cap on the rate at which the variable
component is computed, to allow registrars to better assess their prospects in
the business? 

This would be absolutely needed. 

Q40.  Are the stated grounds of termination appropriate?  Can they be made
more specific while still preserving their utility?

"inaccuracy" should not be an automatic motive for termination, unless a
period to cure is granted, and this being more than the stated 15 days (which
by the way are less than what registered mail will take to deliver documents
form most parts of the world). 30 days seems the minimum.

Q41.  Should additional grounds of termination be added?

Not at this time.

Q42.  Is fifteen days notice sufficient to allow a registrar to initiate arbitration?

30 days would be better. 

Q43.  Is there any type of dispute arising from a registrar's accreditation
agreement that should not be subject to arbitration?

No.

Q44.  What arbitral body should be selected, and what rules of arbitration
should be used?

A specialised international arbitration court, such as UNCITRAL.

Q45.  Should registrars seeking renewal of their accreditation be specially
protected from the effects of raised accreditation standards or changes to the
terms of the accreditation agreement?

The same criteria should apply in principle to all registrars. And we firmly
believe that, over time, and possibly after a relatively short period,
accreditation criteria will be less stringent, not more. Competition will make
the selection, and ICANN will only need to bother about customer's protection
(through mechanisms like data escrow).

Q46.  Should disputes arising from actions on initial or renewal applications
for accreditation be subject to arbitration?   If so, on what grounds should
such actions be reviewable?

Absolutely. In this global environment only arbitration through international
bodies is an acceptable solution.

Q47.  Should the guidelines prohibit accredited registrars from charging
different prices to different categories of customers?

No. Registrars will be in a highly competitive market and this will adjust any
possible abuse.

We insist that this is not at all the case for registries, and it is here
where, in absence of a cost-recovery model guaranteed by open membership to
all registrars price regulation will be of critical importance.

Q48.  Should the guidelines require accredited registrars to offer unbounded
domain name registration services?

If the market proves to be competitive, ICANN should refrain from imposing
such restrictions. Except perhaps to the dominant player.

Q49.  Are the additional requirements for participation in phase 1 too burdensome?

We strongly believe that the testbed experiment makes no sense. But we
acknowledge again that this is not something that ICANN can freely decide, but
a mistake made during the negotiation of the Amendment 11.

Govern that mistake, we understand ICANN's interest in devising more stringent
criteria in order to limit somehow the access to the test. No matter how
objective or fair those criteria could be, they don't seem to be warranted by
any need other than the mentioned policy mistake.

Q50.  Is the primary selection criterion appropriate?

Yes, given the circumstances.

Q51.  In the event that more that five registrars meet the criteria equally,
what mechanism should be used to select among them?

Ask the USGov to modify Amendment 11 and let them all in.

In case this fails, try to devise (beforehand) a set of ·diversity· criteria
(regarding size, type; cultural area served...) principles that would help
making this experiment a little bit more useful, and select the registrars
according to those criteria or slots. But we acknowledge that this will never
sort all the conflicts. So we are back to the first option: if a rule leads to
unsolvable situations, it is better to modify the rule.



Amadeu Abril i Abril
Legal & Policy Advisor     
Nominalia
http://www.nominalia.com