RE: [registrars] White paper on new gTLDs from Philip Sheppard et al

[Tim Ruiz <tim@xxxxxxxxxxx>]

Ross, 

I want to digest your comments more fully, but have a few initial
comments/observations.

> ... abolishing the current distinction between gTLD and sTLD registry operators ...

In general I'd agree with this.

> A gTLD registry operator accreditation program.

Perhaps this should go a step further and keep registry accreditation
completely separate from TLD operation. What I mean is that registry
accreditation would not require a TLD delegation (our friends at
CentralNIC for example), but a TLD delegation would require use of an
accredited registry. Or is that where you were going anyway? If so, I
think I would agree with that as it would help focus any new TLD
allocation process on that process itself (whatever that turns out to
be).

Tim

 
-------- Original Message --------
Subject: Re: [registrars] White paper on new gTLDs from Philip Sheppard
et al
From: "Ross Wm. Rader" <ross@xxxxxxxxxx>
Date: Fri, July 29, 2005 6:29 am
To: Bruce Tonkin <Bruce.Tonkin@xxxxxxxxxxxxxxxxxx>
Cc: registrars@xxxxxxxxxxxxxx

>>I am also concerned with viability. Registry failure is going
>>to happen sooner or later. But there is no point in inviting
>>it by only requiring certain financial and technical
>>requirements for a new TLD. That does not do much for
>>promoting a stable and secure Internet. I think there also
>>needs to be verifiable market research that indicates
>>viability, and clear and well defined marketing plans from
>>any applicant.
>>
>
>
> One approach to a viability test could be to require that a proposed new
> TLD, first establish a second level TLD and show that it is viable in
> terms of number of registrations and business model (ie earns sufficient
> revenue to cover full operating costs - to avoid just offering a large
> number of free registrations initially).


Viability is extremely difficult to appropriately quantify. Typically, 
any attempt to demonstrate market viability results in very compelling, 
but pretty useless, fiction.

Registrars need to advocate for gTLD registry competition and not 
unintentionally limit the number and types of registries that enter into

the market. The surest way to get a fair deal registries is to ensure 
that there is a competitive market for gTLD registry services.

Rather than imposing artificial hurdle rates or pursuing a "full 
employment for intellectual property lawyers" program we should be 
seeking to define what we mean when we talk about gTLD registry 
competition and what we need to do to see that definition implemented in

policy.

For instance, within ICANN, when people talk about gTLD registry 
competition, they talk about some vague choices that registrants make 
when choosing a string. To me, this is bollocks. Historically, we only 
see competition between gTLD registries when a gTLD registry contract 
bid process has been imminent.

Competition can be deemed as being successful when it results in more 
products and better products for consumers. In the case of gTLDs, we 
need to carefully consider who the consumers are. There is an assumption

in most ICANN discussions that the consumer is the registrant. However, 
this is only true in a very limited sense that the registrant is the 
recipient of the finished product that registries manage. On a 
day-to-day basis, Registrars are the primary consumer of registry 
service - we interface with their technical systems, we consume their 
customer service, we deal with their inefficiencies, problems and gaffs.

End-users simply benefit from the output of this process.

Put another way - Registrants are the primary beneficiary of gTLD 
Registrar competition. Registrars are the primary beneficiary of gTLD 
Registry competition.

gTLD Registries do not compete thoroughly or vigorously with one another

today because of the regulations that the community has imposed on them.

We imposed these restrictions on their activity because of the limited 
monopoly that they exercise over the core registry. Over time, these 
regulations have become more onerous, imposing and meddlesome to the 
registries, so they have sought to lessen the burden of these 
restrictions by seeking relief on such items as price caps, compliance 
w/ consensus policy and so on. These actions should be interpreted as a 
request by the gTLD registries indicating that they wish to operate in a

more market driven manner. That they wish to be released from some of 
their obligations and let the markets sort out the winners from the
losers.

I believe that this arrangement will actually benefit registrars over 
the long run, provided that there are a set of minimal checks in place 
to ensure that gTLD registry competition is actually implemented in 
return for lifting some of these restrictions.

Logical checks and processes that we need to implement are:

1. Standardized contracts between ICANN, the registries and registrars. 
Competition presumes a larger number of registries than we have today. 
Using custom contracts, as the registrar experience indicates, does not 
scale. This presumes abolishing the current distinction between gTLD and

sTLD registry operators as the goals of the latter program are 
inherently anti-competitive according to the definition I put forward
here.

2. Finite terms of delegation for TLD strings. Registry operators need 
enough certainty in the term of their delegation that they can 
demonstrate a reasonable rate of return to their investors. The 
community needs enough certainty that the registry operator can be 
replaced in the event that they aren't performing or operating in a 
competitive manner. Implementing contract terms of 7 to 10 years, after 
which the registry contract would be rebid, should satisfy both of these

requirements.

3. A gTLD registry operator accreditation program. Registry operators, 
like registrars, must meet certain technical, operational and financial 
standards. A program similar to the Registrar accreditation program that

allows a potential operator to demonstrate their capabilities, should be

implemented. Once accredited, a Registry Operator would be qualified to 
apply for gTLD delegations through a gTLD allocation process.

4. A gTLD allocation process. Registry operators also need predictable 
processes through which they can apply for and be awarded a gTLD 
delegation. This set of processes needs to be defined and implemented. 
This process usually gets the most attention when people discuss new 
gTLDs, but from a registrar perspective, the attributes of the process 
are largely unimportant so long as the process is fair, predictable, 
scalable and requires applicants to agree to the minimum requirements 
set out in this message. The method of allocation (auction, first-come, 
etc.) is equally unimportant so long as the process encourages a 
suitably large number of applicants to participate.

5. Additional operating requirements for the .com registry operator that

can be lifted when a certain level of competition is achieved and 
current market dominance has been tempered to a degree that satisfies 
the community. Price caps, budget contribution, etc. are examples of 
suitable additional requirements to expect from the com registry
operator.

We have a unique opportunity in front of us - not only is the community 
highly interested in the topic of new gTLDs, they are highly interested 
in the topic of the contracts the existing operators are working under. 
We should seek to promote our long-term interests by using this 
convergence to focus the discussion and implement a program that will 
result in the definition and implementation of a market that will ensure

that we get access to more products and better products. Right now, I 
fear that we are setting ourselves up for worse products at higher 
prices than what we could get in a truly competitive market.

-- 
Regards,


-rwr






"In the modern world the intelligence of public opinion is the one 
indispensable condition for social progress."
- Charles W. Eliot (1834 - 1926)