[comments-gtlds] Comments on the interim report of DNSO Working Group C (New gTLDs)

[HiroHOTTA <h.hotta@hco.ntt.co.jp>]

Comments on the interim report of DNSO Working Group C (New gTLDs)

0. Introduction

  The following are the opinions on the addition of gTLDs discussed by 
a group of volunteers from JPNIC.  Our discussions were held from the 
point of view of "what we should do to make the Internet better".  Here, 
it would be appropriate to define first what "better" means in our 
discussion standpoint.

  Our idea of "better" is as follows.
a. without damaging the stability of the Internet
b. responding to the needs of the market as much as possible

Here are the conclusions we reached in our discussions.
(1) Position Paper A seems to be a well-balanced proposal overall from 
    various points of view. However, revisions shown in (2) to (5) are 
    proposed.
(2) [II. What should the nature of the gTLDs be?]
    We don't take a standpoint disagreeing with limited-purpose TLDs.
    However, if we take into consideration whether we (or ICANN) could 
    check if they are operated according to the charter and whether we 
    (or ICANN) can create the strategies to prevent the proprietary use, 
    it might be extremely difficult to realize limited-purpose TLDs.
(3) [IV. What should the transition to an expanded namespace look like?]
    We agree with the introduction of a small number of gTLDs followed 
    by a period of evaluation.  However, as mentioned in Position Paper 
    B, it would be preferable to make public the number of new gTLDs to 
    be added in the future beforehand, and at the same time to show the 
    scenario for their addition in concrete figures in order to avoid 
    unnecessary confusion over the small number of gTLDs introduced at 
    first.
(4) [VII. What should ICANN's process be for selecting new domains and
     registries?]
    When we consider that the string of the gTLD itself has a value, the
    gTLD to be added should be applied for by registry candidates, not
    determined by ICANN beforehand.
(5) The next issues are also important, and deserving of due discussion.
    First, [Should the business operation of a given gTLD registry be
    transferable between registries?]  We believe that we need to use 
    "periodical forced transfer" as well, with the idea of 
    "transferability" as a basis.  Secondly, [What should the Definition 
    of gTLD be?]  We propose we should discuss the definition of "g" of 
    gTLD before changing it from "generic" to "global."

  This document is summarized by a group of volunteers within JPNIC, and
doesn't represent the policy of the whole members of JPNIC, or their 
policy as an organization.

  Sections 1 to 7 of the following correspond to the questions raised
in the WG-C report.  In sections 8 and 9 we suggest issues that should 
be discussed additionally.

1. Should there be new gTLDs?

  We believe they are necessary.  The ".com" in particular has almost
reached saturation point.  The economic value of .com as the commercial
tool has become very powerful, almost to the extent of being 
unreasonable.  It seems that very unfair sacrifices have been forced on 
enterprises that fell behind due to the pure first come first served 
rule.  Many enterprises use the same business names or trade names in 
different business fields, or the same trademarks in different lines 
of products and/or services in the world.  They are coexisting under 
those same business names or trademarks, and it seems that no actual 
problem between them and in public has occurred.  Proper arrangements 
are necessary to ensure that such enterprises, groups and individuals 
are given a new chance, and fairer competition is secured on the 
Internet.

  Regarding respect for trademark rights, the UDRP that realizes a 
simple, inexpensive and prompt dispute resolution was actually 
introduced in January, 2000, therefore we need to observe the 
circumstances in the future carefully.  However, we can say that a big
problem facing the introduction of new gTLDs has been removed for now.  
Although more well-known or famous trademarks tend to be the target for 
cyberpirates and a considerable burden is laid on the owners of the 
rights to eliminate cyberpiracy compared with others, we can not agree 
to the approach of holding back adoption of new gTLDs until a 
pre-emptive prevention system has been introduced.  We should promote 
concurrent advancement of UDR process maturity and introducing new 
gTLDs.  We think that we need to handle disputes utilizing the UDRP
during this period.  However, we consider more prompt disclosure of
information and development of simpler/cheaper unified search systems 
on registered domain names to be urgent tasks.

2. What should the nature of the new gTLD be?

  General-purpose gTLDs and limited-purpose gTLDs are a controversial 
issue in the WG-C report.  We don't have any objection to establishment 
of general-purpose gTLDs.

  Regarding the establishment of limited-purpose gTLDs, we can 
understand why some people are behind it.  However, we are all agreed in 
the opinion that it would be difficult to make a system for verifying 
whether it is used properly according to the charter of the gTLD.  To 
begin with, who would judge and how would they judge whether each 
charter was ideal for the development of the Internet.  This is a 
troublesome problem, but, even if a given charter is appropriate, 
annoying problems will be caused if operations are not performed 
according to that charter.  The situation that we are most concerned to 
avoid is where a given TLD is made proprietary in real term by the 
organization (or individual) who undertakes the registry business.  If 
the organization (or individual) were malicious, in real terms they 
might privately own a TLD by intentionally breaking the charter and it 
might then be used as the proprietary TLD of a private enterprise (or 
individual), not as a gTLD as originally intended.  In section 3, we 
will describe the dire consequences that would result if this were to 
happen, but anyway, we are in agreement that we can not allow the 
presence of proprietary TLDs.  Therefore, we can not agree to the 
establishment of limited-purpose gTLDs unless systems are created for 
judging the appropriateness of a charter and properly verifying that the 
charter is followed.

  However, no matter whether it is general-purpose or limited-purpose, 
it is crucial that a third party stores the registered data (data 
escrow).  This is important in order to prevent the proprietary use of 
the relevant TLD and also to prepare for disasters and bankruptcy of the 
registry.  However, we can not state definitively that this will be 
sufficient to prevent the proprietary use.  It is also necessary to 
clarify that the gTLD registered data and the gTLD itself belong to 
ICANN.

3. How many new gTLDs should there be?

  Compared with the time when the IAHC (International AdHoc Committee)
issued their Final Report, the technology situation has changed
considerably.  Therefore, it is probably appropriate to say that no 
problems will result even if a much greater number of TLDs than the 
number considered then are added.  As is pointed out in the WG-C Report, 
it is not recommended for the number of new TLDs to be limited 
artificially.  However, the fact is that the more the number of TLDs 
increases, the greater the load on the root server will be, and the 
stability of the Internet would gradually be compromised.  We could 
keep thinking up numbers of TLDs, how about if there were 500, or 1,000 
or 5,000 ..., but we can not draw a clear boundary anywhere, however, 
several hundred thousand TLDs would really be a nightmare.  It would be 
good if we could create a system in which market principles would 
operate and the number of TLDs would be settled down naturally at a 
reasonable figure, without imposing artificial limitations on the 
number.  We discussed what type of conditions would be necessary to 
create such a system.  For example, let's assume that an enterprise with 
the name of aaaaa owns a TLD named ".aaaaa" privately (in real terms).  
Then, another enterprise called bbbbb might hope to own a TLD named 
".bbbbb".  Once this happens, in a worst case example, many registrants 
of ".com" might feel that a TLD of ".xxxxx" was more attractive than 
"xxxxx.com" and rush to get one.  That is to say, we think that we need 
to be wary of the proprietary use of TLDs (proprietary TLDs) as it could 
lead to the nightmare of creating hundreds of thousands of TLDs.  We are 
all agreed that we should not allow this to happen.

  On the other hand, if we create a system that avoids proprietary TLDs, 
it should not be too difficult for market principles to work, allowing 
the number of TLDs to settle down to a reasonable figure naturally.

4. What should the transition to an expanded namespace look like?

  We believe that it is appropriate to consider keeping the speed of
expansion within the allowable range of ICANN processing capability in 
order to expand namespace smoothly.  In this sense, we feel that the 
plan in the WG-C Report for "adding 6 to 10 gTLDs at first, and then, 
investigating the addition of other gTLDs after evaluation" is 
appropriate.  However, if it is not clarified whether other gTLDs will 
definitely be added or not, we suspect that the problems shown below 
might occur.  That is to say,
  + The value of the small initial number of gTLDs might rise 
    abnormally, and there might be a rush of registry applicants and 
    applicants for domain name registration.
　+ There are fears that the vested interests of a gTLD determined at 
    the initial stage might be fixed.  Later addition of further gTLDs 
    might be obstructed and free competition might be disturbed in some 
    cases.
We believe it is necessary to determine and announce the time schedule 
and numbers of additional gTLDs and execute the addition according to 
the schedule in order to avoid such confusion.

  In this sense, we can appreciate the concrete scenario stated in 
Position Paper B as a concrete plan, which says that ICANN should 
declare that they intend to add 500 new gTLDs over the next three years, 
in which they will add 10 of them in the first 6 months, and 40 in the 
following 6 months, 150 in the 2nd year and 300 in the 3rd year (total 
500 in 3 years), and that there should be no limit to the increase in 
numbers after that.

  This means that our discussion can be summarized in the following two 
points.
　+ Add 6 to 10 gTLDs at first and after setting the evaluation period,
    reflect the result in the following gTLDs.
　+ ICANN announces the number of new gTLDs to be added in the future
    beforehand and shows the scenario with concrete figures.

5. Should ICANN require each new gTLD registry to be shared -- that is, 
   to support competing registrars on an "equal access" basis?

  All registries of general-purpose gTLDs should furnish a common 
registry system and provide service to multiple competitive registrars.

  When the number of gTLD registries increases, competition between
registries will result.  However, as each registry holds its own string 
as gTLD, competition between registries has a different nature to the
competition occurring between competitive companies that sell the same
product.  It is not difficult to imagine that this will mean more
registrations concentrated on the popular gTLDs and unpopular gTLDs will 
have a small number of registrations.

  If we wish to determine prices by market principles, we need to 
introduce the competitive principle in another dimension, without 
depending only on competition between registries.  Therefore, all 
registries of general-purpose gTLDs should furnish common registry 
systems and create an environment in which registrars can compete under 
the system.

  However, regarding common registry systems, there is no great merit 
for adopting different systems for each registry.  It is said that RRP
(Registry Registrar Protocol) is going to be standardized as an RFC 
document, so we consider that a system complying with the standardized 
RRP should basically be used by all registries.

6. Should ICANN require that each new gTLD registry be operated on a
   non-profit basis?

  Both profit and non-profit making have merits, so it is not necessary 
to limit ourselves to one or the other.  It is acceptable to choose one 
or the other according to the nature of the gTLD.

7. What should ICANN's process be for selecting new domains and 
   registries?

  The basis should be market-driven.  We think many registry applicants 
will feel a greater incentive towards a registry with a specific string, 
rather than feeling an incentive towards the registry business operation 
itself.  Therefore, we believe it would be preferable for each applicant 
to apply for registry together with the string of the gTLD that they 
wish to deal with.

  However, how will we resolve situations where several registries wish 
to operate the same gTLD?  This is a different issue.  We haven't 
reached a final conclusion on this, however, some were of the opinion 
that we should practice some sort of lot drawing for the sake of 
fairness while others said that submission of tender (bidding) would be 
appropriate, taking into account the idea of keeping things market-
driven.

8. Should registry business operation of a certain gTLD be transferable
   between registries?

  Aside from the 7 questions above, the marking issue is also an 
important one, which needs to be discussed intensively.

  Transferable or nontransferable business operations have their own
respective merits or demerits.

Merits of "non-transferability":
  ICANN will accredit binding of the registry business operation of a 
  given gTLD to a certain organization.  It can then set criteria to 
  accredit registries, and the range of organizations that can carry out 
  registry business operation including those with comparatively small 
  financial strength will be extended.  However, problems will occur 
  when the business operation of a registry can not be continued, or 
  when abuse of monopoly occurs if a gTLD acquires huge business power 
  like ".com".

Merits of "transferability":
  Transferability is the only way to handle cases where business 
  operation can not be continued due to bankruptcy of the registry, 
  etc., or when making business more efficient through M&A or business 
  operation transfer.  This is based on market-driven principles, and
  gTLD management will be consolidated in the more stable and efficient
  registries.  There is a possibility that this will result in abuse, as 
  there is a risk that gTLDs with large business power will acquire a 
  powerful monopoly.

Merits of "periodical forced transfer":
  It would be possible to change the organization in charge of registry
  business for individual gTLDs, for example every 4 years in order to 
  solve "monopoly" problems, which represent an abuse of 
  "transferability".  However, there is a risk that this forced transfer 
  might counter efforts toward efficiency or stability.

  The following are the ideas predominant in our discussions.  For the 
sake of healthy development a market-driven environment is best, so the 
basis should be "transferable".  In addition we should investigate the 
use of "periodical forced transfer" as a method of reducing abuse
through monopoly building.  In this case, registry should be determined
through auction among organizations applying for the same gTLD upon 
accreditation, as financially strong organizations would probably buy 
them back in the end even if the method of lot drawing was used.

9. Definition of "gTLD"

  We have been using the term "gTLD" as an acronym for "generic Top 
Level Domain" to date, but the WG-C report refers "gTLD" as "global Top 
Level Domain" instead.  We are not opposing the change. However, we 
insist that the WG-C should give the reason why they'd like to change it 
to the NC as well as to the public so as to avoid unnecessary 
misunderstanding and confusion.

  In some position papers the other types of gTLD than "general-purpose
gTLD", which are also called "limited-purpose gTLD", "chartered gTLD",
"sponsored gTLD" and so on, are proposed. Those types of gTLD should 
have not been the target of the WG-C discussion at the beginning. We 
think whether it should be remained as "generic" or be changed to 
"global" is not a simple name change but touches the concept of "gTLD".

--
Toshi TSUBO, Domain Name Working Group, Planning/International Working 
  Group, JPNIC <tsubo@global-commons.co.jp>
Hiro HOTTA, NTT, also a member of Planning/International Working Group, 
  JPNIC <h.hotta@hco.ntt.co.jp>
Tsugizo KUBO, Domain Name Working Group, JPNIC <azkubo@interlink.or.jp>
Naomasa MARUYAMA, Planning/International Working Group, JPNIC
  <maruyama@nic.ad.jp>
Shuichi TASHIRO, Planning/International Working Group, JPNIC 
  <tashiro@etl.go.jp>



----------
Hirofumi Hotta
R&D Strategy, NTT