[wg-c-3] Notes on new gTLD registries

[Kent Crispin <kent@songbird.com>]

                  Notes on New gTLD Registries
                         July 9, 1999
                         Kent Crispin


Terminology
-----------

It's probably impossible to get everybody to agree on terminology, 
but at least I want to be sure that people understand what I am using:

database:  (abstract) a formally structured collection of data; 
(concrete) a system of computer software/hardware that implements a 
database.

TLD: One of the entries in the IANA-approved root zone.

gTLD: a TLD that has no enforced criteria for the entities that may 
register in it.  This departs from the rfc1591 definition.

Registry: a database associating DNS information with some person,
legal entity, operational entity, or other referrent.  Note that we
can speak of a registry in the abstract or in the concrete, as per
the definition of "database" above.  To emphasize the abstract
meaning we may use the terms "registry database", or possibly
"registry data". 

gTLD registry: a registry for a particular gTLD ("the .com registry").

Registry operator: the organization or business that operates a 
registry.  This distinction is very important:  NSI is the operator 
of the .com registry;  Emergent was the operator of the prototype 
CORE registry.

Registry administrator:  registry operator.

Registrar: an entity with a direct contractual relationship with, and
special access to, a registry, that inserts, modifies, deletes, or
otherwise manages records in the registry on behalf of others. 

Registration agent: Registrar (the Nominet registry uses the term
registration agent to mean something slightly (but importantly)
different than a "registrar", but I am operating at a fairly high 
level of generality here.)

Shared Registry: a registry that allows access from multiple
distinct registrars.


Premises
--------

While it is possible to argue these in other contexts, I consider 
them to be part of the ground rules of this discussion:

1)  New gTLD registries will be shared registries (mandate from 
white paper)

2) ICANN will accredit all gTLD registrars (white paper; ICANN/USG 
MoU)

3) We are only interested in the IANA root zone.

4) The dns system is part of the public service infrastructure -- it 
  includes governments, schools, museums, and long-term data 
  archives as its users.  With the deployment of the vastly  large IPV6 
  address space, individual traffic lights could be have domain 
  names, and be synchronized over the internet.  Consequently, 
  stability of operation of the dns is the *primary* requirement.  



Profit, non-profit, cost-recovery, public trust/resource
--------------------------------------------------------------

A substantial body of opinion exists that *all* gTLD registries
should be public resources; there is another body of opinion that
states that some gTLD registries could be privately controlled; but
there is no significant body of opinion that states that *all*
registries must be privately controlled.  That is, no one has a
significant problem with there being some "public resource" oriented
gTLDs.

The controversy, therefore, has been over whether there should be
privately controlled gTLDs.  Feelings run very deep on both sides of
this issue, and it seems clear that the controversy will not reach 
a consensus.  

Therefore, the only prospect for getting new gTLDs in the root in any
reasonable time frame is for them to be admitted under the "public
resource" model. 

The characteristics of the "public resource" model are as follows:

  The registry data is considered a public resource, subject to
  privacy limitations, held in trust for the public by ICANN.

  The registry is operated as a shared registry on a not-for-profit
  cost-recovery basis.  The registry operator, however, may be a
  for-profit company, operating the registry under contract to 
  ICANN.  The registry operator may be removed for cause, and the 
  contract would be rebid on a periodic basis.  

  Since the data in the registry is considered a public resource, it
  should be escrowed under different control from the registry
  operator, and in widely dispersed jurisdictions and locations.

  Ideally, there would be several registry operators, any one of
  which could, within a few days, assume operation of a gTLD registry
  from escrowed data.  These registry operators should be distributed
  worldwide.  Presumably each registry operator would operate several
  gTLD registries at the same facility.  

  Even more ideal, the transfer of registries from one registry
  operator to another would be completely routine -- for example, a
  small company in a location with lesser internet access could run
  several very small registries, but transfer a gTLD to another,
  better connected registry operator if the load got to high.  This
  would enable developing countries to develop registries.

  Registry operators can fail; physical disasters can strike a 
  particular installation.  Having multiple dispersed registry 
  sites with multiple operators gives a great deal of robustness to 
  the whole DNS.  A single monolithic site, no matter how secure, 
  can fail, but distributing registries like this, with escrowed 
  copies of the registry data available for quick switchovers would 
  be a far more bombproof and resilient system.

  A requirement of easy transferability of registry data is that the 
  underlying software and protocols be standardized.


Concrete Proposal
-----------------

With the above model in mind, I propose the following:

  1) Six new gTLDs be approved immediately.  I would propose that
  they be chosen from the IAHC gTLD set; and that CORE relinquish any
  intellectual property rights they may have acquired in these names
  to ICANN. 

  2) That a request for proposal for registry operators be tendered
  quickly.  The goal of this rfp would be for three independent
  registry operators from three different regions of the world to
  operate six gTLD registries, two per operator. 

  3) That ICANN support the standardization effort in the IETF for a
  shared registry protocol, and that the six new registries all use
  this protocol.

  4) That the new registries operate according to the public 
  resource model described above.



-- 
Kent Crispin                               "Do good, and you'll be
kent@songbird.com                           lonesome." -- Mark Twain