Yes.
> If yes: How many?
Here's how I answer that question:
There should be no limit on the number of root systems that may be
established by those who which to establish such systems. When I say root
system, I don't mean an individual root server but rather an entire set of
servers that operate much like the current set, a-l.root-servers.net --
one points to 'em via the named.cache/db.cache/cache.db file.
Each such root system operator will try to attract customers based on its
service offering. This service offering will consist to two things: the
inventory of TLDs that it offers and value added services.
Starting with the second thing, value added services - Yes, one can
conceive of a DNS root system in which there are benefits to be obtained
by the choice of root. For example, one can conceive of a system in which
the subscribers to that root will obtain DNS query responses that are
filtered to exclude well know porn sites. (Yes there are other ways to do
this, but why should we prejudge the best way?)
The first thing - the inventory of TLDs: A root system operator will
select which TLD's he/she wants to include. That selection will be based
on what the root operator thinks would be appropriate.
(Overall, I expect all root operators to try to trump one another with
more and more TLDs, the net result being that everybody has every TLD and
the only difference being the value added services.)
This makes the answer to the question: let the root server operators
decide what TLDs they which to recognize and include in their "inventory".
As for what happens if there are two or more different groups that want to
offer a TLD of the same name? Well, I'd let 'em duke it out among
themselves using established legal and economic methods. I know that if I
were a root operator I'd be hesitant to include any disputed TLD in my own
inventory.
Yes, this is a somewhat "radical" proposal, but it is one that replaces a
regulated single DNS with one that is regulated by the competitive forces
and is in line with the actual technology of DNS. It also makes the net
more robust by removing a single point of failure.
And yes, it is not a panacea. It leaves open some of the issues most
important to individual domain name owners - like that of
registry/registrar lock in and domain name portability.
It also trusts to the rational judgement of root zone operators to avoid
doing economically stupid things that would also reduce the degree to
which net users can depend on DNS lookups to give them the answers they
expect.
--karl--
-
This message was sent via the IDNO-DISCUSS mailing list. To unsubscribe,
send a message containing the line "unsubscribe idno-discuss" to
majordomo@idno.org. For more information, see http://www.idno.org/