Re: [ga] New top-level internet addresses come with $100,000-plus price tag

[Karl Auerbach <karl@xxxxxxxxxxxx>]

Roberto Gaetano wrote:
> I am missing something here.
> We have been flooded over the years with emails that have been stating how
> successfule alternate roots are, on how ISPs all over the world are
> increasingly pointing to a system that is able to override what is in the
> "ICANN root" with a different choices, and so on.
(I tend to use the phrase "competing roots".)

The idea of having multiple root systems is to provide a way for users 
and providers to interact, in a way similar to the complicated ways that 
buyers and sellers interact in a multi-vendor marketplace, to arrive at 
a relatively stable and relatively consistent set of TLDs without the 
need for a central decision making authority (such as ICANN.)
The notion is one of "consistency" among the various root systems.  By 
this I mean that the desired goal is a world in which most competing 
root systems offer a core set of TLDs, and in which some root systems 
perhaps offer additional TLDs that are aspiring for market share.  (I 
call these "boutique" TLDs.)  I'll come back to this in a moment.
But for any given TLD name the idea is that ideally that TLD should 
represent the same DNS subtree, i.e. .example in the legacy root should 
be the same as .example in the other competing roots.
How is this "same name, same meaning" obtained?  Through the positive 
application of an idea that has previously been the source of so much 
trouble: the laws of trade and service marks.  If two or more people try 
to offer different versions of .example then they have recourse to the 
legal system to resolve that dispute.
And, of course, while there is a viable dispute, few rational root 
server operators would incorporate what are, in effect, damaged and 
dangerous goods in its root's inventory of TLDs: root server operators, 
like the rest of us, do not want their customers to be surprised and to 
take their custom elsewhere as a consequence.
And, in addition, informed rational domain name buyers will tend to shun 
TLDs that are in dispute, thus giving the disputants an incentive to 
resolve their differences.
(It is only through luck that ICANN was not caught in a legal thicket 
when it chose to create a .biz in conflict with a prior operator of a 
TLD with that name.  Had the prior operator held her ground it is 
possible, even probable, that ICANN may have found itself denied because 
of the prior use.  ICANN may not be so lucky in the future despite warm 
and fuzzy noises from ICANN's has-not-always-been-right law firm that 
would make money from ICANN should such disputes arise.)
It is highly likely that most sane root system operators will pick up 
the current ICANN/NTIA root zone and add to that zone those additional 
boutique TLDs that each wants to offer.
Thus we would probably end up with:

  - a legacy root with the NTIA/ICANN root zone and no boutique names.

  - Other roots with the NTIA/ICANN TLDs plus such uncontested others 
as each root choses to add.
  - A set of TLDs that are in dispute.  We can anticipate that at least 
some roots, either from lack of knowledge of the dispute or from simple 
stupidity or bullheadedness, will pick up each of the disputed versions.
  - There will be some roots that simply chose to be completely 
inconsistent.  These are like some of the alternate roots today - and 
they will garner the same, nearly nil, presence.
  - Root operators will strive to obtain market share.  They will do 
this by offering enhanced services - such as placing a root server onto 
a customer's net (and perhaps that server will do recursion and caching 
in order to reduce name lookup latency for that customer).
    And perhaps that root operator will even purchase market share in 
the same what that Google does - by buying it.  Consider a root operator 
that will pay an ISP $$ if that ISP switches to that root operator. 
That root operator can data mine (just as ICANN permits Verisign, PIR, 
and Afilias to do) the query stream and carve out a share to pay ISPs.
Domain name buyers, if we assume that they are informed and rational, 
will chose whether to buy names in the core set of TLDs that are present 
in all TLDs.  Most will buy names in stable core places - the .com/.org 
kind of places.  Some will chose to buy elsewhere, perhaps due to price 
perhaps for other reasons (for example a tight community of people, such 
as a church, may chose a boutique name for their shared purposes) - it 
will be a marketing effort for those boutique TLDs to obtain customers. 
 It will be an uphill effort but one that is very common in the fight 
for shelf space that occurs with products in the non-internet world.
Some boutique TLDs will grow, most will wither and die or simply remain 
tiny and of no real consequence except to its users.
If a boutique TLD grows more and more roots will decide that that TLD 
should be included into that root's inventory of TLDs.  Thus a boutique 
TLD can naturally grow into a core TLD.  (And, of course, core TLDs that 
fade or become tainted may fade out of the core and become boutiques.)
I know that this is an inadequate description.  However, it is merely a 
copy of the kind of brand-building efforts that go on every day in 
real-life non-internet market places as new products strive for market 
share and shelf space.  It is a system that works.  It is a system that 
does have conflicts when products stumble upon the same name (or when 
someone tries to steal market share by forging and usurping an 
established product name.)
This kind of system eliminates the need for a top-down imposition of 
names, i.e. ICANN becomes just one of several root zone definitions.  It 
is a truly bottom-up system because it works through the aggregation of 
independent choices by those who wish to create TLDs and those who chose 
to use them.
This kind of system allows the natural growth and death of TLDs.  And 
that life cycle depends on the efforts of the TLD providers and the 
choices of users rather than life supporting decisions from a 
centralized body of internet governance.
It is good that ICANN is finally opening up to new TLDs.  It is bad that 
ICANN is creating a belief that new TLDs must jump through ICANN defined 
hoops in which business plans and financial depth are evaluated by ICANN 
rather than by the marketplace.  (Last year I took a stab at putting 
together what I think should be ICANN's TLD application form.  Take a 
look at the latter part of 
http://www.cavebear.com/cbblog-archives/000324.html )
By-the-way, one would hope that ICANN will finally grant the 
applications of the 40 remaining applicants from year 2000.  They have 
already paid their fees and have been patiently waiting for 8 years.
                --karl--