we already have that... the current tld/root server system.

once upon a time, all we had was a file called hosts.  it told us where we 
would find other computers on the (arpa)net, and it scaled to a few 
thousand hosts, as updated copies of the file were passed around manually. 
then someone said "this would be much better with a central database of 
these name to number mappings," some code got written, and the dns was 
born.  .gov, .mil, .edu, .com, .org, .net, and most of the cctlds were 
registered.  then someone wrote some more code, and viola:  domains cost 
money, but you could only buy a 2nd level domain, and only certain ones. 
great amounts of money were transferred because more and more people 
wanted to find their computers on the (inter)net, but the money was 
transferred only to those who had the key to the lock placed on the dns... 
the authority to add a entry to what grew from good old /etc/hosts.
and now keys are for sale.

let us make no mistake here:  the dns is now more than a way to find other 
computers easily on the internet.  the dns is a mechanism of governance. 
it is also a perpetual money machine, applied in a capitalistic fashion. 
will anyone actually make any money off of new tld's?  those providing the 
tld's, of course.  a few lucky and talented individuals who manage to get 
a tld with reasonable commercial viability AND develop it over a period of 
a decade or more will turn a profit.  after all, this is the standard 
telecom model:  sell once, bill in perpetuity.  that is what is really 
going on here... a successful model is being franchised.  will more than 
5% of the franchisees survive?  doubtful.  there is no shortage of 
names... no finite limitation on the size of a zone file.  supply 
and demand does not apply... there is infinite supply to fulfill a finite 
demand.

Enjoy,
Scott





On Thu, 18 Mar 2010, Marc Perkel wrote:


> However - another model is where you are just using the TLD for private name 
> resolution and you and not offering public registration but rather allowing 
> an organization to use it like a second level domain name. For example, .ibm, 
> .microsoft, .google, .catholic, etc. That way - for example, instead of going 
> to google.com you can just go to google. In this situation the .google TLD 
> would be owned by google and they would control it just like they control 
> google.com.
>
> In the second model the TLD is no more complex than running ant other 
> standard name server. It would be equivelene to running a "." (nothing) TLD 
> where instead of registering .com, .net .org or .info you would just register 
> with "."
>
> You would need a registrar to handle the "." TLD.
>