[Politech] Karl Auerbach replies on U.N. Net control and many, many top-level domains

From: Declan McCullagh (declan@private)
Date: Wed Oct 05 2005 - 23:21:14 PDT

Previous Politech message:

-------- Original Message --------
Subject: Re: [Politech] Weekly column: Will the U.N./Bush administration 
split the Net?
Date: Wed, 5 Oct 2005 00:12:40 -0700 (PDT)
From: Karl Auerbach <karl@private>
To: Declan McCullagh <declan@private>
References: <4343237E.60503@private>

> Power grab could split the Net
> By Declan McCullagh

I think that you might have misstated a few subtle points.

It is possible that even Paul Vixie and I would agree that one solution to
the question of DNS governance would be for all the root server groups to
use the identical suite of top level domains with exactly the same
delegation information.  In other words there would simply be replica root
server systems that all serve up the same menu of identical TLDs.

Such a world could be commercially interesting if some root server
providers were to specialize on the basis of accuracy, responsivity, or
availability.  But the standard of service provided by today's
*.root-servers.net is nothing less than superlative.  A new offering would
have to be extraordinary to induce anybody to switch.

However, many countries (or regions) will view it as a matter of national
pride, if not of national security, to offer root servers that are not
perceived as being in a position to do data mining or poison responses at
the behest of the US (or some other country that is perceived to have
divergent interests.)

Clone roots may also have reason to come into existance during the
recovery from natural or human disasters as the victims try to build their
network infrastructure from the inside-out as the rest of the world tries
to rebuild from the outside-in.  But those would usually be short-lived.

So there are reasons why clone roots could come into existance.  But as
long as they are exact data clones then my guess is that there won't be
much of a fight.

(I am playing a bit lose with the fact that DNSSEC might introduce some
issues even with cloned roots, but that question has not been fully aired,
much less answered.)

Where the troublesome differences and concerns start to grow are in the
following two issues:

Issue #1. Whether different root server groups could offer different menus
of TLDs.

Issue #2. Whether TLDs with the same name must contain the same contents.
(In other words, might there be a .com with contents different than that
maintained by Verisign, the ICANN designated registry for the its instance
of .com.  This isn't hypothetical - this kind of situation was created
when ICANN selected .biz even though there was a prior existing TLD with
the same name.)

I think that issue #2 contains the outcome we all most fear.  And it is a
situation that I think we can all agree deserves to be prevented.

Yet do we need technical limitations or institutions of internet
governance to ensure consistency of same-name TLDs?  I think the answer is

Inconsistency between same-named TLDs can be prevented simply by
considering each TLD to be a brand and allowing the registry for that TLD
to "police its brand" through well established mechanisms for trade and
service mark enforcement.  I don't think we need to invent any new
mechanisms of worldwide internet governance for that.  (Nor do I think we
need put any new kinds of legal weapons into the hands of intellectual
propery owners - they have demonstrated a predisposition to swing them
without looking.)

With issue #1 the question gets muddy and I believe that issue #1 is the
proper frame for where we ought to be debating.

As a practical matter with multiple roots there probably would be
considerable overlap from root to root - pretty much every root group
would have the same offerings of core TLDs, such as .com, .net, .org ...
and probably the entire suite of NTIA/Verisign/ICANN TLDs [and ccTLDs]
known today.)

The crucial point is that there would be some areas of non-overlap.  For
example, some root server groups might have my .ewe on the menu, some
might not.

Whether that is acceptable or not is the point of contention.

One might say that the question is moot: There's really no technical way
to prevent this from happening short of some worldwide ban - but where
would the authority from that come from and how could it be enforced?
And besides, doesn't such rigidity fly in the face of the end-to-end
principle, not to mention the principle that everyone should be able to
use the internet in the way he/she choses as long it doesn't harm someone

And there are many people out there who want to have new TLDs so they can
make "the big bucks" that others have made.  That is a pressure that is
going to be hard to resist.  In fact ICANN's reluctance to grant new TLDs
has create a rising flood of pent-up demand.

On the other hand there would be situations in which you would receive a
domain name (or URL/URI) that you could not resolve through whatever
system of roots you subscribe to.  For instance I could ask you to call me
at sip:karl@private

Some people find that to be unacceptable because it means that someone,
somewhere might find it hard to communicate with me.

I, personally, have the opposite point of view.  I find it to be an aspect
of freedom of choice - If I make myself known via a domain name that is
one of the lesser, boutique TLDs that are not widely honored, then that is
one of the consequences of the choice that I have made.  I do not consider
it my duty and the duty of every user of the internet to make ourselves
accessible to everyone in the world.  Rather, I consider it a matter of
choice whether we want that or not.

What is of interest to me in this is that the question is not a technical
one but rather one of social and individual rights and values.

Yet much of the debate is cast in the frightful language of technology
with overtones of internet catastrophe.  That tends to scare away those
who believe they have nothing to say in these matters or who believe that
their opinion is irrelevant because the lack deep technical credentials.


Politech mailing list
Archived at http://www.politechbot.com/
Moderated by Declan McCullagh (http://www.mccullagh.org/)

This archive was generated by hypermail 2.1.3 : Wed Oct 05 2005 - 23:40:10 PDT