[ga-roots] A Unique, Authoritative Root for the DNS

["Patrick Corliss" <patrick@quad.net.au>]

On Tue, 29 May 2001 23:54:23 +0700,
Stefan Probst <stefan.probst@opticom.v-nam.net> wrote:
To: ICANN@icann.org
Subject: Comments to ICANN Discussion Draft: A Unique, Authoritative Root
for the DNS
Cc: ISTF Discussion List <istf-participants@lyris.isoc.org>,
        domain-policy@open-rsc.org


Dear unknown Author, All,

hesitantly I accept the terminology of "authoritative root" for ICANN's
root, because this is what it is called in the DoC contracts, although I
think this term is neither significant nor useful.

In order to make the scenario furthermore a bit more clear for outsiders, I
would like to explain, that at present, there are in fact only very few
"alternative" roots, which - with a bit of simplification - can be
described as follows:

- one big group of numerable TLD operators, which resolved most of their
   previously conflicting TLDs, and are effectively operating now one joint
root,
   the "SuperRoot". The degree of community governance vs. individual rule
over
   the TLDs varies. This SuperRoot claims the historical legitimacy of being
   the originally planned "Test Root" for new TLDs.
   Organisations in this group seem to be: ORSC, TLDA, PacificRoot, OpenNIC,
   AlterNIC, IRSC, NARSC, AURSC, EURSC, CINCIS (?)
- New.net, a completely commercial operation, which "sells" their root very
   aggressively during the last months.
- Namespace.com, with its over 500 (!) mostly conflicting TLDs.

(I have to admit, that I didn't check iDNS yet).

Your discussion paper IMHO mixes two issues, which raise - although somehow
connected - different questions:

1. Competing Roots, i.e. roots which host domains, which are also in other
roots.
2. Inclusive Roots, i.e. roots which host additional domains.

It seems to me, that nearly nobody doubts the final infeasibility of
competing roots. That was obviously the reason, why all the a.m.
organizations (except New.net and Namespace.com) finally agreed to
co-operate and to avoid competing TLDs. In a scenario of competing roots,
there is no winner, only loosers.


All of the a.m. "alternative roots" are however "inclusive roots" towards
the ICANN root.

In a scenario of a SINGLE "alternative root" being assigned authority by a
growing number of users, your concerns regarding wrong delivery of eMail,
cache poisoning, not uniqueness of an URL, etc. are not applicable.
The only technical issue, which stays in this scenario is "service or not
service" / "resource not found": Users who delegated authority to this
SINGLE other root would reach everything, users who stay with the ICANN
root would reach only other resources under the ICANN root, not the other
ones. Hosts "en route" might make it more difficult, but it is said, that
the Net routes around point of failures....

In this scenario of a SINGLE "alternative root", users or ISPs by switching
roots would - at least in the short run - loose nothing, but potentially
gain. Once a critical mass switched roots, it will proof difficult to hold
back the rest.

ICANN's concern will therefore have to be to prevent a single other
"alternative root" being chosen as the authoritative root by a critical
mass of Internet users, without ICANN's agreement.
If New.net's numbers are to be believed, then IMHO there is a considerable
threat in this direction. Your paper doesn't mention how you intend to
counter that.

You have valid concerns in principle regarding the "public trust",
especially against only commercially minded companies like New.net or
Namespace.com, which don't bother at all about anything other than their
own business interests.

However, your argument about representative legitimacy looses weight when
dealing with TLD operators who are willing to cooperate to ensure the
stability of the Net, i.e. obviously most of the organizations under the
SuperRoot. The question which remains then is "hierarchical structure" vs.
"flat structure" of the Name Space.

As long as the system in itself, and the operators of the other TLDs are
able to keep the system stable, I see not much reason in keeping the space
strictly hierarchical as it is now. Why does the name space need only to be
frequented by a few buses (a few models fit all), and not by many
individual cars? Your paper doesn't provide answers to this.

With regards to legitimacy, public trust etc. you win in theory, at least
against companies which behave like New.net. However, technical history has
shown us, that VHS won over the better BetaMax, Windows over the better
MacOS. Once there is a critical mass of users, then virtually nothing will
stop them to go for the principally inferior solution.

All in all, your paper sounds to me very much like Sam Lanfranco's "early
leaders" vs. "second wave followers". You fail IMHO to produce a viable
plan on how to counter a threat like New.net.

And by not showing the disadvantages of a flat name space structure (if
there is), you fail IMHO to produce convincing arguments about why not
looking for possible cooperation with (at least some of) the organizations
under the SuperRoot.

Yours sincerely,

Stefan Probst




--
This message was passed to you via the ga-roots@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga-roots" in the body of the message).
Archives at http://www.dnso.org/archives.html