Re: [ga] Re: ICANN & Stability

[Alexander Svensson <alexander@svensson.de>]

At 17.09.2002 14:58, DannyYounger@cs.com wrote:
>Stuart,
>
>As a reasonable man, do you plan to protect the stability of the Internet by 
>acceding to the reasonable request for these nameserver updates while the 
>ICP-1 policy is re-examined?

This standoff is painted a bit too dramatic by both sides.

ICANN is insisting on a principle (first a zone transfer, /then/
a nameserver), although it probably does not expect that giant
ccTLDs like .de/DENIC have malformed root zones which the ccTLD
operator does not check for RFC compliance.

The ccTLDs are insisting on a principle (no zone transfer to 
IANA), although they probably do not expect ICANN to use the
data for illicit purposes and they can get written assurances
required by national privacy laws.

The main problem is the lack of detailed and accepted procedures.
RFC 1591 is not clear: "There must be a primary and a secondary 
nameserver that have IP connectivity to the Internet and can be 
easily checked for operational status and database accuracy by 
the IR and the IANA." Operational status and database accuracy
can be checked in various ways. The controversial ICP-1 is 
supposed to contain only existing policies; however, this 
specific policy has only been enforced from time to time during 
1999-2001. It is understandable that ccTLD operators perceive
this as a new requirement -- for them, it /is/.

The issue behind the problem is of course the ccTLD-ICANN 
relationship. Some ccTLDs continue to talk to ICANN like customers
of a provider of root services, because that's what they want
it to be. ICANN continues to talk to ccTLDs like the Guardian 
of the DNS, rigidly applying its requirements even -- or rather:
especially -- in cases where swift action is needed.
ICANN seems to think: If we allow one change of a nameserver
entry without exercising our rights, we might lose them.
Some ccTLD seem to think: If we allow ICANN to inspect our
zone files, we accept some form of ICANN "supremacy" forever.

In the end, we need to discuss the ICP-1 requirement again,
without the question of precedents obscuring the view.
CENTR states: "Whilst some tests could be conducted on the zone 
file - such as for syntactic correctness - we do not believe 
they are of sufficient importance to justify the need for zone 
transfers." (http://www.centr.org/news/ICANN-AXFR.html)
ICANN a.k.a. IANA states: "Proper DNS configuration of all TLD 
zones is a global, not just a local, issue." 
(http://www.iana.org/faqs/tld-zone-access-faq.htm)

Why not set up a small, technically oriented task force with
people from ICANN and ccTLDs to find out what checks ICANN 
would have performed, how these checks could be performed
without transferring zone files or at least in a much more
secure way. If the ccTLDs have to be content with a written
assurance by ICANN that their data will not be misused, 
ICANN should equally be content with a written assurance by
the ccTLD operator that a suite of consistency tests has 
been performed according to mutually agreed procedures.
Of course, some ccTLDs may actually prefer to have ICANN 
perform the tests independently.

Best regards,
/// Alexander

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