Christian Huitema | 1 Aug 2002 02:24
Picon
Favicon

RE: Trees have one root


> Mr. Hall's original and current thesis is that adding TLDs as 
> an act is harmless, and that the real point of concern is the number
of queries.
> This point has not yet been disproven. The counter-argument 
> that a busy ("popular") TLD will result in more queries only proves
the point,
> although the advocates of such an argument refuse to admit that the
> frequency of queries -- the number of queries over time, the 
> measure of "popularity" -- is the issue, not the number of TLDs.

Except that the cache misses hitting the root are by no means *linearly*
proportional to the number of queries. As the popularity of the servers
in a domain increase, the load on the root grows asymptotically towards
1 request per TTL per resolver, and in fact it approaches the limit
quite fast. Split .COM in two, both halves will be very the asymptote,
both would generate almost the same load on the root as the previous
combination. Adding more TLD and splitting the existing load across
these new TLD would definitely have an impact on the root.

-- Christian Huitema

Eric A. Hall | 1 Aug 2002 06:58

Re: Trees have one root


on 7/31/2002 7:24 PM Christian Huitema wrote:

> Split .COM in two, both halves will be very the asymptote, both would
> generate almost the same load on the root as the previous combination.

If a truly generic TLD such as .com were randomly split, sure. You're less
correct for categorical classifications (.com versus .edu again).

This is not a call for a categorization of TLDs, although self-forming
categorizations would likely emerge from a free market.

> Adding more TLD and splitting the existing load across these new TLD
> would definitely have an impact on the root.

The number of queries would definitely go up, but not to a harmful level.
If it were that bad, we would have consolidated the ccTLDs long ago.

--

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/

Einar Stefferud | 1 Aug 2002 08:39

Re: delegation mechanism, Re: Trees have one root

I have edited the to/cc to eliminate the duplicates.  Enough is enough;-)...
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

I just want to clarify one aspect of what Ed Gerck said, by replying 
to his message to the IETF as included below.

ORSC has never added any TLD that already existed in the IANA cum ICANN root.

Perhaps some predecessors committed such acts before ORSC was formed, 
but certainly never since ORSC was established.

One of our policies has always been to reject any attempt of anyone 
to introduce any collisions among all known TLDs.  We have also 
worked successfully to to resolve all conflicts that threatened to 
"cleanliness" of the ORSC root.  This was done mostly be Richard 
Sexton tracking down the historical facts of each case and cajoling 
conflict prone parties to find some way to resolve their conflicts. 
I musty say I am very impressed with his results, and of course with 
his efforts.

But, we also understand that completeness of inclusion is a very 
important aspect, thus, the ORSC root servers also carry all the TLDs 
that exist in the ICANN root.
Root service users should be required to decide between two 
non-intersecting global root subsets!  At least one alternative 
should be a full meld of the whole DNS namespace.

So, the ORSC objective is (and has been since the ORSC confederation 
was activated) active in being both inclusive, and opposed to 
allowing conflicts.  Thus users of the ORSC roots see "the Whole 
(Continue reading)

Dave Crocker | 1 Aug 2002 09:34

Re: delegation mechanism, Re: Trees have one root

At 11:39 PM 7/31/2002 -0700, Einar Stefferud wrote:
>All of this went very well until ICANN chose to add a new TLD that 
>conflicted with our ORSC .BIZ TLD which was a long time member in good 
>standing in the ORSC "family.

Stef,

This federation that you are describing was formed without any involvement 
of IANA/ICANN.  It was done entirely independently.

So the idea that IANA/ICANN was somehow obligated to coordinate its own 
activities with an independent effort is a very basic non sequitur.

d/

----------
Dave Crocker <mailto:dave <at> tribalwise.com>
TribalWise, Inc. <http://www.tribalwise.com>
tel +1.408.246.8253; fax +1.408.850.1850

Valdis.Kletnieks | 1 Aug 2002 10:38
X-Face
Picon
Favicon

Re: delegation mechanism, Re: Trees have one root

On Thu, 01 Aug 2002 00:34:55 PDT, Dave Crocker said:
> This federation that you are describing was formed without any involvement 
> of IANA/ICANN.  It was done entirely independently.

Very true.  It's also true that if I refuse to answer the phone, I will
not have any phone conversations. 

> So the idea that IANA/ICANN was somehow obligated to coordinate its own 
> activities with an independent effort is a very basic non sequitur.

A case could be made that if ICANN was even *pretending* to serve the public
interest, they could at least enter into a *discussion* with the ORSC people.
Of course, we all know they won't even do that with their OWN board of
directors without a court order.

JFC (Jefsey) Morfin | 1 Aug 2002 16:19

way out of the DNS problems? (former Re: delegation mechanism, Re: Trees have one root)

On 10:38 01/08/02, Valdis.Kletnieks <at> vt.edu said:
>On Thu, 01 Aug 2002 00:34:55 PDT, Dave Crocker said:
> > This federation that you are describing was formed without any involvement
> > of IANA/ICANN.  It was done entirely independently.
>
>Very true.  It's also true that if I refuse to answer the phone, I will
>not have any phone conversations.
>
> > So the idea that IANA/ICANN was somehow obligated to coordinate its own
> > activities with an independent effort is a very basic non sequitur.
>
>A case could be made that if ICANN was even *pretending* to serve the public
>interest, they could at least enter into a *discussion* with the ORSC people.
>Of course, we all know they won't even do that with their OWN board of
>directors without a court order.

I am sorry to repeat it (and I will try not to say it too many times 
again): the terms of the ".arpa" sub global namespace delegation are 
described in RFC 920 by Jon Postel himself. The creation of ".biz" (as 
.info and others) are not permitted under that delegation. However ICANN 
obviously acted in this along with the DoC recommendation. The question is: 
should ICANN obey the USG or comply with the initial ITU community 
agreement - FCC included. ".biz" could have been a way to make 
Justice/Congress commit on this.

We all are confronted to the age of a stantard agreement which works well 
but of which the involved parties forgot a long ago the standard, the 
rationales, their rights and their duties.

Solution 1. we forget about it and build it anew. Then ICANN, ccTLDs, gTLDs 
(Continue reading)

Sean Jones | 1 Aug 2002 18:22
Picon

RE: way out of the DNS problems? (former Re: delegation mechanism, Re: Trees have one root)

Hi

This is my first posting to the IETF mailing lists, so please forgive me if I do something wrong.

<snipped to get to the crux>

> Solution 3. we keep going but we rebuild anew. This seems to be the 
> impossible ICANN policy in building their Intercontract system. It 
> obviously does not work and leads to strong oppositions.

> The reason why I join the IETF debate is that I think 
> technology may be the 
> response. The same as the namespace agreement has aged, the same the 
> technology has aged. We are endlessly arguing on old stuff. Aftyer 20 
> years, we need a clean sheet review of the DNS, based on 
> today and future 
> users' needs. IMHO this goes into two directions: a new DNS 
> core system 
> analysis (DNS.2) and an extended DNS services (DNS+) logic. 
> IMHO in going 
> ahead in that two directions (as the iDNs show the path) we will soon 
> discover that the ICANN preoccupations and solutions are 
> totally outdated 
> by what we will uncover and specify.

> Obviously we will meet the same kind of oppositions to 
> DNS.2/DNS+ than was 
> met for IPv6. The question is then to know if this effort 
> will be carried 
> within the IETF or not. I think DNS.2 cannot be specified 
(Continue reading)

Dave Crocker | 1 Aug 2002 19:16

Re: delegation mechanism, Re: Trees have one root

At 04:38 AM 8/1/2002 -0400, Valdis.Kletnieks <at> vt.edu wrote:
>A case could be made that if ICANN was even *pretending* to serve the public
>interest, they could at least enter into a *discussion* with the ORSC people.

One of the tests of an idea that is required in the IETF pertains to 
scaling.  So, please apply your thought to the general case.  It means that 
ICANN would be legitimizing anyone and everyone who chooses to start an 
independent activity.  This is a very bad idea for the stability of the DNS.

It makes far more sense for ICANN to literally ignore independent 
activities.  And that is exactly what ICANN has done.  To ignore them means 
that ICANN does not pay any attention at all to those other 
activities.  This includes ICANN treating the IANA/ICANN name space as 
self-contained, as in fact it is.  Hence, the only "collisions" that ICANN 
should pay attention to are collisions within the IANA/ICANN DNS.

Independent activities are, after all, independent.  They choose to work 
without coordinating with ICANN and they are free to do that.  (Contrary to 
Stef's interpretation, my comments are about registries and root 
administrators, not about users.)

What makes no sense at all is for these folks to act independently and then 
make a post hoc claim that ICANN is somehow obligated to coordinate with 
them.  Apply hour logic equally.  Why were these folk not, themselves, 
obligated to first do pre hoc coordination with ICANN?

>Of course, we all know they won't even do that with their OWN board of
>directors without a court order.

1.  On the contrary, the Board seems to work quite well amongst itself.  If 
(Continue reading)

Picon
Favicon

Re: delegation mechanism, Re: Trees have one root

This statement quoted below, in which it is alleged that Karl Auerbach got
only what he'd previously been offered, is flatly contradicted by the text
of the judicial decision. It is, quite simply, utterly false.  In fact,
almost the reverse is true: what ICANN got is almost exactly what Karl
offered them months ago (the only differences being he has to give 10 days
notice to disclose instead of 7, and some documents must be inspected on
the premises). The judge stated that in several material respects, what
ICANN sought violated California law. 

See the text of the decision for yourself at
http://cryptome.org/auerbach-icann.htm

See also http://www.icannwatch.org/article.php?sid=883 for a subsequent
development.

One does not have to be a lawyer to understand what the following means:
the court said ICANN's position "violates both section 6334 and Bylaws
Article V, Section 21 because it deprives Auerbach of the inspection
rights he has under law and imposes such unreasonable requirements as
having to sign a confidentiality agreement and having to pursue burdensome
review in any effort to enforce his inspection rights".

A few choice quotes from the court decision (note, Respondent == ICANN and
the "Inspection Procedures" were ICANN's attempt to restrict a director's
access to info):

[A]
"Respondent contends that inspection rights of directors may be restricted
by corporate inspection procedures and cites Chantiles v. Lake Forest II
Master Homeowners Ass'n (1995) 37 Cal.Ap.4th and Havlicek v.
(Continue reading)

Karl Auerbach | 1 Aug 2002 21:08
Favicon

Correcting an incorrect assertion. Was: Re: delegation mechanism...


The following concerns my lawsuit against ICANN and highly incorrect
misstatements made on this list about the outcome of that action.

I have no intention of turning this into a thread, but I do feel that I'm
within the bounds of etiquitte by sending a corrective e-mail.

On Thu, 1 Aug 2002, Dave Crocker wrote:

> 2.  Note that Karl won the suit, but lost the war.  He is now subject to 
> the conditions that ICANN had originally wanted to apply for his 
> access.  The only difference is that now Karl is under a court order to 
> conform to those rules.

The writer of the above paragraph has his facts completely wrong.  He is
doing nothing more than parrotting back ICANN's quite misleading press
release.

ICANN lost utterly and totally - not a single issue advocated by ICANN was
adopted by the court.  The court granted my motion for summary judgement.  

ICANN's so-called "procedures" and the "restrictions" issued under those
procedures were rejected by the court with a kick that would be the envy
of a world cup champion.

ICANN is trying to spin its complete loss into some sort of Pollyanna
story about how the result is what ICANN wanted all long.  If this is
really what ICANN wanted, then ICANN ought to have acquiesed to my offer
to them back in September of last year.

(Continue reading)


Gmane