Re: [wg-c] reposted for Harald Tveit Alvestrand

[Paul Garrin <pg@name-space.com>]

David,

> 
> At 05:13 PM 12/19/1999 , Paul Garrin wrote:
> >There is no need to fragment the internet.  Multiple roots
> >can and should exist, but must be coordinated.  This is not
> 
> If they are coordinated, they are not "roots" in the real sense.
>

I would say in the "legacy" sense, not the "real" sense.

> Or, rather, the coordination authority is really the root.
>

A coordination authority would be applicable in the administrative
sense, but not necessarily in the technical sense.  Administrative
coordination is important for quality control and accreditation of
operators and in the setting of and enforcement of standards.  This
could be done by ICANN, or it could be the IETF or the IAB.
The most important criteria rather than who does it is that is is
done with competence and understanding of the underlying technical
issues, and with fairness.

The technical coordination is the duty of the peering networks
that operate the decentralized and coordinated root systems in
accordance with the standards set through engineering and development
of the shared root system.  Decentralizing the root (or better,
the decentralized control of the root.zone) is easier than the
decentralized sharing of TLD operations, assuming that updates
to the root.zone (adding TLDs) happen at a low frequency and
updates to the TLD zones (adding SLDs) happens at a much higher
frequency.  Decentralized sharing of TLD zones is technically
more difficult and demanding to reduce the number of collisons,
or conflicting applications for the same sld strings, but it is
not an impossible task to achieve.

 
> 
> >only technically possible, but necessary to provide for the
> 
> Please explain the details of the insight you have in these matters, which 
> permits you such a basic difference of technical analysis from that of the 
> IAB.
> 

I don't make such grand claims as to have more experience than the IAB.
I did however discuss the issues with Paul Mockapetris at length some
time ago and his response was that competent computer scientists could
build it.  I have such talent on my team and we are doing it.

> It would be helpful to provide some history of your design and 
> implementation with large-scale distributed systems, as well as operations 
> experience with them, to contrast with that of the IAB's somewhat awesome 
> aggregate experience.
> 

Again, I don't make such grand assertions.  However, my engineers,
one of whom is a PhD candidate at Columbia U. and has worked for
NASA and DARPA and who holds several patents, and another who has worked for 
Lucent, are competent in these areas and are working on just such a system
which we will open up to a testbed in the coming months.  If you are
interested in participating please keep in touch with us and we will
keep you up to date on our progress.

regards,

Paul Garrin