Re: [wg-c] Re: multiple_roots

[Dave Crocker <dcrocker@brandenburg.com>]

At 01:12 AM 8/31/99 , Karl Auerbach wrote:
>As for your concern about what holds the net together - I'm relying on
>enlightened self interest or what you call "anarchy.".  Which, by the way,
>is exactly what holds the net's routing system together.  If it's good
>enough for routing, its certainly good enough for DNS.

Karl,

You are confusing USE with ASSIGNMENT.

Autonomous System numbers and IP net numbers hold the net together and 
those are assigned uniquely, specifically to avoid the kind of referential 
ambiguity that you want to encourage for naming.

The enlightened self-interest you cite pertains to ISPs using those numbers 
through peering arrangements.  The equivalent for the DNS would be to have 
unique assignment, as we do today, but have root operators decide which 
numbers among the single, unique set, they want to carry.

By the way, the DNS magnifies such selectivity:  Given the hierarchical 
nature of the DNS, a selective filter at the top of the DNS has a 
disproportionate effect since it prunes an entire branch rather than a more 
limited portion.

As odd as all that would be, I'd probably favor having DNS administration 
replicate the AS and IP net number assignment system...  The top of the 
assignment hierarchy is 3 non-profits coordinating among themselves.

>I've shown this working group a method that can get it out of its bind
>(pun intended).  If the working group choses to explore it, great.  If the

Karl, you are making a technical and operations proposal.  That needs to be 
put forward to the IETF and/or the net operations groups like NANOG and RIPE.

>No matter what this working group or ICANN does, multiple roots can grow
>and flourish.

So can world peace.  All we need to do is want it.

>If that is the case then the net is in sad shape because today we have
>multiple roots, they are deployed, they are running, and they work.

We have multiple roots in the global Internet?  Other than some fringe 
efforts that have garnered a fraction of a percent, where are this 
invisible multiple root services?

>So?  We have paper telephone books, we have CD-rom phone directories, we

Telephone numbers are assigned through a hierarchically centralized 
administrative service.

Although citing White Pages is useful to explain the simple mapping nature 
of the DNS, the DNS is not really comparable to the White Pages, since 
people do not require use of the White Pages to make a phone call.

The DNS is part of the operational infrastructure, rather than being a 
useful support service.  Telephone services resolve ALL telephone references.

>So we have divergency, so what?  As long as it is divergence without a
>difference, who cares.

Then "as long as" show a willingness to pursue a pretty remarkable degree 
of risk to net stability.

d/

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Dave Crocker                                         Tel: +1 408 246 8253
Brandenburg Consulting                               Fax: +1 408 273 6464
675 Spruce Drive                             <http://www.brandenburg.com>
Sunnyvale, CA 94086 USA                 <mailto:dcrocker@brandenburg.com>