Address Ownership Draft comments

["Albert E. Manfredi" <[e-mail]>]

I read with great interest draft-ietf-cidrd-addr-ownership-07.txt. A few 
comments, no disagreement from me, came to mind.

1. It is not by accident that the telcos around the world have been 
practicing what this document calls "address lending" for such a long 
time. I would agree that constraining the topology is unrealistic and
that the Internet isn't going to get any smaller, so there seems to be no
more credible solution right now.

2. The functionality that some sort of "address portability" scheme would 
provide is really no different from what the existing Domain Name System 
(DNS) provides. As one who has "been renumbered" twice already, I can 
vouch that the hassle is purely internal and short-term. Inventing an 
address portability scheme, in my opinion, would just duplicate the DNS 
with a numbers-only database.

3. Given the above, I would be more insistent on an "address lending" 
policy.

4. The Internet has many service providers already, which is why a 
topology tied to geography alone is not useful. At first, one might find 
this reason enough to discard any links with the telco schemes (E.164). 
But in fact, the telcos will also have to accommodate multiple providers 
in the local loop, so as things stand now, i.e. with CIDR in the Internet 
and with multiple providers in the telco local loop, I see no distinction 
in the goals and in the design of these addressing schemes. The
differences in the near future, as far as I can tell, are only in the
nitty gritty of formats.

5. The implications to ip-atm's ROLC, ND, NHRP, ... , of this evolving
state of affairs are interesting indeed. In a positive sense.

Bert
[e-mail]