Re: Technical (was: security, authorization, definition, etc.)

[Olivier Guillard / AFNIC <Olivier.Guillard@nic.fr>]

le 11 Jan, budi@alliance.globalnetlink.com a écrit :


> This one entity could be the registry, could be a
> registrar, could be the registrant, what have you
> depending on the (busines) model.
> From the technical point of view, we don't really
> care. As long as we can make sure that only *one*
> entity can make changes, right.

I really do agree with you, that why "rrp" is *not*
the proper name for a protocole aimed to update the
DNS database.

DNS is a db, we know what information is required
to delegate a new domains (basically 2 ns) on the
net.

On the other hand, under some TLD, there might not
have have any registrars but a single entity (.mu),
under other their might not have any delegation
without a validation done by a third party that is
not a registrar nor the registry and so on.
In all case their should have an interface plugged
to the db to update it be it should not be formally
designated as a "registry-registrar protocol".

Regards,

-- 
Olivier Guillard
---
||    A.F.N.I.C.
|| http://www.nic.fr
|| mailto:guillard@nic.fr