Re: Nameserver as object/entity or not ?

["James Seng/Personal" <James@Seng.cc>]

This is a very reasonable requirement. If the original requirement is to
overcome this problem, then I suggest rewording it to something like
what Bill said below.

-James Seng

----- Original Message -----
From: "Bill Manning" <bmanning@isi.edu>
To: "Brian W. Spolarich" <briansp@walid.com>
Cc: "Patrick" <patrick@gandi.net>; <ietf-provreg@cafax.se>
Sent: Tuesday, February 06, 2001 11:57 PM
Subject: Re: Nameserver as object/entity or not ?


> %
> % | >From experience, I can tell you that it is used. I was working
before
> % | in an ISP. One day we had to change our Internet connexion, which
> % | changed the IP of our nameservers, thus the need to change the
> % | properties of all domains using these nameservers.
> % | I'm working in a Registrar now, and it happens that customer ask
us
> % | to change 50 or 100 domains at the same time because they change
> % | nameservers or they change IP of one nameserver.
>
> A case in point.  A server was recently renumbered and the host record
> at the registrar was modified.  This server was/is hosting a number
> of distinct zone entries.  Apparently the internal protocol tracked
> the new number at a number of levels but did not work across the
board.
> This server also serves as a server for a number of ccTLDs and the
"glue"
> records for this server at those ccTLDs was -NOT- automatically
updated...
> leading to "lame" server indications.   So the protocol ought to
ensure
> that when a "host" record is modified, that it can be tracked
throughout
> the system.
>
> --
> --bill