RE: Handling of External Host Objects

[Bruce Tonkin <Bruce.Tonkin@melbourneit.com.au>]

This makes sense to me.

An external host could be referred to in the same way that an email address
field is used.  
The registry is not authoritative for the email address, nor is it
authoritative for an external host where no glue records are created.

Regards,
Bruce


> -----Original Message-----
> From: Hollenbeck, Scott [mailto:shollenbeck@verisign.com]
> Sent: Sunday, November 03, 2002 12:08 PM
> To: 'ietf-provreg@cafax.se'
> Subject: RE: Handling of External Host Objects
> 
> 
> One of the reasons I think that this external host thing 
> continues to be an
> issue is because it's a kludge.  I firmly believe that the 
> only objects that
> should exist in a repository are objects for which the repository is
> authoritative.  Other needed data should exist as attributes 
> of existing
> objects.
> 
> I suggested this concept related to external hosts back when 
> the topic first
> came up.  Some folks had issues with it, but I'll suggest it 
> again as an
> alternative.  In a nutshell:
> 
> - The only objects that should exist in a repository are 
> objects for which
> the repository is authoritative.
> 
> - Host objects should only be created in a repository that is 
> authoritative
> for the host.  In the case of hosts as name servers, 
> "authoritative" means
> that data in the repository (host name and address(es)) is 
> used to publish
> DNS glue and the repository is the legitimate source for that data.
> 
> - If an external host is needed for delegation purposes, it can be
> associated with a domain object as an attribute of the object 
> with no host
> object needed in advance.  There's no need to create an 
> external host object
> ahead of time, no need to worry about IP addresses, etc.
> 
> This solution does not allow object-based management of 
> external hosts,
> which means that renaming the external host would need to be done on a
> per-name basis.  It may address the other issues that people 
> have talked
> about on this thread, though.
> 
> I know this means that a domain can be associated with hosts 
> as objects and
> host as attributes and some people think that's inconsistent. 
>  I don't think
> it is if you agree with the first point above.
> 
> -Scott-
>