Re: [dhcwg] REMINDER: dhc WG last call on draft-ietf-dhc-relay-id-suboption-04.txt

[Mark Stapp <[e-mail]>]

Hi Kim,

thanks for the great comments. I'm posting an updated version of the draft.

Kim Kinnear wrote:
> 1.  In section 3.2, Bulk Leasequery, the world has moved on since
> these words were written. [...]
>   
done - converged to the one-true-draft.
> 2.  In section 5, Relay Identifier Types, it would be useful to
> simply include a table with the defined numeric values for
> RELAY_IDENTIFIER_DUID and RELAY_IDENTIFIER_ASCII.  IANA isn't
> going to change these values, and it seems difficult (and
> unusual) to send the implementor off to the IANA section to find
> the actual values of these constants.
>   
Ah - I've actually gotten pushed the other way recently about the 
handling of IANA values. The current 'best practice' seems to be to keep 
all the actual number assignments in one place - in the IANA section - 
and to use tokens elsewhere in the document. Unless you feel it's really 
confusing, I'd like to leave the IANA material as-is: I'm afraid I'd 
just have to change it back during IESG review.
> 3.  In section 7, Identifier Stability, it might be useful to
> discuss something about Identifier Continuity in the context of
> stability.  You have used the Industrial Ethernet as a use case
> (which I think was a good idea), and one of the difficulties of
> this particular use case is what to do when a relay agent has to
> be replaced.  This would seem to be the place to specify some
> behavior as to how to handle one device taking over for a failed
> device.  It would seem to me that allowing (or requiring) devices
> to be able to generate a DUID that matches that of a previous
> plug-compatible (or at least port-compatible) device might be
> useful here.
>   
Yes, thanks - that's a very good point. I've added another paragraph to 
that section to describe the situation where a relay device is replaced, 
and explain that this is a reason for implementors to permit admin 
configuration of the identifier value.

Thanks,
Mark

_______________________________________________
dhcwg mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/dhcwg