On Wed, Jun 30, 2004 at 09:04:01AM -0700, Ted Lemon wrote:
> As a practical matter, I think that the ISC server is the only DHCP 
> server that treats hardware_addr = 1:0:3:4:5:9:7 as not the same client 
> identifier as uid = 1:0:3:4:5:9:7, when no client identifier option is 
> present.   So I think you would be on firm ground if you tweaked the 
> ISC server's lease allocation behavior to match this.   I have never 
> heard of a problem report being submitted as a result of treating the 
> hardware address as if it were a client identifier when no client 
> identifier is present.

On my 'wish list' for 3.1.0 is configuration syntax to allow the
entity who maintains the dhcpd.conf to do what they please.

Wether that be fashioning a mock client-id for certain clients
which the server uses from then out to match leases, or stripping
client ids from certain clients so that the server does not use
them for lease matching.

I see this approach as more robust since the future is
draft-ietf-dhc-3315id-for-v4, and the reality I perceive is that
there may only be a lot more confusion, particularly with the
pervasiveness of PXE or similar (two-stage bootstraps that don't talk
to each other, and are inter-inconsistent).


I'm still not of a mind that the behaviour you suggest will be
out-of-the-tarball default.

Although it seems like a statement to that effect could appear in
the aforementioned draft, now that you mention it.

-- 
David W. Hankins		"If you don't do it right the first time,
Operations Engineer			you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins


-- Attached file included as plaintext by Ecartis --

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (FreeBSD)

iD8DBQFA4u67cXeLeWu2vmoRAo4HAJ4kALNMQyyD/twK1iZQ142SdiNB5ACgkDuh
Datj2v2axs027FzvygYznt0=
=+8Y7
-----END PGP SIGNATURE-----