Re: [dhcwg] Client behavior with Parameter Request List

Ted Lemon <mellon@nominum.com> Thu, 20 May 2004 18:40 UTC

Received: from optimus.ietf.org (www.iesg.org [132.151.1.19]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA10681; Thu, 20 May 2004 14:40:47 -0400 (EDT)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org) by optimus.ietf.org with esmtp (Exim 4.20) id 1BQsE5-0008Ma-Cs; Thu, 20 May 2004 14:25:05 -0400
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org) by optimus.ietf.org with esmtp (Exim 4.20) id 1BQsAz-0007LN-L3 for dhcwg@optimus.ietf.org; Thu, 20 May 2004 14:21:54 -0400
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA08952 for <dhcwg@ietf.org>; Thu, 20 May 2004 14:21:50 -0400 (EDT)
Received: from ietf-mx.ietf.org ([132.151.6.1] helo=ietf-mx) by ietf-mx with esmtp (Exim 4.32) id 1BQsAw-00079h-IK for dhcwg@ietf.org; Thu, 20 May 2004 14:21:50 -0400
Received: from exim by ietf-mx with spam-scanned (Exim 4.12) id 1BQs9w-00073V-00 for dhcwg@ietf.org; Thu, 20 May 2004 14:20:49 -0400
Received: from toccata.fugue.com ([204.152.186.142]) by ietf-mx with esmtp (Exim 4.12) id 1BQs90-0006xK-00 for dhcwg@ietf.org; Thu, 20 May 2004 14:19:50 -0400
Received: from [128.177.194.120] (dhcp-120.engr.nominum.com [128.177.194.120]) by toccata.fugue.com (Postfix) with ESMTP id E62DE1B2765; Thu, 20 May 2004 13:06:08 -0500 (CDT)
In-Reply-To: <EJEFKKCLDBINLKODAFMDAEOJDEAA.rbhibbs@pacbell.net>
References: <EJEFKKCLDBINLKODAFMDAEOJDEAA.rbhibbs@pacbell.net>
Mime-Version: 1.0 (Apple Message framework v613)
Content-Type: text/plain; charset="US-ASCII"; format="flowed"
Message-Id: <481EBF5A-AA8A-11D8-9337-000A95D9C74C@nominum.com>
Content-Transfer-Encoding: 7bit
Cc: Dhcwg <dhcwg@ietf.org>
From: Ted Lemon <mellon@nominum.com>
Subject: Re: [dhcwg] Client behavior with Parameter Request List
Date: Thu, 20 May 2004 11:19:49 -0700
To: rbhibbs@pacbell.net
X-Mailer: Apple Mail (2.613)
X-Spam-Checker-Version: SpamAssassin 2.60 (1.212-2003-09-23-exp) on ietf-mx.ietf.org
X-Spam-Status: No, hits=0.0 required=5.0 tests=none autolearn=no version=2.60
Content-Transfer-Encoding: 7bit
Sender: dhcwg-admin@ietf.org
Errors-To: dhcwg-admin@ietf.org
X-BeenThere: dhcwg@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/dhcwg>, <mailto:dhcwg-request@ietf.org?subject=unsubscribe>
List-Id: <dhcwg.ietf.org>
List-Post: <mailto:dhcwg@ietf.org>
List-Help: <mailto:dhcwg-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/dhcwg>, <mailto:dhcwg-request@ietf.org?subject=subscribe>
Content-Transfer-Encoding: 7bit

On May 20, 2004, at 9:47 AM, Richard Barr Hibbs wrote:
> "A client MUST send a consistent Parameter Request List if
> it expects the DHCP server to provide consistent responses
> to DHCPDISCOVER and DHCPREQUEST messages.  If a client does
> not send a consistent Parameter Request List, the server MAY
> choose to resolve any conflicts according to [local]
> implementation policy, which is not specified by this memo."

Yes, we should definitely update the text.   I think it needs a bit 
more thought, though - what are we really trying to accomplish?   I 
think there is very little we actually can require here without placing 
an undue burden on implementation.   I would suggest that the only time 
that consistency between DHCP server responses is even desirable is 
between the initial DHCPOFFER and DHCPACK.   As long as the client 
sends the same options in its DHCPDISCOVER that it sends in its 
DHCPACK, it should get the same answers back.   The server should not, 
however, be required to implement this behavior across restarts or 
reconfiguration events, although it certainly may.   The client should 
be allowed to restart the configuration process if it receives 
inconsistent results, and the server should be required to be 
consistent in its response if it has _not_ been reconfigured.   
Consistency from the initial DHCPACK to the renewal DHCPACK isn't 
really possible, IMHO, so we shouldn't require it.   That is, we could 
just save what we sent to the client last time, but that actually would 
be really bad, and we've seen the badness that it causes, because old 
versions of Windows assumed parameters would not change.   What happens 
is that when network server addresses change, the clients don't pick up 
the change, and this results in a management nightmare.


_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
https://www1.ietf.org/mailman/listinfo/dhcwg