Re: How many folks ...

Rainer Bittermann <bitterm@dbktest1.net.db.de> Mon, 07 October 1996 08:01 UTC

Received: from cnri by ietf.org id aa22390; 7 Oct 96 4:01 EDT
Received: from marge.bucknell.edu by CNRI.Reston.VA.US id aa04138; 7 Oct 96 4:01 EDT
Received: from reef.bucknell.edu by mail.bucknell.edu; (5.65v3.2/1.1.8.2/17Jul96-0109PM) id AA25674; Mon, 7 Oct 1996 03:56:16 -0400
Date: Mon, 7 Oct 1996 03:56:16 -0400
Message-Id: <199610070655.IAA14268@dbktest1.net.db.de>
Errors-To: droms@bucknell.edu
Reply-To: dhcp-v4@bucknell.edu
Originator: dhcp-v4@bucknell.edu
Sender: dhcp-v4@bucknell.edu
Precedence: bulk
From: Rainer Bittermann <bitterm@dbktest1.net.db.de>
To: Multiple recipients of list <dhcp-v4@bucknell.edu>
Subject: Re: How many folks ...
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
X-Comment: Discussion of DHCP for IPv4
Mime-Version: 1.0

Hi everybody,

we evaluated DHCP for about 4 weeks and decided NOT to use it. The main reason
for skipping DHCP was the unsatisfying interaction between DHCP and DNS in the
implementations we considered to use (WinNT 4.02 and AIX 4.2).
In our environment we have to run about 30,000 PCs and we have to make an 
accounting of IP traffic based on IP-Adresses - so we depend absolutely on 
up-to-date entries in the DNS. The accurate DNS update is a "MUST" in our
environment (may be that's only our problem but I think every provider has the
same problems when using DHCP for his customers) - the DHCP servers we found do
not support any feature that allow to configure the server for a mandatory
DNS update upon a DHCPACK, with other words: you can distribute and use IP
addresses without an existing valid DNS entry.

Furthermore the lack of any authorization and authentification was the next
knock-out point. Anybody in a network can set up a DHCP server and clients 
are not able to distinguish "good guy" and "bad guy".

We decided to wait for new products providing these features which are about
to be defined just right now (see the latest drafts around these topics).
I think that we will have to wait at least to the mid / end of 1997 before
we can see some useable implementations - am I wrong ?

Regards

Rainer Bittermann
Nortel Dasa GmbH
Germany