Redundant DHCP Configuration

I have two domain controllers one of which is running 2003 SP1 and the
other is running 2003 R2. We want to have both running DHCP for
redundancy on our network so that either server can provide DHCP if the
other is not running.

Which of the following options is possible or correct to follow:

1. Have the two servers each control a different subnet. This
apparently requires specifying the entire range of both subnets in each
server's scope and excluding the range of addresses of the opposite
server's subnet so that the opposite server can service requests on one
subnet (The subnet mask is set so that both subnets are included
together

e.g.
server 1 issues addresses on 192.168.0.x
server 2 issues addresses on 192.168.1.x
subnet mask is 255.255.254.0

2. Have the two servers both work on the same subnet, if that is
possible. The assumption being that the information can be replicated
just as it can with DNS.
This would probably be simpler to set up as our current organisation
has always been around one subnet.

T.I.A.

Both configured identically with the ability to serve the whole LAN (all
subnets).
Configure Exclusions so that each server gives out 50% of the addresses.
One server goes down remove the Exclusions on the remaining one,...when the
one is repaired, put the Exclusions back as they were.
It will always be a manual process.

--
Phillip Windell [MCP, MVP, CCNA]
www.wandtv.com

The views expressed are my own (as annoying as they are), and not those of
my employer or anyone else associated with me.
-----------------------------------------------------

"NZSchoolTech" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) ups.com...
>I have two domain controllers one of which is running 2003 SP1 and the
> other is running 2003 R2. We want to have both running DHCP for
> redundancy on our network so that either server can provide DHCP if the
> other is not running.
>
> Which of the following options is possible or correct to follow:
>
> 1. Have the two servers each control a different subnet. This
> apparently requires specifying the entire range of both subnets in each
> server's scope and excluding the range of addresses of the opposite
> server's subnet so that the opposite server can service requests on one
> subnet (The subnet mask is set so that both subnets are included
> together
>
> e.g.
> server 1 issues addresses on 192.168.0.x
> server 2 issues addresses on 192.168.1.x
> subnet mask is 255.255.254.0
>
> 2. Have the two servers both work on the same subnet, if that is
> possible. The assumption being that the information can be replicated
> just as it can with DNS.
> This would probably be simpler to set up as our current organisation
> has always been around one subnet.
>
> T.I.A.
>