Troy Thompson said:
Here's the problem. We're running out of IP address on our 10.53.1.0/24 subnet, and its
my job to research and figure out the most efficient way to add more.
We have 1 domain, and 2 sites, and 2 DCHP servers (one in each site) seperated by a
WAN. Our main site is using 10.53.1.0/24 and our remote site (not even a factor in this
equation really) is using 10.53.1.0/24
Unless that is a typo, both sites are in the same subnet so yes, the remote
site *is* a factor in this. Layer3 addressing doesn't care about
Geography,...they are all the same subnet, so logically they are all in one
big happy place together.
Ok. Well first, to avoid confusion, don't use the word Domain or Active
Directory in the converstation. There is absolutly no relationship to
Domains/AD and what you are doing (or want to do) with IP addressing.
So far, I think I have a few options. 1) Superscope- Don't really want to
do that because even Microsoft admits that it may cause network performance
issues.
After dealing with this myself, and after much researching, I have yet to
find a practical use for a Superscope. A Superscope "combines all the
Scopes inside it into one big fat Scope, therefore a client (any client) may
get an address from any one of those scopes within it,..therefore all of
those member Scopes must be logically in the same subnet or disasters will
happen.
2) Multiscope- This would add a new subnet thus causing tons of work
recofiguring devices that have static IP settings.
This is one way. Yes, it would mean a lot of work and possibly a redesign of
the physical topology and the addition of router(s) to route between the
subnets.
3) Adding a child domain (like for our shop floor PCs) and giving its own
seperate scope. I'm not even sure yet how well this would work, and any
problems it might cause other us having to reconfigure all the PCs that we'd
want on the child domain.
A Child Domain is still a "AD Domain" and is not relevant to the Layer3
addressing & routing scheme. Yes, you can design them to "parallel" each
other, but that is only for "human consumption",....the truth is that your
Layer3 addressing & routing scheme couldn't care less what you do with
Active Directory,...and Active Directory couldn't care less what you do with
your Layer3 addressing & routing scheme. Think about it,..what if you were
running a 100% Unix system which has no AD Domains and were running out of
IP#s?...what would be different?,...well absolutely nothing would be
different. You would still have the same problem and would still use the
same solution.
#4. You didn't have a number 4, but here's my other option. You can leave
your existing addressing in place and just change the subnet mask from
255.255.255.0 to 255.255.0.0. By changing this in the Scope and on
Statically assigned machines will vastly increase the number of host well
beyond what a smoothly function LAN could support with a single subnet.
This would probably work on your existing system and not even cause
disruption during the transition because a machine with a 255.255.255.0 mask
and a machine with a 255.255.0.0 mask with still talk to each other as long
as none of the currently existing addresses go higher than 10.53.1.254.
This of course would only exist briefly during the transition and you would
still want to get the masks to all agree as quickly as you can.
Your current DHCP Scope (10.53.1.1 -- 10.53.1.254) will have to be
re-created with the new Range of 10.53.1.1 -- 10.53.255.254.
You could do a few other tricky things with the mask to avoid wasting so
many address by using a mask like 255.255.252.0 (or similar) but things can
get complicated in a hurry if you aren't careful and plan it well,...So I'm
not even gonna go there,..that is too much for me to deal with while trying
to communicate using newsgroup messages.
