Windows Forum

Ask something
Discussion Locked

Adding a 'subnet' to a private class c

0 helpful
7 replies

by bitbucketboy

Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

Okay - this isn't really adding a true class-c subnet, but rather more like treating a private class c as a class b.

Our small network is reaching the available threshold of dhcp/all addresses. We have a private network at address: 192.168.1.0.All of the servers, several print servers, and the router/switches occupy static addresses in the lower numbers. Above about .70, we have a scope of DHCP addresses.

Question: If we want to avoid migrating to a class b (or a) private address, canwe simply add the 192.168.0.0 or 192.168.2.0 addresses to the dhcp server, and change the subnet mask on the DHCP server from 255.255.255.0 to 255.255.254.0 or drop more bits for future ranges?

Reason: We have devices spread city-wide with staticaddresses. Migration would cost some down-time. Our firewall supports addresses: 192.168.any.any for internal routing. If we could do this, we would have more addresses than necessary.

Possible Implications: Would devices assigned a .0.0 address communicate with the .1.0 devices, as long as the subnet mask is the same and allows for both?

7 replies

by cavedweller

Subnetting is great for isolating local traffic from you inter-site links. I would try to assign each site it's own "subnet" (3rd octet) and continue to use class C netmasks. Connect all the sites together with a switch. Even if you only add a bit or two to the node address you're going to have to update each machine with a new netmask. Setup a DHCP pool for each site on a local server for maximum reliability. Update your DNS/WINS server and convert one site at a time to it's new "subnet". The rest of the network should work interrupted.

Whatever you do, plan for expansion or you'll find yourself in the same situation again.

by bowieb

You may be able to get away with an illegal subnet like that, but you could run into problems with equipment that doesn't recognize it. If you change the subnet, you are going to have to make the change on all of the PCs anyway.

What I would suggest is to migrate to a class B (172.16.0.0). You can configure a router to route between the two subnets on your LAN while you are making the switch. This would allow all of your computers to communicate even when some have been changed and some haven't.

If you have a DNS server, don't forget to change the addresses there as well.

by bitbucketboy

Much apprec. I'll likely try this on a limited scale. Won't DHCP fill in the new subnet on hosts at lease time? Tks for the advice!

R

by ebob

Even if your equipment could handle this, you're heading for lots of unforeseen problems. Your Internet connection will not like you, any MAN/WAN links you have across "public" networks will not like you, as soon as omebody else brings in a box thatlooks at the network address 192.168.a.b, ignores the "non-standard" subnet mask you will hqave connectivity problems, and so on. Soon you'll go mad, take up drinking excessively, and wind up living in a dumpster.

No, no.

It's much better to go through the pain of redesigning your address allocation, and get that out of the way. By the way, it's not necessarily as simple as migrating to a class B, as they are likely not available for use on the Public network(s).

Here's what I would do:
- Pick one of the RFC1918 non-routable address spaces
- Subnet it to match your organisation (preferrably with one subnet per location - use variable length subnet masks (VLSM), if your routers support this)
- Use your existing class C to definethe MAN/WAN links and your internet access
- use 255.255.255.252 to define your links
- use something suitable to define your Internet link
- Use a firewall that supports NAT and PAT to connect to the Internet, using a subnet provided by yourISP (so as not to expose your internal structure)

This way you'll have room for growth. Short term pain for long term gain.

Here's some good info on subnetting:
http://www.microsoft.com/windows2000/library/resources/reskit/samplechapters/cnbb/cnbb_tcp_pnmz.asp

by bitbucketboy

Yeah, we are using a type of NAT now on our firewall to support the "private" class-c. Wouldn't suggest our idea for public routing, but do you forsee equipment problems within our own network? Maybe our own print servers will hate it. Tks for the advice!

R

This discussion is closed to new replies.