Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I'm using 192.168.x.x and have exhausted the IP pool. Since is all private, why can't I just change the mask to and start allocating out 192.168.x+1.x IPs?

Edit: I realise that this would preclude me from linking to any other 192.168.x.0/24 networks

share|improve this question
Who told you that you couldn't? –  NickW Apr 24 '14 at 14:42
No-one except convention. I've never seen a 192.168/16 internal range and I'm nervous. –  BlueCompute Apr 24 '14 at 14:43
The only issue you might have would be your router (if it's not serious) choking on NAT sessions, a /16 is huge though, why not try a /22 first? –  NickW Apr 24 '14 at 14:47
Why not add static routes for other /16 networks? –  Bert Apr 24 '14 at 14:49
You are helping Cisco, Juniper, and others promote their "flat network" back to basics approach without even knowing it. –  TheCleaner Apr 24 '14 at 15:08

3 Answers 3

up vote 9 down vote accepted

Of course you can use a /16, or /23, /22 etc. if you want, I know we have at least a couple of /23's in place and it works fine. Obviously you'll have to alter your netmasks everywhere and especially on your router/s but this is very commonly done.

I personally wouldn't go the whole hog and use a /16, as you say it could cause problems if you're using 192.168.x.y addresses elsewhere, so just choose a /23 or /22 and that should last you a long while.

share|improve this answer
Also keep in mind to update odd and end devices like printers and multifunction devices. Seems those get forgotten the most. –  Nixphoe Apr 24 '14 at 14:48
Agree, and regarding the /23 or /22, since the OP didn't state WHICH 192.168.x.0/24 network they are using, they need to be aware of what it would mean to switch to such a mask, or even the /16. If they started with a and changed it to a /22 they need to be aware that the range is going to encompass now and won't start with 2.0. The OP just needs to understand basic subnetting to handle their need. –  TheCleaner Apr 24 '14 at 15:00
Thanks, I was stuck in a classfull IP addressing frame of mind and looking at going to 172.16 or 10. As it happens the network is on 192.168.0.x, so I can go to /23 and only change the subnet masks which will hopefully make life slightly easier. –  BlueCompute Apr 24 '14 at 15:13
@BlueCompute Since you actually hit the max, just doubling your pool (/23) seems a bit low. I agree you probably don't need a /16, but as long as you're changing the mask anyway, might as well make it a /22 or even /20. –  Kevin Apr 24 '14 at 17:40

Nothing stops you doing this - with private address ranges, such as the netblock and the netblock. Technically you could use non private address space - especially behind an address translating firewall - but it's rude and may become a massive nuisance. is a common address range to use. But that's all. People commonly use /24 address blocks, because it neatly maps to the IP address range octets. A netmask is a bit mask that says the first (24 in this case) bits are 'network' and the remainder are 'local subnet'.

That basically means is - Another common convention is to adopt '.1' as the gateway, and '.255' as the broadcast. Drop your netmask to 23 bits though, and you double the size of your subnet - - Each bit doubles the size. So whilst a /24 contains 256 addresses (8 'bits worth') - a /16 leaves 16 bits for host - 65,535 addresses.

The same is true of /16 and /8 - they map nicely to the octets of the dotted quad IP format. They are known as class A, class B and class C address ranges and once upon a time, when the internet was a much smaller place, that's all there was. Which is why you'll see a lot of (established) organisations have a /16 or class B, despite - probably - not using most of it on the public Internet.

share|improve this answer
While I applaud your detail, you should start with the "answer" and then include the necessary or extraneous details after the answer itself. So I would suggest making your "Nothing stops you doing this" your first sentence and explain why the OP can do this, rather than giving an explanation of what subnetting is. –  TheCleaner Apr 24 '14 at 15:04
Fair point. Edited for clarity. –  Sobrique Apr 24 '14 at 15:07
Also, it is not entirely true. It's not the subnet size /8 /16 /24 that defined classes A, B, C, it's also the address space assigned to it. Have a read here. And there is no such Thing as class A, B, C anymore. –  MichelZ Apr 24 '14 at 15:13

The whole subnet class thing (A, B, C) is something from the past See wikipedia. It was superseded by CIDR (Classless Inter Domain Routing) Wikipedia.

This means that in "today's" world, you can subnet your private network with whichever network mask that suits you.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.