I am looking to upgrade our company's server from SBS to 2xServer 2012 R2 in Hyper-V.
This will be quite a major move forward both in capacity and cost, so I am reviewing the current internal IP addressing structure whilst we are in the process.
Currently we have a single subnet of 10.25.0.0 / 20 (255.255.240.0)
I understand this is far too big and shouldn't be in a single subnet ideally. We only have around 40 devices on the network currently, so broadcasts probably not yet an issue though.
With the view that the company will possibly be going international, with branch offices / servers, I have looked at splitting our existing head-office network into 4 subnets.
This would also make cost effective use of our router+switches which support VLANs and max of 4 LAN's on the router. (Draytek Vigor 3200, Netgear GS724TS).
I believe I would then setup each port on the switch to assign VLAN "tag", and relevant settings on the router to match the ID, or could it be setup with all 4 subnets on the same physical network without the use of VLANs? (I think this might defeat the object).
The idea I had for our head office was to use:
10.25.0.0 / 23 - Network devices+servers
10.25.2.0 / 23 - Printers
10.25.4.0 / 23 - DHCP clients
10.25.6.0 / 23 - VPN clients
If we were to take out a branch office, I would possibly use 10.26.0.0, 10.27.0.0 etc...
Bearing in mind I would rather not totally change the class and addresses of every existing static device on the network (printers, servers, etc), what would your recommendation be to achieve this with minimal disruption? I'm hoping that changing just the subnet mask on servers / devices would be less of a job than a total IP address change on DNS/DHCP/WINS/Exchange etc.
We are not at the stage of having IPV6 public address yet, and a couple of devices on the network apparently don't support it either. This was another thought...
Thanks in advance.