Changing a subnet can be a daunting task so when you've come to the point where you have to change your company's subnet you want to be sure that it will take care of your needs now and into the future. Here are some problems to avoid:

- Subnet is public
- Subnet too large
- Subnet too small
- Subnet too common

Public IP addresses are assigned to devices on the Internet.  If your internal subnet is a public address range then your router will not route packets to those addresses on Internet and they will instead remain on you local network.  I worked with one comany that purchased an HP minicomputer.  It came pre-configured with an IP address and the company used the HP mini's address and subnet to build their IP infrastructure.  Whenever that company tried to access www.HP.com the packets never left the local network.  That made it tough to download updates from HP.  A Private IP address range should always be used for internal networks,  The private ranges are:

10.0.0.0 - 10.255.255.255
172.16.0.0 - 172.31.255.255
192.168.0.0 - 192.168.255.255

Company's are becoming more and more connected.  Companies connect when there are mergers or buy-outs, when there are collaberative projects and when one company provides a service for another that requires connectivity.  The problem with picking a subnet that is too large is that you increase the chances that you are going to collide with another subnet belonging to an organization that you connect to. For expample, Company-A has 200 IP devices on their network and their subnet is 10.0.0.0/8.  Company-B has 250 devices and their subnet is 10.25.14.0/24.  Now if Company-A and Company-B connect there will be routing problems becuase Network-A includes the same address space as does Network-B.

Network-A has a subnet that uses 16,777,214 addresses; way more addresses than is needed for the 200 devices on their network.  A Class B network mask yields 65,534 addresses so even that is many times more than is needed.  A Class C network mask gives us 254 address and that is too small to allow for growth and DHCP redundancy in this case.  A good rule of thumb is to create a subnet that is 5 times the number of IP addresses currently in use and then round up to the next subnet.  So for Company-A that comes to (200*5=1,000) addresses.  Round that to the next subnet number and you can determine that the subnet you need is 22 bits (yielding 1022 addresses) or a mask of 255.255.252.0.

You should probably not create a subnet any larger than 1022 addresses. Half these addresses are for redundancy so a subnet of this size should have only 511 devices on it but in my opinion that's pushing it. With Ethernet Switches segmenting a network we don't have to worry too much about collisions on large networks any longer but keep in mind that broadcast traffic crosses all segements in a network so every station is going to see every broadcast from every other station on the subnet. I would not put more than 511 devices in the same subnet and would feel more comfortable if it was more like 250 per subnet. I shudder to think of what it would be like to have a broadcast storm take down a subnet with 500 devices on it.

Why multiply the current IP devices by five?  First you want to have at least twice the number for DHCP redundancy.  If a DHCP server fails then you want to have a backup DHCP server with enough addresses available to cover all your devices.  Plan for growth by doubling the existing IP addresses.  You don't have to double the size of the business to double the number of addresses.  More devices are becoming IP enabled.  The future of the phone industry is VoIP so you can expect that you will need an IP address for every phone in your organization in the future.  So that's 2X for DHCP and 2X for growth or 4X the addresses.  I add one more multiplier for unanticipated growth and because IP addresses don't cost anything; that gives me a multiplier of 5X.  Evaluate your own situation, if you think there will accelerated growth in your company then add more addresses; just don't get carried away.

The obvious problem with selecting a subnet that is too small is that you soon run out of addresses and you have to redo your network.  Keep to the 5X rule.

You don't want to pick a subnbet that everyone elses is using.  It is human nature to start at the begining so you can expext there are a lot of subnets out there that are 192.168.0.0, 192.168.1.0, 172.16.0.0, 172.16.1.0., etc...  Engineers also like to make networks easy to type and remeber so I'd stay away from 10.10.0.0, 10.1.1.0 and the like.  Try to pick something obscure like 10.173.72.0/21.  Never use 192.168.0.0/24 or 192.168.1.0/24 as one of your company's internal subnets.  These subnets are widely used by home networks.  You're users will have problems when they VPN into your company network if the subnets are the same at home and in the office.

Serving These Central Wisconsin Communities and Surrounding Areas

Computer Network Services & Phone Systems

Stevens Point - Wausau - Marshfield - Waupaca -Schofield - Mosinee - Wisconsin Rapids - Shwano - Rhineladner - Amherst - Tomahawk - Plover - Merrill - Wood County - Marathon County - Portage County - Lincoln County - Waupaca County - Outagamie County - Winnebago County - Waushara County - Oneida County - Adams County - Marquette County