Designing Home and SME Networks 7 – Local Network IP Addressing Strategies
OK, so we use private IP addresses for the local network. What are the available private address blocks?
These are defined in standard RFC1918, and there are 3 private blocks available:
192.168.x.x, (which provides up to 65,536 addresses)
172.16.x.x – 172.31.x.x, (which provides up to 1,048,576 addresses)
10.x.x.x, (which provides up to 16,777,216 addresses)
(where x is any number from 0 to 255)
Any one of these blocks provide far more private IP addresses than any but the largest organisation (such as IBM!) could ever use!
So, which do I use?
Normally, you should use the first one. Why? No real reason – there’s no standard to cover it – but it’s the smallest block and you almost certainly won’t want to use it all. Hence, choose the 192.168.x.x block.
How do I use it, or rather, a bit of it?
Here you need to be introduced to what, in the trade, is called ‘subnetting‘ (technically ‘Classless Inter-Domain Routing’ or ‘CIDR’).
A router has a clever bit of software built in to it that allows it to define a ‘subnet mask’ for each address block, which is used to determine the size of the address block. This has the extremely useful function of being able to define blocks of addresses of varying sizes that are all treated the same. The subnet mask (netmask) is a 32-bit number, normally written in the same 4 octet format as an IP address, with each bit that represents a subnet address set to 1 and every other bit, that represents a device or node address, set to a zero. The blocks of 1s and 0s must be contiguous. So, for example, a subnet that allows 256 addresses (254 device nodes) would be written 255.255.255.0 (Binary 1111 1111 1111 0000) and one that allows four addresses (two device nodes) would be 255.255.255.252 (Binary 1111 1111 1111 1100).
Subnets that support only two node addresses would be used for implementing point-to-point links, but this is rarely done in domestic or SME installations.
Traditionally, the first address in each subnet is reserved for the network ID and the last address in each subnet is reserved as the Broadcast address (the address to use to send a message to all members of the subnet) so are not used for hosts (attached devices).
The notation used to define a subnet is either Network ID and netmask or the Network ID followed by a slash and the number of bits in the subnet address. So, for instance, the first 256 address block in the 192.168.x.x private IP address block could be defined as 192.168.0.0 255.255.255.0 or 192.168.0.0/24. The second representation is obviously simpler.
Typically, a home or SME network would use one block of 256 addresses as a subnet. This would allow up to 254 hosts (PCs, networked peripherals, routers etc) to be connected to a single segment. Most home networks use one /24 block out of the 192.168.x.x private address space – frequently 192.168.0.0/24 or 192.168.1.0/24.
There is no real reason to choose any one subnet in preference to another, unless you connect directly to other people’s networks or some of your devices have pre-set and unchangeable IP addresses (very rare nowadays).
Of course, it couldn’t be that simple and the pending introduction of IPv6 (q.v. article 6 in this series) will change much of the traditional methodology described above. Don’t worry about it ’till it happens!