(SUBNETwork mask) The technique used by the IP protocol to create a subnet address. The subnet mask is a binary pattern that is stored in the client machine, server or router. It is matched with the IP address of a packet to determine which network segment the packet is destined for. See subnet.

A Tradeoff Between Hosts and Subnets
Depending on the network class (A, B or C), some number of bits are reserved for hosts and subnets, and these bits become a tradeoff. The more hosts, the fewer the subnets can be created; the more subnets, the fewer the hosts can be individually addressed.

A Subnet Mask Example
The subnet mask below is a Class C address, which uses the first 24 bits for network ID and the last 8 for host ID. These last 8 can be divided between hosts and subnets. In the default mask below, the 0 means there are no subnets and up to 254 hosts can be addressed.

In the second example, the 224 reserves the three high-order bits of that field for subnets, leaving the remaining 5 bits for hosts. This 224 pattern creates subnets in the range of 001 to 110 (000 and 111 are reserved). As such, 6 subnets of up to 30 hosts in the range of 00001 to 11110 can be addressed. Once again, 00000 and 11111 are reserved: all zeros mean "this" node, and all ones mean "all" nodes (broadcast), which is why calculations for maximum hosts and subnets are always subtracted by 2. For a list of all Class B and C subnet masks, see subnet mask tables.

  Class C Default Submask
  255.255.255.0
  11111111.11111111.11111111.00000000

  Class C Subnet Mask (6 Subnets/30 Hosts Each)
  255.255.255.224
  11111111.11111111.11111111.11100000