Illustrated TCP/IP by Matthew G. Naugle Wiley Computer Publishing, John Wiley & Sons, Inc. ISBN: 0471196568   Pub Date: 11/01/98

Chapter 48
Subnetting Examples (Classes A, B, and C)

Any of the classes can be subnetted, although some are easier than others. The slide shows the three classes of networks, each with an address. This time, each of the addresses has been assigned a subnet mask. A mask is a series of bits that are applied (known as ANDing) to a portion of the address. This portion is what we are subtracting from the original address. It indicates how many bits we are masking out of the original host portion of an address to use as a subnet address. A subnet address is a real network number, but simply a network under the class address. Subnet masks are variable in length and move from the first host bit to the last. In other words, they move to the right of the address. Moving a mask to the left of the network address, beyond its natural mask, is known as supernetting (this concept will be discussed in a moment).

In this example of the Classes A and B addresses, I have shown all available bits following the network ID portion of the address used to indicate a subnet. The Class C address uses the first 3 bits of the host portion of the address for the subnet. With any of the addresses, any of the host bits (except for 2 bits at the end of the address; there must be at least one host on a network) may be used for subnetting. For example, a Class B address may use all of the third octet and 2 bits of the fourth octet for subnetting. This would give 1024 possible subnetwork numbers—yes, 1024. Those who are paying attention here should have caught the fact that in order to have 1024 subnet addresses we must use all 0s and all 1s in the subnet field as valid subnet addresses. This may seem contrary to host and network ID assignment, but it is not. All 0s and all 1s are allowed to be used in the subnet portion of any address (they still cannot be used in the host or network portions of the address as unique addresses). Refer to RFC 1812. This causes problems with subnet broadcasts, which I’ll explain later. Using the preceding example (10–bit subnet on a Class B), each subnet can support up to 62 hosts (63 would indicate a broadcast).

Subnetting Examples (Classes A, B, and C)

Subnet considerations:

1.  Hosts and routers must implement subnetting (there is a way around this discussed under Proxy ARP) and locally must have the same mask.

2.  The router must be able to distinguish between all 1s as a subnet address and a subnet broadcast.

3.  In some situations, the routing update protocol must support it.