Subnetting

In the mid-1980s, RFCs 917 and 950 were released. These documents proposed a means of solving the ever-growing problem posed by the relatively flat, two-level hierarchy of IP addressing. The solution was termed subnetting. The concept of subnetting is based on the need for a third level in the Internet’s hierarchy. As internetworking technologies matured, their acceptance and use increased dramatically. As a result, it became normal for moderate and large-sized organizations to have multiple networks. Frequently, these networks were LANs. Each LAN may be treated as a subnet.

In such multiple-network environments, each subnetwork would interconnect to the Internet via a common point: a router. The actual details of the network environment are inconsequential to the Internet. They comprise a private network that is (or should be) capable of delivering its own datagrams. Thus, the Internet need only concern itself with how to reach that network’s gateway router to the Internet. Inside the private network, the host portion of the IP address can be subdivided for use in identifying subnetworks. Subnetting, as specified in RFC 950, enables the network number of any classful IP address (A, B, or C) to be subdivided into smaller network numbers. A subnetted IP address actually consists of three parts:
• Network address
• Subnetwork address
• Host address

The subnetwork and host addresses are carved from the original IP address’s host address portion. Thus, your ability to subnet depends directly on the type of IP address being subnetted. The more host bits there are in the IP address, the more subnets and hosts you can create. However, these subnets decrease the number of hosts that can be addressed. You are, in effect, taking bits away from the host address to identify subnetwork numbers. Subnets are identified using a pseudo-IP address, known as a subnet mask. A subnet mask is a 32-bit binary number that can be expressed in dotted-decimal form. The mask is used to tell end systems (including routers and other hosts) in the network how many bits of the IP address are used for network and subnetwork identification. These bits are called the extended network prefix. The remaining bits identify the hosts within the subnetwork. The bits of the mask that identify the network number are set to 1s and the host bits are set to 0s.

For example, a mask of 11111111.11111111.11111111.11000000 (255.255.255.192 in dotted-decimal notation) would yield 64 mathematically possible host addresses per subnet. The values of the right-most six bits, the ones set equal to zero, sum to 64 in the base 10 number system. Thus, you can uniquely identify 64 devices within this subnet. Only 62 of these addresses, however, are actually usable. The other two host addresses are reserved. The first host number in a subnet is always reserved for identifying the subnet itself. The last host number is also reserved, but is used for IP broadcasts within the subnet. Thus, you must always subtract 2 from the maximum number of hosts in a subnet to get the maximum number of usable host addresses per subnet.

The number of mathematically possible subnets, however, depends on what class of IP address is being subnetted. Each class reserves a different number of the available bits for the network number. Thus, each class offers a different number of bits that can be used for subnetting.