Subnetting
How IP Addressing Became Less Wasteful

Chapter1

When the Internet Protocol (or IP) was first introduced, there was a great deal of inflexibility in how IP addresses could be assigned. Though there were over 4 billion IP addresses available, they were split up in such a manner that there was a great deal of waste. As the Internet began evolving from a research tool used by the government and universities, the waste needed to be eliminated. A new version of addressing was in the works, but it would be years before it was ready and years after that until it could be implemented.

Something was needed to stem the tide of IP address waste until a new addressing scheme could be implemented. There had to be a way to distribute addresses in a way that was more closely tailored to the needs of groups receiving them. For this reason, subnetting was introduced.

IP subnetting is defined in RFC 950. It was originally introduce in 1985 and was later updated in RFC 1812 in the middle of 1995 to include Classless Inter Domain Routing (or CIDR). Its purpose was simple: to break down large blocks of IP addresses into smaller, more easily manageable blocks.

The original subnets were known as classful subnets. This refers to the boundaries between the sections of an IP address. When writing out an IP address, there are four sections, known as octets because each contains 8 binary digits. In base ten (or decimal) format, they are written out along the following lines: 192.168.100.100. The octet boundaries became known as Class A, Class B, and Class C.

In a Class A subnet, only the first octet remains the same. In other words, if you have a Class A block starting with 192.0.0.0, you would have every address up to 192.255.255.255. That's over 16 million addresses, which is many more than all but the largest of organizations would ever need. Since there are only 256 Class A blocks possible, it was hardly feasible to give them out to anyone who wanted them.

Class B and Class C subnets are much the same, except that a Class B's first two octets are static and a Class C's first three blocks are static. This provides 65 thousand or 256 addresses, respectively. While slightly more manageable, this still provided many more addresses than small companies would need and placed an undue burden on Internet Service Providers (or ISPs) who were responsible for routing traffic to them.

Subnet masks are the method by which a router is able to tell how many IP addresses are in a block. They are written out to show how many octets are free in a particular block. They are written out in a form similar to the IP addresses themselves, as they are in four octets, separated by periods. A Class A subnet mask is 255.0.0.0, stating that the first octet is 'full' and the other three are empty. Likewise, a Class B mask is 255.255.0.0 and a Class C mask is 255.255.255.0.

Continue in Chapter - 2