Network Class

In the early days of IP, networks were delegated in classes. There was no such thing as a netmask that we use today. Nowadays, however, network classes have become obsoleted by better methods of addressing (such as CIDR). Nevertheless, learning about network classes can help one's understanding of the IPv4 address-space and how it became the way it has.

In addition to the number of possible hosts on a network, information about the number of possible networks was embedded into the IP address itself. An address consisted of a network part and a host part. The network part was itself partitioned to describe the network's class, by the leading number of bits. For each class (decreasing A through E), the host part was in multiples of eight bits (starting from 24 and going 0). The leading bits represented numbers that were multiples of two (with the exception of Class E which has four bits representing the number seven), starting from one bit that was zero.

Examples
For example, a Class A network had one leading bit that was necessarily 0. Then, there were 7 bits dedicated to the network. The rest (24 bits) were for hosts. This means that there were only 128 possible Class A networks, each with 24 bits allocatable by the owner. Class A networks were therefore highly sought after and highly limited by the IANA. So any IP address in the range 0.0.0.0 to 127.255.255.255 was considered as being part of a Class A network. For example, if someone told you that their IP address was 93.35.54.123, you would immediately knew that they were on the Class A network with a first decimal of 93 (in modern CIDR notation, we would call that network 93.0.0.0/8).

Similarly, any IP address in the range 128.0.0.0 to 191.255.255.255 was in a Class B network, and now, there were only 16 bits dedicated to hosts.

For Class C networks, the range was therefore 192.0.0.0 to 223.255.255.255 with only eight bits dedicated to hosts.

Special Classes
Now Classes A, B, and C described all the normal addresses anyone would use, but there are two more classes.

A Class D network had a range of 240.0.0.0</tt> to 239.255.255.255</tt>. What makes this unusual is that there are zero bits dedicated to hosts. How can that be you ask? Well, all Class D addresses are used only for multicast. This makes sense, as one multicast address can be shared by hosts, and each expects to receive it.

The last network class is Class E, which includes the rest of the address-space (240.0.0.0</tt> to 255.255.255.255</tt>) and is reserved by the IANA to experimentation.