Classful IP addressing was the original method used in IPv4 to divide the 32-bit address space into fixed-size network blocks called classes. The first few bits of an address determined its class, which in turn defined a default network mask and the number of possible hosts on that network.
There were five classes in total:
- Class A addresses began with a leading 0 bit (ranges 1.0.0.0 to 126.255.255.255) and used a default mask of 255.0.0.0 (or /8), providing over 16 million host addresses per network. These were intended for very large organisations or major internet backbones.
- Class B started with bits 10 (ranges 128.0.0.0 to 191.255.255.255) and used the mask 255.255.0.0 (/16), offering up to 65 534 hosts per network, suitable for medium-sized enterprises.
- Class C began with bits 110 (ranges 192.0.0.0 to 223.255.255.255) and had a mask of 255.255.255.0 (/24), allowing 254 hosts, ideal for small business or departmental LANs.
- Class D (224.0.0.0 to 239.255.255.255) was reserved for IP multicast groups and did not follow standard host numbering.
- Class E (240.0.0.0 to 255.255.255.255) was set aside for experimental or future use and not assigned to hosts.
Under classful rules, every network used its default mask. There was no way to carve out smaller or larger subnets without borrowing bits from the host portion, leading to significant waste when a network’s needs did not match the rigid class sizes. For example, a department requiring a few hundred addresses still had to consume an entire Class B block of 65,000 addresses.
By the early 1990s the inefficiency of fixed classes contributed to rapid depletion of available IPv4 space. This led to the adoption of classless inter-domain routing (CIDR), which allows variable-length prefixes and more precise allocation of addresses. Although classful addressing is largely obsolete today, understanding it provides historical context for how IP addressing and subnet masks evolved into the flexible, scalable system in use now.