Classful vs Classless: The History That Changed IP Addressing Forever

Today we take variable-length subnet masks for granted. But for the first two decades of the internet, every IP address came with a fixed, predetermined network size — and that rigidity nearly caused the internet to collapse under its own growth. Understanding the shift from classful to classless routing explains why CIDR exists and why every modern network engineer needs to understand it.

The Classful Era (1981–1993)

RFC 791, published in 1981, defined three main classes of IPv4 network:

• Class A (first bit 0): addresses 0.0.0.0 to 127.255.255.255 — 128 networks of 16 million addresses each
• Class B (first bits 10): addresses 128.0.0.0 to 191.255.255.255 — 16,384 networks of 65,536 addresses each
• Class C (first bits 110): addresses 192.0.0.0 to 223.255.255.255 — 2,097,152 networks of 256 addresses each

The class was encoded in the first bits of the address itself — routers could determine the network boundary without any additional information. It was elegant and simple. It was also catastrophically wasteful.

The Waste Problem

Imagine a university needing addresses for 5,000 hosts. A Class C block held only 254 usable addresses — far too few. The only option was a Class B block with 65,534 usable addresses. The university used 5,000 of them and the remaining 60,000 were stranded, allocated but unusable by anyone else. By the early 1990s, Class B addresses were being consumed at an alarming rate and routing tables were exploding in size.

CIDR's Solution (1993)

RFC 1519, published in September 1993, introduced Classless Inter-Domain Routing. Instead of encoding the network boundary in the address itself, CIDR used an explicit prefix length — the slash notation you use every day. A network needing 5,000 addresses could receive a /19 block (8,190 usable addresses) rather than a full Class B. This dramatically reduced waste and allowed ISPs to aggregate multiple /24 blocks into a single /20 announcement, keeping routing tables manageable.

Supernetting and Route Aggregation

CIDR also enabled supernetting — combining multiple smaller networks into one routing advertisement. An ISP holding eight /24 networks could advertise a single /21 to upstream providers. This reduced the size of global BGP routing tables and made internet routing more scalable. See our guide on CIDR aggregation and supernetting for the mechanics.

Legacy of the Classful Era

Classful thinking still surfaces in documentation and tooling. You will see references to "Class A private range" meaning 10.0.0.0/8 or "Class C" meaning any /24 network. The IPv4 address classes guide covers the legacy details. Today these are historical descriptions — no modern routing protocol uses class boundaries.