The Old Way: Classful Addressing's Rigid Tiers
Imagine buying T-shirts, but you can only choose Small, Medium, or Large. Need something in between? Too bad. This was the world of classful addressing, the original system for organizing IPv4 addresses introduced in 1981. The entire address space was carved
into five fixed classes: A, B, C, D, and E. For network administrators, only A, B, and C mattered for assigning addresses to devices. The class of an IP address was determined by the first few bits of the address itself. This meant you didn't even need to state the subnet mask; it was implied. A Class A address always had a /8 mask (over 16 million hosts), Class B was always /16 (65,534 hosts), and Class C was /24 (254 hosts). This system was simple, but its rigidity was its downfall.
The Breaking Point: Waste and Inefficiency
The classful model quickly created two massive problems. The first was staggering address waste. If your company needed 500 IP addresses, a Class C block was too small. Your only option was to get a Class B block, which gave you over 65,000 addresses. This resulted in over 99% of your allocated addresses sitting unused, a disaster for the finite IPv4 pool. The second problem was the explosion of the internet's routing tables. With classful routing, every individual network needed its own entry in a router's table. As the internet grew, these tables became unmanageably large, threatening to slow global routing to a crawl. The internet was becoming a victim of its own success, and the rigid, wasteful class system was the culprit.
The Fix: Classless Addressing and CIDR
To solve this crisis, the Internet Engineering Task Force (IETF) introduced Classless Inter-Domain Routing (CIDR) in 1993. CIDR is the system we all use today, and it’s based on a simple but revolutionary idea: flexibility. CIDR decoupled IP addresses from their implied class, allowing administrators to define a network of any size using a prefix. This is what the slash notation (e.g., /24, /26, /19) represents. Instead of fixed Small, Medium, and Large T-shirts, you could now get fabric cut to your exact measurement. This is achieved through Variable Length Subnet Masks (VLSM), which let engineers create subnets of different sizes from a single block, allocating addresses with surgical precision. If you needed 500 addresses, you could get a /23 block, which provides 510 usable hosts—a near-perfect fit with minimal waste.
In Production: Explicit vs. Implied Information
This is where the difference becomes tangible in a production network. Old, classful routing protocols like RIPv1 didn't send subnet mask information with their routing updates. A router running RIPv1 would receive an update for 10.1.1.0 and assume it belonged to the entire 10.0.0.0/8 network because the address starts with '10'. It had no concept of a more specific subnet like 10.1.1.0/24. This made complex subnetting across a large network impossible. In stark contrast, modern classless routing protocols like OSPF and BGP are explicit. When they share routing information, they include the prefix length (the /24 part). The router receives an update for 10.1.1.0/24 and knows exactly the size and boundary of that network. This allows for precise routing and a powerful technique called route summarization (or supernetting), where multiple smaller networks can be advertised as a single, larger block, drastically shrinking routing tables and making the internet more efficient.
