Are IP classes still used?

Only informally and as historical reference. Routing protocols, allocation systems, and operating systems all use CIDR notation. The class names persist in conversation but the technical reality is classless since 1993.

Why is the 10.x.x.x range called private?

RFC 1918 reserved it (along with 172.16/12 and 192.168/16) for private internal networks not reachable from the public Internet. The 10/8 was an early Class A reserved for this purpose; the size made it usable for very large internal networks.

Can someone still get a Class A?

Not from a Regional Internet Registry — the original free pool is exhausted. The only path is buying an existing /8 on the secondary market, which costs hundreds of millions of dollars. Realistically, no organization is acquiring new Class A equivalents.

Why doesn't 127.0.0.1 use just one address?

Historical accident. When 127/8 was reserved for loopback in 1981, the smallest reservable unit was a Class A. The 16 million addresses are effectively wasted on a function that needs only one (or a few). IPv6 corrected this with ::1/128 — a single address for loopback.

What changed when CIDR replaced classful?

Routing protocols had to be updated to carry prefix length (RIPv2, OSPF, BGP4). Allocation rules became finer-grained. Network engineers had to learn variable-length subnet masks. Internet routing tables became more efficient (CIDR aggregation reduced the number of routes). The transition took years; the technical and educational changes were substantial.

IPv4 Classful Addressing: The History Before CIDR

Before CIDR (Classless Inter-Domain Routing) showed up in 1993, IPv4 addresses were divided into rigid "classes" — A, B, C, and a few special purposes. The classful era only lasted 12 years from IPv4's standardization but its terminology persists in casual conversation. Understanding what classful addressing was clarifies the history of how the Internet got built.

Classful addressing was the IPv4 address allocation scheme from RFC 791 (1981) until effectively replaced by CIDR in RFC 1517-1520 (1993). The 32-bit IPv4 address space was divided into fixed-size classes based on the leading bits. The model was simple to understand and disastrous for address conservation.

The classes

Class A — leading bit 0. Network portion is 8 bits, host portion is 24 bits. Range 0.0.0.0 through 127.255.255.255. Each Class A network has 16,777,216 addresses. Only 128 possible Class A networks.
Class B — leading bits 10. Network portion is 16 bits, host portion is 16 bits. Range 128.0.0.0 through 191.255.255.255. Each Class B has 65,536 addresses.
Class C — leading bits 110. Network portion is 24 bits, host portion is 8 bits. Range 192.0.0.0 through 223.255.255.255. Each Class C has 256 addresses (254 usable).
Class D — leading bits 1110. Reserved for multicast addressing. Range 224.0.0.0 through 239.255.255.255.
Class E — leading bits 1111. Reserved for experimental use. Range 240.0.0.0 through 255.255.255.255.

The allocation disaster

The early Internet handed out IPv4 addresses according to which class fit the requester's expected size:

Need fewer than 256 addresses? Get a Class C.
Need 257-65536? Get a Class B.
Need more? Get a Class A.

The problem: a university expecting 500 addresses had to take a Class B (65,536) because Class C was too small. The remaining 65,036 addresses sat unused. Early Class B holders included many universities, research labs, and corporations that needed maybe a few thousand addresses but got allocations 20× larger.

The cumulative waste was enormous. By the late 1980s, the IPv4 free pool was visibly depleting, with most of the consumed addresses sitting idle in oversized allocations.

The famous early /8 holders

Class A blocks (each containing 16.7 million addresses) were assigned to a small number of early Internet players. Notable holders:

MIT — 18.0.0.0/8 (until returned in 2017)
IBM — 9.0.0.0/8
HP / Compaq — 15.0.0.0/8, 16.0.0.0/8
Xerox — 13.0.0.0/8
US DoD — many /8s
AT&T — 12.0.0.0/8
Apple — 17.0.0.0/8
Ford — 19.0.0.0/8

These allocations were made in the 1980s when /8 felt reasonable. Today they're worth tens of millions of dollars on the secondary market.

CIDR: the response

By 1992, address exhaustion was clearly coming. The CIDR proposal eliminated the class boundaries: any prefix length was allowed (now familiar as /24, /22, /15, etc.). Allocations could be sized to actual need rather than rounded up to the next class.

CIDR was deployed in 1993. The immediate effect: a university needing 500 addresses got a /23 (510 usable) instead of a Class B. The aggregate impact bought enough time for IPv6 development that the IPv4 free pool didn't fully exhaust until 2011.

Why "Class C" still gets used colloquially

Despite CIDR's 30+ year deployment, you'll hear network engineers casually say "give them a Class C" meaning "a /24." The terminology survived because /24 happens to be the size of the original Class C, and "Class C" is shorter to say. It's not technically correct after 1993 — there are no classes — but it's understood in context.

Strictly accurate terminology uses CIDR (e.g., "a /24") rather than class names. The class names are vestigial.

The reserved ranges that came from classful days

Some special-purpose blocks have classful-era origins:

127.0.0.0/8 — loopback. Originally a full Class A reserved for loopback (one address would have sufficed, but Class A was the smallest allocation).
10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16 — RFC 1918 private addressing. The 10/8 is a Class A; 172.16/12 is a contiguous block within the original Class B range; 192.168/16 is in the Class C range.
169.254.0.0/16 — link-local autoconfiguration (used when DHCP fails). Within the original Class B range.
224.0.0.0/4 — multicast. Originally the whole Class D.

The takeaway

Classful addressing was a 12-year era that ended with CIDR. Knowing about it explains why some terminology persists, why some legacy networks have unusual allocation patterns, and why companies like Apple still hold /8s. For day-to-day networking in 2026, CIDR notation and concepts are what's relevant — see our subnetting article.

Frequently asked questions

Are IP classes still used?: Only informally and as historical reference. Routing protocols, allocation systems, and operating systems all use CIDR notation. The class names persist in conversation but the technical reality is classless since 1993.
Why is the 10.x.x.x range called private?: RFC 1918 reserved it (along with 172.16/12 and 192.168/16) for private internal networks not reachable from the public Internet. The 10/8 was an early Class A reserved for this purpose; the size made it usable for very large internal networks.
Can someone still get a Class A?: Not from a Regional Internet Registry — the original free pool is exhausted. The only path is buying an existing /8 on the secondary market, which costs hundreds of millions of dollars. Realistically, no organization is acquiring new Class A equivalents.
Why doesn't 127.0.0.1 use just one address?: Historical accident. When 127/8 was reserved for loopback in 1981, the smallest reservable unit was a Class A. The 16 million addresses are effectively wasted on a function that needs only one (or a few). IPv6 corrected this with ::1/128 — a single address for loopback.
What changed when CIDR replaced classful?: Routing protocols had to be updated to carry prefix length (RIPv2, OSPF, BGP4). Allocation rules became finer-grained. Network engineers had to learn variable-length subnet masks. Internet routing tables became more efficient (CIDR aggregation reduced the number of routes). The transition took years; the technical and educational changes were substantial.