In this episode of PING, APNIC Chief Scientist Geoff Huston discusses the complicated history of the CIDR Report. Geoff has been running the CIDR Report continuously for over two decades, following on from the work of Tony Bates and Philip Smith.
Classless Inter-Domain Routing (CIDR) was introduced in the 1990s to replace the original fixed‑size network model defined in RFC 791. That earlier model divided IP address space into Class A, Class B, and Class C (there were also Classes D and E, but they’re not relevant here).
The key idea behind classless addressing was to remove the requirement that the top three bits of an IP address determine its class and size. Under the old system, each class came with a fixed number of addresses: 17 million for Class A, 65,000 for Class B, and just 256 for Class C. CIDR eliminated these fixed boundaries, allowing address blocks to be sized much more flexibly.
The older class-based mechanism worked fine for the early life of the Internet, but under the stresses of exponential growth in the 1990s, this new method for allocating addresses was defined, which exploited this classless model and allowed people to be given sizes between 17 million and 65,000 or between 65,000 and 256. This fixed two problems: Access to addresses into the future (through the Regional Internet Registry (RIR) model of justified need for addressing) and the scaling problems of the routing mechanism.
Routing has roots going back to the 1950s, with the development of methods for exchanging information about paths in a network, most notably the Bellman‑Ford algorithm. Variants of this approach continued to evolve alongside the growth of the Internet, eventually displacing earlier routing models used in other networks, such as DECnet from the Digital Equipment Corporation.
Today, this lineage lives on in the Border Gateway Protocol, version 4 (BGP4), an appropriate name for the fourth version of a protocol that itself evolved from BGP3. BGP4’s key innovation was support for CIDR, allowing routing to work with prefixes of variable length rather than fixed‑size address classes.
The CIDR Report grew out of a need to understand what was putting stress on BGP — a shared public system that collects everyone’s routing announcements. While it was often perfectly rational for an individual network to improve its own performance by announcing more routes, doing so imposed a real cost on every other BGP speaker.
The report made these impacts visible. It showed which Autonomous Systems (ASes) were the ‘noisy’ speakers, how much strain they were adding to the routing system, and how they could achieve the same engineering goals with far fewer announcements. In effect, it was an early form of ‘nudge’ theory: A public record that used transparency (using naming‑and‑shaming) to highlight the damage any one participant could do to the shared routing commons.
But carried into the modern era, after years of discussion about the declining importance of BGP routing in a world increasingly driven by name‑based content steering, how relevant is the CIDR Report today? In this episode of PING, Geoff and I discuss whether the CIDR Report still holds value.
Learn more about the CIDR Report:
- The CIDR Report (website, daily update)
- The why and what of the CIDR Report (APNIC Blog, 2026)
- We don’t need subnets any more (PING podcast, 2024)
- RFC 4632 Classless Inter-Domain Routing (IETF RFC)
- IPv6 Prefix Lengths (Geoff Huston, APNIC Blog, 2024)
Subscribe and share your story
You can stream and subscribe to PING via the following channels:
If you’re interested in sharing your insights or research, please get in touch — we’re always looking for great stories from the community. Please let us know what you think of the podcast and the APNIC Blog so we can keep improving.
The views expressed by the authors of this blog are their own and do not necessarily reflect the views of APNIC. Please note a Code of Conduct applies to this blog.