Classless Inter-Domain Routing

An Internet addressing and routing architecture that represents contiguous IP address blocks by arbitrary-length prefixes, enabling hierarchical allocation, longest-prefix forwarding, and route aggregation beyond fixed address classes.

Version

v2 · 2026-09-06 · History

Domain-specific #

1478

Origin domain

computing

Subdomain

Internet addressing and inter-domain routing

Aliases

CIDR, Classless routing, Classless addressing

Core Idea

Classless Inter-Domain Routing replaces fixed IPv4 network classes with an address prefix plus explicit prefix length, conventionally written address/length. A /p prefix fixes the first p bits and identifies a contiguous block of 2^(32-p) IPv4 addresses before reservations and operational subdivision. Routing advertisements carry such prefixes, and forwarding chooses the most specific matching route.^[1]

Arbitrary prefix lengths support hierarchical allocation and aggregation: a provider can advertise one shorter covering prefix rather than every customer subprefix, reducing routing state. More-specific routes remain possible for traffic engineering, multihoming, or exceptions, but deaggregation increases global table size. CIDR does not by itself conserve addresses, authorize ownership, secure route origin, or make an arbitrary set of networks aggregatable; binary contiguity and aligned boundaries matter.

Structural Signature

• The fixed-length address space. IPv4 or an analogous binary namespace supplies addresses.
• The prefix. A leading bit string names a contiguous block.
• The prefix length. An integer states how many leading bits are fixed.
• The containment hierarchy. Longer prefixes nest inside shorter covering prefixes.
• The allocation plan. Blocks are delegated along administrative/network topology.
• The route advertisement. Reachability is announced for a prefix.
• The longest-prefix rule. Forwarding selects the most specific matching advertisement.
• The aggregation condition. Aligned contiguous subprefixes with common routing policy can be summarized.
• The exception/deaggregation path. More-specific routes override aggregates with scaling cost.

What It Is Not

• Not subnetting alone. CIDR also changed global allocation and inter-domain aggregation.
• Not an address-ownership system. Prefix delegation and registry policy are separate.
• Not route security. Origin validation and path security require additional mechanisms.
• Not guaranteed aggregation. Topology and policy can force more specifics.
• Not only IPv4 notation. IPv6 also uses prefix-length addressing, though CIDR's historical crisis was IPv4.
• Not dotted-decimal arithmetic without binary alignment. Valid block boundaries follow fixed prefix bits.

Scope of Application

CIDR is literal in IP address planning, routing policy, forwarding, registry allocation, and Internet-scale route aggregation.

• Address allocation. Assigning variable-sized contiguous blocks.
• Subnet design. Nesting operational networks under a covering prefix.
• BGP routing. Advertising aggregated or more-specific reachability.
• Forwarding tables. Applying longest-prefix match.
• Route summarization. Compressing aligned subprefixes with compatible policy.
• Capacity planning. Estimating address counts and growth boundaries.
• Security operations. Binding route-origin authorizations to exact prefixes as a separate layer.

Clarity

State address family, canonical network address, prefix length, mask, inclusive range, usable-address convention, containing/contained prefixes, routing origin, aggregation assumptions, and more-specific exceptions. Perform calculations in binary or with verified tools. Distinguish a configured subnet, an allocated block, and an advertised route.

State address family, prefix, prefix length, allocation context, and whether the operation is assignment, aggregation, advertisement, filtering, or forwarding. A prefix denotes all addresses sharing its leading bits; host usability can be reduced by protocol or operational reservations without changing the mathematical block. Longest-prefix match is a forwarding rule, while aggregation is an advertisement strategy and variable-length subnetting is an internal allocation practice. CIDR introduced classless prefixes across addressing and inter-domain routing, but the terms are not interchangeable. A shorter aggregate can cover address space not actually reachable through one component unless advertisements and policy are constructed carefully. Route selection also depends on routing protocol attributes before forwarding uses the selected prefix. IPv6 is classless but has different address length and historical context from the IPv4 class-replacement plan.

Manages Complexity

Prefix hierarchy compresses many addresses and routes into one representation and lets the same containment rule drive allocation and forwarding. Longest-prefix match localizes exceptions. Aggregation loses value when address assignment ignores topology, policies differ, multihoming proliferates, or operators announce unnecessary specifics.

Classful addressing coupled network size to a few fixed boundaries and caused both allocation waste and rapid routing-table growth. CIDR replaces those coarse classes with arbitrary-length hierarchical blocks. Allocation can match demand more closely, adjacent blocks can be summarized by a common shorter prefix, and routers can forward by choosing the most specific applicable route. The same flexibility creates obligations: aggregation must respect bitwise contiguity, more-specific exceptions must remain visible where reachability differs, and deaggregation can re-expand the global table. The hierarchy compresses many routes only when topology, allocation, and policy align. CIDR manages complexity by making address containment explicit in the prefix itself, but it does not solve routing policy, multihoming, renumbering, or table growth automatically.

Abstract Reasoning

1. Choose the address family and required block size.
2. Derive a valid prefix length and aligned boundary.
3. Represent the block canonically.
4. Place it within a delegated covering hierarchy.
5. Allocate nonoverlapping subprefixes.
6. Advertise the least-specific route consistent with policy and reachability.
7. Apply longest-prefix match for exceptions.
8. Monitor table growth, leaks, and authorization separately.
9. Prefix reasoning should be performed bitwise rather than by visual similarity of decimal addresses. A block with prefix length p fixes its first p bits and leaves the rest variable; containment follows by comparing the fixed leading bits and lengths. Two equal-length blocks are either identical or disjoint, while blocks of unequal length can nest only when the shorter prefix matches the longer prefix's leading bits. Aggregation asks for a shorter common prefix whose represented block is exactly supported by reachability or is safely covered by more-specific exceptions. Forwarding asks a different question: among installed matching routes, which has the greatest prefix length? Separating those questions prevents a valid mathematical supernet from being mistaken for a valid operational advertisement. It also shows why address allocation, route policy, and packet lookup can use one hierarchy without being one process.

Knowledge Transfer

CIDR is a canonical hierarchical address: an address carries a nested prefix path, and specificity selects the closest applicable rule. Hierarchical Address is the strict parent; binary IP blocks, delegation, routing aggregation, and longest-prefix forwarding supply the accent.

Hierarchical Address is the strict parent because a prefix locates an address within nested blocks whose lengths express containment and delegation. The transferable pattern is variable-depth shared prefix → nested address region → inheritance with more-specific override. It applies to tries and naming systems when prefix containment is literal. The network residual includes IP bit strings, address allocation, route advertisements, longest-prefix forwarding, and aggregation across administrative domains. A subnet mask encodes a boundary, but CIDR is the architecture that removed fixed classes and integrated arbitrary prefixes into allocation and routing.

Examples

Canonical

The IPv4 prefix 192.0.2.0/24 fixes 24 leading bits and contains addresses from 192.0.2.0 through 192.0.2.255. Two aligned adjacent /25 blocks can be represented by that /24 when their routing policy and reachability admit aggregation.^[1]

Mapped back: fixed leading bits + explicit depth → contiguous address block → hierarchical allocation and summary route.

Applied / In Practice

A provider announces a covering /20, while one customer prefix is announced as a /24 through another path. Longest-prefix forwarding selects the /24 for matching addresses and the /20 otherwise; the exception consumes extra routing state.

An operator receives one contiguous prefix and divides it into unequal internal subprefixes matching different needs. At an external boundary it advertises the covering aggregate because all included destinations share reachability. One subprefix later moves to another path, so a more-specific advertisement overrides the aggregate for that block while the remainder stays summarized. Routers first select routes under protocol policy and then forward packets by the longest matching installed prefix. If the allocated blocks were noncontiguous, a single aggregate would also claim unrelated space and would be invalid without additional handling. The example separates address arithmetic, advertisement, and forwarding.

Mapped back: aggregate + more-specific exception → specificity-based forwarding → scaling tradeoff.

Structural Tensions

• Aggregation vs. policy independence. Summaries scale but can hide divergent routing. Diagnostic: Do all covered blocks share reachability and policy?
• Address efficiency vs. hierarchy. Tight allocation can fragment future aggregation. Diagnostic: Does the plan preserve aligned growth space?
• Specificity vs. table growth. More-specific routes add control and state. Diagnostic: Is each exception operationally necessary?
• Reachability vs. authorization. A route can be syntactically valid but unauthorized. Diagnostic: What origin/path validation is applied?
• Autonomous architecture vs. generic hierarchy. Many identifiers nest; bit-prefix allocation plus longest match defines CIDR. Diagnostic: Are prefix length and IP routing load-bearing?

Structural–Framed Character

CIDR is structural. Binary containment and matching are formal; allocations and advertisements are administratively framed. It is evaluatively neutral but affects Internet scaling. Hierarchical Address supplies nesting and specificity; IP routing supplies the concrete namespace and action.

Bit prefix, explicit length, contiguous block, hierarchical delegation, aggregation, more-specific exception, and longest-prefix forwarding are structural. Address family, registry policy, routing protocol, operator, topology, and particular block are framed. A slash notation is only a representation; the semantics lie in leading-bit agreement. Aggregation quality depends on framed topology and ownership, yet arbitrary prefix length remains invariant. This split prevents every slash-form address from being credited with successful inter-domain aggregation.

Structural Core vs. Domain Accent

The skeleton is prefix path + explicit depth → nested address set + specificity rule. The accent is IPv4/IPv6 bits, route advertisements, allocation registries, aggregation, and longest-prefix match. Remove those and one has hierarchical addressing generally.

The portable core is encode nested scope through variable-length leading symbols and let more-specific scopes override broader ones. The Internet accent is fixed-length binary addresses, globally delegated blocks, routing advertisements, and forwarding lookup. Remove this accent and the concept becomes Hierarchical Address. Keep subnetting but retain class-based inter-domain advertisements and the historical CIDR architecture is absent. The autonomous residual joins allocation efficiency and routing-table aggregation through one classless prefix representation.

Instantiates / Related Primes

Hierarchical Address is the strict parent because a CIDR prefix locates a block at one depth in a nested namespace and supports inheritance plus more-specific override.

The prospective workspace queue contains one strict upward edge to prime:hierarchical_address. No live DAG mutation is authorized.

Relationships to Other Abstractions

Current abstraction Classless Inter-Domain Routing Domain-specific

Parents (1) — more general patterns this builds on

• Classless Inter-Domain Routing is a kind of Hierarchical Address Prime

  Hierarchical Address is the strict parent because a CIDR prefix locates a block at one depth in a nested namespace and supports inheritance plus more-specific override.

Hierarchy paths (4) — routes to 4 parentless roots

• Classless Inter-Domain Routing → Hierarchical Address → Hierarchy → Network → Reservoir-Flux Network → Conservation Laws → Invariance

Show alternative paths (3)

Neighborhood in Abstraction Space

Classless Inter-Domain Routing sits in a sparse region of the domain-specific corpus (95^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (1565 abstractions)

Nearest neighbors

• Trie — 0.77
• Branch Table — 0.77
• Administrative Distance — 0.77
• Address space — 0.77
• Flat memory model — 0.77

Computed from structural-signature embeddings · 2026-09-08

Not to Be Confused With

• Subnet mask. An alternate representation of a prefix length.
• Variable-length subnet masking. Use of different subnet sizes, closely related but often intradomain.
• Route aggregation. One benefit/operation enabled by aligned prefixes.
• Longest prefix match. The forwarding selection rule.
• Classful networking. The fixed A/B/C boundary system CIDR replaced.
• NAT. Address translation, not prefix allocation or route aggregation.

References

[1] Vince Fuller and Tony Li, ‘Classless Inter-domain Routing (CIDR): The Internet Address Assignment and Aggregation Plan,’ RFC 4632 (Internet Engineering Task Force, August 2006), https://doi.org/10.17487/RFC4632. registry ↩a ↩b