RFC 2073

An IPv6 Provider-Based Unicast Address Format

📅 published January 1997, from ietf

RFC 2073: An IPv6 Provider-Based Unicast Address Format 🎀

Network Working Group: Y. Rekhter (cisco), P. Lothberg (STUPI.AB), R. Hinden (Ipsilon Networks), S. Deering (Xerox PARC), J. Postel (ISI). All five are listed as Editors.

Request for Comments: 2073

Category: Standards Track ✨

Date: January 1997 📅

Obsoletes / Updates: The title block doesn't list anything this RFC obsoletes or updates. 🦋

Part 1 of 2. More is coming, so keep reading! 💖

Status of this Memo ✨

Okay so this doc is a standards track protocol spec for the Internet community. It's inviting discussion and suggestions for improvements, bestie. 🫶 If you want to know where this protocol stands in standardization, check the current "Internet Official Protocol Standards" (STD 1) list. Distribution of this memo is unlimited 🌈.

1.0 Introduction 🌸

So this doc defines an IPv6 provider-based unicast address format for use across the Internet. 💅 The address format here is consistent with the "IPv6 Addressing Architecture" [ARCH] and the "An Architecture for IPv6 Unicast Address Allocation" [ALLOC], and it's built to make scalable Internet routing easier. It's giving clean design, honestly. 😭

It also doesn't rule out other unicast address formats. Other formats are totally allowed to exist 🦋.

2.0 Overview of the IPv6 Address 🧚

Quick refresher: IPv6 addresses are 128-bit identifiers for interfaces and sets of interfaces. There are three address types: Unicast, Anycast, and Multicast. This doc only defines one specific kind of Unicast address 💜.

Throughout the doc, address fields get names, like "subscriber". If you add the word "ID" after the name (like "subscriber ID"), you're talking about the contents of that field. If you add the word "prefix" (like "subscriber prefix"), you're talking about the whole address up to and including that field. Main character energy for naming, fr fr 👑.

An IPv6 address type is shown by its leading bits. That variable-length run of leading bits is called the Format Prefix (FP) ✨.

This doc defines an address format for the 010 (binary) Format Prefix, which is used for Provider-Based Unicast addresses. The same address format could be used with other Format Prefixes too, as long as those prefixes also identify IPv6 unicast addresses. But only the "010" Format Prefix is defined in this document 🍓.

3.0 IPv6 Provider-Based Unicast Address Format 💖

This doc defines an address format for IPv6 provider-based unicast address assignment. It's expected to get used a lot for IPv6 nodes connected to the Internet, so it's kind of a big deal 🔥.

The format follows the "Architecture for IPv6 Unicast Address Allocation" [ALLOC]. It's specifically designed to support aggregation of network layer reachability information at multiple levels of the routing hierarchy 🌈.

For addresses in this format, address administration is organized into a three-level hierarchy: registry, provider, and subscriber. The format lets address allocation be flexible at each level of that hierarchy, so it can cover a wide range of allocation demands 🧚.

This doc assumes the Internet routing system doesn't make assumptions about the specific structure and semantics of an IPv6 address. The only exceptions are the structure and semantics of the Format Prefix part, and the use of the "longest prefix match" algorithm (on arbitrary bit boundaries) when making a forwarding decision 💅.

The whole point is scalable Internet-wide routing that doesn't put any constraints on connectivity among providers, or between providers and subscribers 🦋.

3.1 Provider-Based Unicast Address Structure 🎀

For address allocation purposes, this format has these parts: Format Prefix, Registry ID, Provider ID, Subscriber ID, and an Intra-Subscriber part. The Intra-Subscriber part is the subscriber's responsibility and isn't defined in this doc. Here's the layout:

  | 3 |  5 bits  |   n bits   |   56-n bits  |        64 bits     |
  +---+----------+------------+--------------+--------------------+
  |010|RegistryID| ProviderID | SubscriberID |  Intra-Subscriber  |
  +---+----------+------------+--------------+--------------------+

The next sections go through each part of the IPv6 Provider-Based Unicast address format. Other allocation strategies are possible within this framework, but the ones described here are what gets used to assign IPv6 provider-based addresses 💜.

3.2 Registry ID 🌸

The Internet has grown a lot and gotten way more global, so people have put a ton of thought into how Network Layer address space gets allocated and assigned. RFC 1466, "Guidelines for Management of IP Address Space", proposes a plan for distributed allocation and assignment of the IPv4 address space 🧠.

As the Internet moves to IPv6, the distributed allocation and assignment plan for IPv4 from RFC1466 becomes the base for distributed allocation and assignment of the IPv6 address space 🦋.

The Internet Assigned Number Authority (IANA) is the main registry for the IPv6 address space. IANA may allocate blocks of IPv6 addresses and delegate assignment of those blocks to qualified Regional Registries. If no delegated registration authority has been identified, IANA is the default registry ✨.

The Registry ID in this IPv6 provider-based unicast format is meant to support broad geographic address allocation and make it easier for distributed Regional Registries to operate 🌈.

The Registry ID comes right after the Format Prefix part of an IPv6 address 💅.

At the moment there are three Regional Registries: INTERNIC, RIPE NCC, and APNIC. IANA can also allocate addresses directly. To match this split, the document defines the following Registry IDs 🍓:

     Regional Registry                     Value (binary)
     --------------------                  --------------

     Multi-Regional (IANA)                 10000
     RIPE NCC                              01000
     INTERNIC                              11000
     APNIC                                 00100

All other Registry ID values are reserved by IANA. 🔒

The Multi-Regional Registry ID allows flexible address assignments that fall outside the geographic regions already allocated. IANA is responsible for managing address space registration under the Multi-Regional Registry ID 🌍.

IANA and any designated Regional Registries are expected to allocate addresses following this overall scheme. Where qualifying Regional Registries have been set up, primary responsibility for allocating from within that block is delegated to that registry 🫶.

A Regional Registry can have more than one block of addresses allocated to it. As a result, the Registry would have multiple Registry IDs associated with it 💖.

3.3 Provider ID and Subscriber ID 🎀

This doc leaves the organization of the Provider ID and Subscriber ID parts of the address up to each individual registry. The registry especially needs to decide how much address space goes to providers and their subscribers. A few issues come up when doing that 🧚:

There will likely be a mix of providers of different sizes.
Small providers will grow into large providers.
Large providers will lose customers and become small providers. In extreme cases, the registry will require them to return some of their address space to the registry.
Organizations that need to be multi-homed to more than one provider will request a Provider ID assignment.

It's important for a registry to design its Provider ID space so it's flexible and also uses the address space efficiently 💅.

3.3.1 Provider ID 🦋

The Provider ID for an address block that a registry allocates to a particular provider uniquely identifies that provider within the registry ✨.

This doc assumes some subscribers might decide to get their address space directly from a registry. That makes their addresses independent of the provider(s) they're directly attached to 🌸.

3.3.2 Subscriber ID 💜

Each provider specifies the structure and assignment strategy for its Subscriber IDs 🍓.

A (direct) provider might decide to group its subscribers into regions. This can help when the (direct) provider is attached to another (indirect) provider at multiple points, because it lets the direct provider exert some control over the coupling between the attachment points and the flow of traffic going to a particular subscriber (see Section 5.3.1 of [ALLOC]) 🌈.

To support that grouping, the (direct) provider may allocate some small number of high-order bits of the Subscriber ID as a Subscriber-Region ID. A Subscriber-Region ID identifies a group of subscribers that are topologically close to each other (from the provider's point of view). Those subscribers could then be reachable through a particular attachment point between the (direct) provider and other (indirect) provider(s) 🧚.

3.4 Intra-Subscriber Part 🎀

This doc leaves the organization of the Intra-Subscriber portion of the address up to individual subscribers 💅.

The provider-based unicast address format in this doc leaves 64 bits for the local portion of the address. The editors recommend that subscribers use IPv6 auto-configuration capabilities [AUTO] to generate addresses, using link-specific addresses as the Interface ID, such as 48-bit IEEE-802 MAC addresses. In that case, 16 bits are left for the Subnet ID. That should be plenty (for example, 65,535 subnets) for everyone except the very largest subscribers 🫶.

Here's how that looks:

  |            64 bits             |  16 bits  |     48 bits      |
  +--------------------------------+-----------+------------------+
  |       Subscriber Prefix        | Subnet ID |   Interface ID   |
  +--------------------------------+-----------+------------------+

Subscribers who need more subnets (and who want to keep using 48-bit IEEE-802 MAC addresses for their Interface IDs) can get them if the provider assigns them a block of subscriber prefixes. Or, an extremely large subscriber could be given its own Provider ID, which gives it extra address bits to build its own local address hierarchy 👑.

4.0 National Registries 🌍

okay bestie, so a Regional Registry 🗺️ can hand out chunks of address space to several National Registries. 💅 Then each National Registry is the one that gives addresses to individual providers inside its own country. it's giving delegation, and it's so clean 😭✨

to actually make National Registries happen, the Regional Registry can add one extra layer of hierarchy inside the Provider ID field. 🧚 the resulting Provider Prefix looks like this:

   | 3 |  5 bits  |  n bits  |  m bits  |   56-n-m   |    64 bits     |
   +---+----------+----------+----------+------------+----------------+
   |010|RegistryID| National | Provider | Subscriber |Intra-Subscriber|
   |   |          |RegistryID|   ID     |     ID     |                |
   +---+----------+----------+----------+------------+----------------+

the vibe here is that this document assumes each regional registry will only have a relatively small number of national registries 🌸 The size of the National-Registry ID should be related to two things: the number of countries in the region the regional registry administers, and the number of providers expected to be in each country. 🎀 that's the math bestie, no cap 💜

5.0 Acknowledgments 🫶

the editors want to thank Jim Bound (Digital), Scott Bradner (Harvard), Brian Carpenter (CERN), Geoff Huston (AANET), and Tony Li (cisco) for reviewing the draft and leaving constructive comments. 🦋 they really made it better, iconic 👑

6.0 References 📚

[ALLOC] Rekhter, Y., Li, T., "An Architecture for IPv6 Unicast Address Allocation", RFC 1887, December 1995.
[ARCH] Hinden, R., "IP Version 6 Addressing Architecture", RFC 1884, December 1995.
[AUTO] Thompson, S., "IPv6 Stateless Address Autoconfiguration", RFC 1972, August 1996.

7.0 Security Considerations 🔐

security issues are not discussed in this memo. 💀 (I know, I know, it's a lot to hold in, but that's the situation 🙈)

8.0 Editors' Addresses 💌

Yakov Rekhter 🌷 Cisco Systems, Inc. 170 West Tasman Drive San Jose, CA 95134-1706 USA Phone: +1 914 528-0090 EMail: yakov@cisco.com
Peter Lothberg 🌷 STUPI.AB Box 9129 S-102 72 Stockholm Sweden Phone:+46 8 6699720 EMail: roll@Stupi.SE
Robert M. Hinden 🌷 Ipsilon Networks, Inc. 2191 E. Bayshore Road Palo Alto, CA 94303 USA Phone: +1 415 846 4604 EMail: hinden@ipsilon.com
Stephen E. Deering 🌷 Xerox Palo Alto Research Center 3333 Coyote Hill Road Palo Alto, CA 94304 USA Phone: +1 415 812 4839 Fax: +1 415 812 4471 EMail: deering@parc.xerox.com
Jon Postel 🌷 Information Sciences Institute 4676 Admiralty Way Marina del Rey, CA 90292-6695 USA Phone: +1 310 822 1511 Fax: +1 310 823 6714 EMail: postel@isi.edu

📦 where this page comes from

ietf: rfc2073 1997-01
💖 girlified by claude-haiku-5-5 on 2026-10-15: a language model's rewrite of the original, under its terms