64:ff9b:1::/48: Local-Use IPv4/IPv6 Translation Prefix
A reserved IPv6 prefix, 64:ff9b:1::/48, that operators can use locally for any IPv4/IPv6 translation mechanism.
Anderson, T., "Local-Use IPv4/IPv6 Translation Prefix", RFC 8215, DOI 10.17487/RFC8215, August 2017, <https://www.rfc-editor.org/info/rfc8215>.
@misc{rfc8215,
series = {Request for Comments},
number = 8215,
howpublished = {RFC 8215},
publisher = {RFC Editor},
doi = {10.17487/RFC8215},
url = {https://www.rfc-editor.org/info/rfc8215},
author = {T. Anderson},
title = {{Local-Use IPv4/IPv6 Translation Prefix}},
pagetotal = 7,
year = 2017,
month = aug,
} RFC text
Contents
Internet Engineering Task Force (IETF) T. Anderson
Request for Comments: 8215 Redpill Linpro
Category: Standards Track August 2017
ISSN: 2070-1721
Local-Use IPv4/IPv6 Translation Prefix
Abstract
This document reserves the IPv6 prefix 64:ff9b:1::/48 for local use
within domains that enable IPv4/IPv6 translation mechanisms.
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Further information on
Internet Standards is available in Section 2 of RFC 7841.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
http://www.rfc-editor.org/info/rfc8215.
Copyright Notice
Copyright (c) 2017 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 2
4. Why 64:ff9b:1::/48? . . . . . . . . . . . . . . . . . . . . . 3
4.1. Prefix Length . . . . . . . . . . . . . . . . . . . . . . 3
4.2. Prefix Value . . . . . . . . . . . . . . . . . . . . . . 4
5. Deployment Considerations . . . . . . . . . . . . . . . . . . 4
6. Checksum Neutrality . . . . . . . . . . . . . . . . . . . . . 5
7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6
8. Security Considerations . . . . . . . . . . . . . . . . . . . 6
9. References . . . . . . . . . . . . . . . . . . . . . . . . . 6
9.1. Normative References . . . . . . . . . . . . . . . . . . 6
9.2. Informative References . . . . . . . . . . . . . . . . . 7
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 7
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 7
1. Introduction
This document reserves 64:ff9b:1::/48 for local use within domains
that enable IPv4/IPv6 translation mechanisms. This facilitates the
coexistence of multiple IPv4/IPv6 translation mechanisms in the same
network without requiring the use of a Network-Specific Prefix
assigned from the operator's allocated global unicast address space.
2. Terminology
This document uses the following terms:
Network-Specific Prefix (NSP)
A globally unique prefix assigned by a network operator for use
with an IPv4/IPv6 translation mechanism [RFC6052].
Well-Known Prefix (WKP)
The prefix 64:ff9b::/96, which is reserved for use with the
[RFC6052] IPv4/IPv6 address translation algorithms.
3. Problem Statement
Since the WKP 64:ff9b::/96 was reserved by [RFC6052], several new
IPv4/IPv6 translation mechanisms have been defined by the IETF, such
as those defined in [RFC6146] and [RFC7915]. These mechanisms target
various different use cases. An operator might therefore wish to
make use of several of them simultaneously.
The WKP is reserved specifically for use with the algorithms
specified in [RFC6052]. More recent RFCs describe IPv4/IPv6
translation mechanisms that use different algorithms. An operator
deploying such mechanisms cannot make use of the WKP in a legitimate
fashion.
Also, because the WKP is a /96, an operator preferring to use the WKP
over an NSP can do so for only one of their IPv4/IPv6 translation
mechanisms. All others must necessarily use an NSP.
Section 3.1 of [RFC6052] imposes certain restrictions on the use of
the WKP, such as forbidding its use in combination with private IPv4
addresses [RFC1918]. These restrictions might conflict with the
operator's desired use of an IPv4/IPv6 translation mechanism.
In summary, there is a need for a local-use prefix that facilitates
the coexistence of multiple IPv4/IPv6 translation mechanisms in a
single network domain, as well as the deployment of translation
mechanisms that do not use the [RFC6052] algorithms or adhere to its
usage restrictions.
4. Why 64:ff9b:1::/48?
4.1. Prefix Length
One of the primary goals of this document is to facilitate multiple
simultaneous deployments of IPv4/IPv6 translation mechanisms in a
single network. The first criterion is therefore that the prefix
length chosen must be shorter than the prefix length used by any
individual translation mechanism.
The second criterion is that the prefix length chosen is a multiple
of 16. This ensures the prefix ends on a colon boundary when
representing it in text, easing operator interaction with it.
The [RFC6052] algorithms specifies IPv4/IPv6 translation prefixes as
short as /32. In order to facilitate multiple instances of
translation mechanisms using /32s, while at the same time aligning on
a 16-bit boundary, it would be necessary to reserve a /16. Doing so,
however, was considered as too wasteful by the IPv6 Operations
Working Group.
The shortest translation prefix that was reported to the IPv6
Operations Working Group as being deployed in a live network was /64.
The longest 16-bit-aligned prefix length that can accommodate
multiple instances of /64 is /48. The prefix length of /48 was
therefore chosen, as it satisfies both the criteria above, while at
the same time avoids wasting too much of the IPv6 address space.
4.2. Prefix Value
It is desirable to minimise the amount of additional "pollution" in
the unallocated IPv6 address space caused by the reservation made by
this document. Ensuring the reserved prefix is adjacent to the
64:ff9b::/96 WKP already reserved by [RFC6052] accomplishes this.
Given the previous decision to use a prefix length of /48, this
leaves two options: 64:ff9a:ffff::/48 and 64:ff9b:1::/48.
64:ff9a:ffff::/48 has the benefit that it is completely adjacent to
the [RFC6052] WKP. That is, 64:ff9a:ffff::/48 and 64:ff9b::/96
combine to form an uninterrupted range of IPv6 addresses starting
with 64:ff9a:ffff:: and ending with 64:ff9b::ffff:ffff.
64:ff9b:1::/48 is, on the other hand, not completely adjacent to
64:ff9b::/96. The range starting with 64:ff9b::1:0:0 and ending with
64:ff9b:0:ffff:ffff:ffff:ffff:ffff would remain unallocated.
This particular drawback is, however, balanced by the fact that the
smallest possible aggregate prefix that covers both the [RFC6052] WKP
and 64:ff9a:ffff::/48 is much larger than the smallest possible
aggregate prefix that covers both the [RFC6052] WKP and
64:ff9b:1::/48. These aggregate prefixes are 64:ff9a::/31 and
64:ff9b::/47, respectively. IPv6 address space is allocated using
prefixes rather than address ranges, so it could be argued that
64:ff9b:1::/48 is the option that would cause special-use prefixes
reserved for IPv4/IPv6 translation to "pollute" the minimum possible
amount of unallocated IPv6 address space.
Finally, 64:ff9b:1::/48 also has the advantage that its textual
representation is shorter than 64:ff9a:ffff::/48. While this might
seem insignificant, the preference human network operators have for
addresses that are simple to type should not be underestimated.
After weighing the above pros and cons, 64:ff9b:1::/48 was chosen.
5. Deployment Considerations
64:ff9b:1::/48 is intended as a technology-agnostic and generic
reservation. A network operator may freely use it in combination
with any kind of IPv4/IPv6 translation mechanism deployed within
their network.
By default, IPv6 nodes and applications must not treat IPv6 addresses
within 64:ff9b:1::/48 differently from other globally scoped IPv6
addresses. In particular, they must not make any assumptions
regarding the syntax or properties of those addresses (e.g., the
existence and location of embedded IPv4 addresses) or the type of
associated translation mechanism (e.g., whether it is stateful or
stateless).
64:ff9b:1::/48 or any more-specific prefix may only be used in inter-
domain routing if done in accordance with the rules described in
Section 3.2 of [RFC6052].
Note that 64:ff9b:1::/48 (or any more-specific prefix) is distinct
from the WKP 64:ff9b::/96. Therefore, the restrictions on the use of
the WKP described in Section 3.1 of [RFC6052] do not apply to the use
of 64:ff9b:1::/48.
Operators tempted to use the covering aggregate prefix 64:ff9b::/47
to refer to all special-use prefixes currently reserved for IPv4/IPv6
translation should be warned that this aggregate includes a range of
unallocated addresses (see Section 4.2) that the IETF could
potentially reserve in the future for entirely different purposes.
6. Checksum Neutrality
Use of 64:ff9b:1::/48 does not in itself guarantee checksum
neutrality, as many of the IPv4/IPv6 translation algorithms it can be
used with are fundamentally incompatible with checksum-neutral
address translations.
Section 4.1 of [RFC6052] contains further discussion about IPv4/IPv6
translation and checksum neutrality.
The Stateless IP/ICMP Translation algorithm [RFC7915] is one well-
known algorithm that can operate in a checksum-neutral manner, when
using the [RFC6052] algorithms for all of its address translations.
However, in order to attain checksum neutrality, it is imperative
that the translation prefix be chosen carefully. Specifically, in
order for a 96-bit [RFC6052] prefix to be checksum neutral, all the
six 16-bit words in the prefix must add up to a multiple of 0xffff.
The following non-exhaustive list contains examples of translation
prefixes that are checksum neutral when used with the [RFC7915] and
[RFC6052] algorithms:
o 64:ff9b:1:fffe::/96
o 64:ff9b:1:fffd:1::/96
o 64:ff9b:1:fffc:2::/96
o 64:ff9b:1:abcd:0:5431::/96
7. IANA Considerations
The IANA has added the following entry to the "IANA IPv6 Special-
Purpose Address Registry":
+----------------------+---------------------+
| Attribute | Value |
+----------------------+---------------------+
| Address Block | 64:ff9b:1::/48 |
| Name | IPv4-IPv6 Translat. |
| RFC | RFC 8215 |
| Allocation Date | 2017-06 |
| Termination Date | N/A |
| Source | True |
| Destination | True |
| Forwardable | True |
| Globally Reachable | False |
| Reserved-by-Protocol | False |
+----------------------+---------------------+
The IANA has also added the following footnote to the 0000::/8 entry
of the "Internet Protocol Version 6 Address Space" registry:
64:ff9b:1::/48 reserved for Local-Use IPv4/IPv6 Translation
[RFC8215].
8. Security Considerations
The reservation of 64:ff9b:1::/48 is not known to cause any new
security considerations beyond those documented in Section 5 of
[RFC6052].
9. References
9.1. Normative References
[RFC6052] Bao, C., Huitema, C., Bagnulo, M., Boucadair, M., and X.
Li, "IPv6 Addressing of IPv4/IPv6 Translators", RFC 6052,
DOI 10.17487/RFC6052, October 2010,
<https://www.rfc-editor.org/info/rfc6052>.
9.2. Informative References
[RFC1918] Rekhter, Y., Moskowitz, B., Karrenberg, D., de Groot, G.,
and E. Lear, "Address Allocation for Private Internets",
BCP 5, RFC 1918, DOI 10.17487/RFC1918, February 1996,
<https://www.rfc-editor.org/info/rfc1918>.
[RFC6146] Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful
NAT64: Network Address and Protocol Translation from IPv6
Clients to IPv4 Servers", RFC 6146, DOI 10.17487/RFC6146,
April 2011, <https://www.rfc-editor.org/info/rfc6146>.
[RFC7915] Bao, C., Li, X., Baker, F., Anderson, T., and F. Gont,
"IP/ICMP Translation Algorithm", RFC 7915,
DOI 10.17487/RFC7915, June 2016,
<https://www.rfc-editor.org/info/rfc7915>.
Acknowledgements
The author would like to thank Fred Baker, Mohamed Boucadair,
Brian E. Carpenter, Pier Carlo Chiodi, Joe Clarke, David Farmer,
Suresh Krishnan, Warren Kumari, Holger Metschulat, Federico
Santandrea, and David Schinazi for contributing to the creation of
this document.
Author's Address
Tore Anderson
Redpill Linpro
Vitaminveien 1A
0485 Oslo
Norway
Phone: +47 959 31 212
Email: tore@redpill-linpro.com
URI: http://www.redpill-linpro.com
AI ELI5
On this page
§1What the RFC Reserves
RFC 8215 reserves the IPv6 prefix 64:ff9b:1::/48 for local use within domains that enable IPv4/IPv6 translation mechanisms. The aim is to let several translation mechanisms live in the same network without requiring a Network-Specific Prefix (NSP) taken from the operator's own allocated global unicast address space.
Two terms matter. An NSP is a globally unique prefix assigned by a network operator for use with a translation mechanism. The Well-Known Prefix (WKP) is 64:ff9b::/96, reserved for the address translation algorithms of RFC 6052.
The RFC puts aside a block of special addresses that a network can use for translating between old and new internet addresses, so nobody has to borrow addresses from the normal pool.
§2The Problem with the Well-Known Prefix
Since RFC 6052 reserved the WKP, the IETF has defined more translation mechanisms, such as those in RFC 6146 and RFC 7915. They serve different use cases, so an operator may want several at once.
The WKP does not fit well for three reasons:
- It is reserved specifically for the algorithms of RFC 6052. Newer mechanisms that use different algorithms cannot use it legitimately.
- It is a /96, so an operator who prefers it over an NSP can use it for only one translation mechanism. All others must use an NSP.
- Section 3.1 of RFC 6052 restricts its use, for example by forbidding its use with private IPv4 addresses (RFC 1918). Those limits may conflict with what the operator wants.
The RFC concludes that a local-use prefix is needed so that multiple mechanisms can coexist in one network domain, including mechanisms that do not use the RFC 6052 algorithms or follow its usage restrictions.
There was only one special set of addresses, and it came with strict rules and room for just one user. A network that wanted several translators had nowhere to put the rest.
§3Why 64:ff9b:1::/48
The prefix length had two criteria. First, it must be shorter than the prefix used by any individual translation mechanism, so several can fit inside it. Second, it must be a multiple of 16, so that it ends on a colon boundary in text and is easy for operators to handle.
RFC 6052 allows translation prefixes as short as /32, and fitting several of those on a 16-bit boundary would need a /16. The IPv6 Operations Working Group considered that too wasteful. The shortest translation prefix reported as deployed in a live network was /64, and the longest 16-bit-aligned prefix that can hold multiple /64s is /48. So /48 was chosen.
For the value, the goal was to keep the reservation next to 64:ff9b::/96 so as to disturb as little unallocated space as possible. That left two options, 64:ff9a:ffff::/48 and 64:ff9b:1::/48.
64:ff9a:ffff::/48is completely adjacent to the WKP, but the smallest aggregate covering both is64:ff9a::/31.64:ff9b:1::/48leaves a gap of unallocated addresses, but the smallest aggregate covering both is the much smaller64:ff9b::/47. Its text form is also shorter to type.
After weighing these, 64:ff9b:1::/48 was chosen.
The authors needed a block big enough to hold many translator teams but not so big that it wasted lots of addresses. They picked the size and the spot that were easiest to type and that sat right beside the existing special block.
§4How Operators Use the Prefix
The prefix is intended as a technology-agnostic, generic reservation. An operator may freely use it with any kind of IPv4/IPv6 translation mechanism deployed in their network.
By default, IPv6 nodes and applications must not treat addresses in 64:ff9b:1::/48 differently from other globally scoped IPv6 addresses. They must not assume anything about the address syntax or properties, such as whether an IPv4 address is embedded, or about the translation type, such as stateful or stateless.
The prefix or any more-specific prefix may be used in inter-domain routing only under the rules in Section 3.2 of RFC 6052. It is distinct from the WKP, so the restrictions on the WKP in Section 3.1 of RFC 6052 do not apply to it.
Operators should not use the covering aggregate 64:ff9b::/47 to mean all translation prefixes. It includes unallocated addresses that the IETF could reserve in future for entirely different purposes.
Network owners can use these addresses with any kind of translator. Everyone else should treat them like ordinary addresses and not try to guess what is inside them.
§5Checksum Neutrality
Using 64:ff9b:1::/48 does not by itself guarantee checksum neutrality, because many translation algorithms it can be used with are fundamentally incompatible with checksum-neutral address translation. RFC 6052 discusses this in Section 4.1.
The stateless translation algorithm in RFC 7915 can work in a checksum-neutral way when it uses the RFC 6052 algorithms for all its address translations. To get there, the translation prefix must be chosen carefully: for a 96-bit RFC 6052 prefix to be checksum neutral, its six 16-bit words must add up to a multiple of 0xffff.
The RFC lists some non-exhaustive examples that are checksum neutral with RFC 7915 and RFC 6052: 64:ff9b:1:fffe::/96, 64:ff9b:1:fffd:1::/96, 64:ff9b:1:fffc:2::/96 and 64:ff9b:1:abcd:0:5431::/96.
Packets carry a little check number. If a translator changes the addresses, that number can come out wrong, unless the new address is chosen so the sums still match.
§6IANA Registry Entries
IANA added an entry to the IANA IPv6 Special-Purpose Address Registry for 64:ff9b:1::/48, named "IPv4-IPv6 Translat.", with an allocation date of 2017-06 and no termination date. The entry says Source and Destination are True, Forwardable is True, Globally Reachable is False and Reserved-by-Protocol is False.
IANA also added a footnote to the 0000::/8 entry of the Internet Protocol Version 6 Address Space registry: "64:ff9b:1::/48 reserved for Local-Use IPv4/IPv6 Translation".
The official address list at the internet's record office got a new line saying these addresses are set aside for translators, and that they work inside a network but are not meant to be reached from everywhere.
§7Security Considerations
The reservation of 64:ff9b:1::/48 is not known to cause any new security considerations beyond those documented in Section 5 of RFC 6052.
Setting aside these addresses is not known to make anything less safe than before.