RFC 6619






Internet Engineering Task Force (IETF)                          J. Arkko
Request for Comments: 6619                                      Ericsson
Category: Standards Track                                      L. Eggert
ISSN: 2070-1721                                                   NetApp
                                                             M. Townsley
                                                                   Cisco
                                                               June 2012


Scalable Operation of Address Translators with Per-Interface Bindings

Abstract



   This document explains how to employ address translation in networks
   that serve a large number of individual customers without requiring a
   correspondingly large amount of private IPv4 address space.

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 5741.

   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/rfc6619.

Copyright Notice



   Copyright (c) 2012 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.






Arkko, et al.                Standards Track                    [Page 1]

RFC 6619                      Scalable NATs                    June 2012


1.  Introduction



   This document explains how to employ address translation without
   consuming a large amount of private address space.  This is important
   in networks that serve a large number of individual customers.
   Networks that serve more than 2^24 (16 million) users cannot assign a
   unique private IPv4 address to each user, because the largest
   reserved private address block reserved is 10/8 [RFC1918].  Many
   networks are already hitting these limits today -- for instance, in
   the consumer Internet service market.  Even some individual devices
   may approach these limits -- for instance, cellular network gateways
   or mobile IP home agents.

   If ample IPv4 address space were available, this would be a
   non-issue, because the current practice of assigning public IPv4
   addresses to each user would remain viable, and the complications
   associated with using the more limited private address space could be
   avoided.  However, as the IPv4 address pool is becoming depleted,
   this practice is becoming increasingly difficult to sustain.

   It has been suggested that more of the unassigned IPv4 space should
   be converted for private use, in order to allow the provisioning of
   larger networks with private IPv4 address space.  At the time of this
   writing, the IANA "free pool" contained only 12 unallocated unicast
   IPv4 /8 prefixes.  Although reserving a few of those for private use
   would create some breathing room for such deployments, it would not
   result in a solution with long-term viability.  It would result in
   significant operational and management overheads, and it would
   further reduce the number of available IPv4 addresses.

   Segmenting a network into areas of overlapping private address space
   is another possible technique, but it severely complicates the design
   and operation of a network.

   Finally, the transition to IPv6 will eventually eliminate these
   addressing limitations.  However, during the migration period when
   IPv4 and IPv6 have to coexist, address or protocol translation will
   be needed in order to reach IPv4 destinations.

   The rest of this document is organized as follows.  Section 2 gives
   an outline of the solution, Section 3 introduces some terms,
   Section 4 specifies the required behavior for managing NAT bindings,
   and Section 5 discusses the use of this technique with IPv6.








Arkko, et al.                Standards Track                    [Page 2]

RFC 6619                      Scalable NATs                    June 2012


2.  Solution Outline



   The need for address or protocol translation during the migration
   period to IPv6 creates the opportunity to deploy these mechanisms in
   a way that allows the support of a large user base without the need
   for a correspondingly large IPv4 address block.

   A Network Address Translator (NAT) is typically configured to connect
   a network domain that uses private IPv4 addresses to the public
   Internet.  The NAT device, which is configured with a public IPv4
   address, creates and maintains a mapping for each communication
   session from a device inside the domain it serves to devices in the
   public Internet.  It does that by translating the packet flow of each
   session such that the externally visible traffic uses only public
   addresses.

   In many NAT deployments, the network domain connected by the NAT to
   the public Internet is a broadcast network sharing the same media,
   where each individual device must have a private IPv4 address that is
   unique within this network.  In such deployments, it is natural also
   to implement the NAT functionality such that it uses the private IPv4
   address when looking up which mapping should be used to translate a
   given communication session.

   It is important to note, however, that this is not an inherent
   requirement.  When other methods of identifying the correct mapping
   are available, and the NAT is not connecting a shared-media broadcast
   network to the Internet, there is no need to assign each device in
   the domain a unique IPv4 address.

   This is the case, for example, when the NAT connects devices to the
   Internet that connect to it with individual point-to-point links.  In
   this case, it becomes possible to use the same private addresses many
   times, making it possible to support any number of devices behind a
   NAT using very few IPv4 addresses.

   There are tunneling-based techniques that can obtain the same
   benefits by establishing new tunnels over any IP network [RFC6333].
   However, where the point-to-point links already exist, creating an
   additional layer of tunneling is unnecessary (and even potentially
   harmful due to effects on the Maximum Transfer Unit (MTU) settings).
   The approach described in this document can be implemented and
   deployed within a single device and has no effect on hosts behind it.
   In addition, as no additional layers of tunneling are introduced,
   there is no effect on the MTU.  It is also unnecessary to implement
   tunnel endpoint discovery, security mechanisms, or other aspects of a
   tunneling solution.  In fact, there are no changes to the devices
   behind the NAT.



Arkko, et al.                Standards Track                    [Page 3]

RFC 6619                      Scalable NATs                    June 2012


   Note, however, that existing tunnels are a common special case of
   point-to-point links.  For instance, cellular network gateways
   terminate a large number of tunnels that are already needed for
   mobility management reasons.  Implementing the approach described in
   this document is particularly attractive in such environments, given
   that no additional tunneling mechanisms, negotiation, or host changes
   are required.  In addition, since there is no additional tunneling,
   packets continue to take the same path as they would normally take.
   Other commonly used network technologies that may be of interest
   include Point-to-Point Protocol (PPP) [RFC1661] links, PPP over
   Ethernet (PPPoE) [RFC2516] encapsulation, Asynchronous Transfer Mode
   (ATM) Permanent Virtual Circuits (PVCs), and per-subscriber virtual
   LAN (VLAN) allocation in consumer broadband networks.

   The approach described here also results in overlapping private
   address space, like the segmentation of the network to different
   areas.  However, this overlap is applied only at the network edges
   and does not impact routing or reachability of servers in a negative
   way.

3.  Terms



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in [RFC2119].

   "NAT" in this document includes both "Basic NAT" and "Network Address
   Port Translation (NAPT)" as defined by [RFC2663].  The term "NAT
   Session" is adapted from [RFC5382] and is defined as follows.

      NAT Session - A NAT session is an association between a transport
      layer session as seen in the internal realm and a session as seen
      in the external realm, by virtue of NAT translation.  The NAT
      session will provide the translation glue between the two session
      representations.

   This document uses the term "mapping" as defined in [RFC4787] to
   refer to state at the NAT necessary for network address and port
   translation of sessions.

4.  Per-Interface Bindings



   To support a mode of operation that uses a fixed number of IPv4
   addresses to serve an arbitrary number of devices, a NAT MUST manage
   its mappings on a per-interface basis, by associating a particular
   NAT session not only with the five tuples used for the transport
   connection on both sides of the NAT but also with the internal
   interface on which the user device is connected to the NAT.  This



Arkko, et al.                Standards Track                    [Page 4]

RFC 6619                      Scalable NATs                    June 2012


   approach allows each internal interface to use the same private IPv4
   address range.  Note that the interface need not be physical; it may
   also correspond to a tunnel, VLAN, or other identifiable
   communications channel.

   For deployments where exactly one user device is connected with a
   separate tunnel interface and all tunnels use the same IPv4 address
   for the user devices, it is redundant to store this address in the
   mapping in addition to the internal interface identifier.  When the
   internal interface identifier is shorter than a 32-bit IPv4 address,
   this may decrease the storage requirements of a mapping entry by a
   small measure, which may aid NAT scalability.  For other deployments,
   it is likely necessary to store both the user device IPv4 address and
   the internal interface identifier, which slightly increases the size
   of the mapping entry.

   This mode of operation is only suitable in deployments where user
   devices connect to the NAT over point-to-point links.  If supported,
   this mode of operation SHOULD be configurable, and it should be
   disabled by default in general-purpose NAT devices.

   All address translators make it hard to address devices behind them.
   The same is true of the particular NAT variant described in this
   document.  An additional constraint is caused by the use of the same
   address space for different devices behind the NAT, which prevents
   the use of unique private addresses for communication between devices
   behind the same NAT.

5.  IPv6 Considerations



   Private address space conservation is important even during the
   migration to IPv6, because it will be necessary to communicate with
   the IPv4 Internet for a long time.  This document specifies two
   recommended deployment models for IPv6.  In the first deployment
   model, the mechanisms specified in this document are useful.  In the
   second deployment model, no additional mechanisms are needed, because
   IPv6 addresses are already sufficient to distinguish mappings from
   each other.

   The first deployment model employs dual stack [RFC4213].  The IPv6
   side of dual stack operates based on global addresses and direct
   end-to-end communication.  However, on the IPv4 side, private
   addressing and NATs are a necessity.  The use of per-interface NAT
   mappings is RECOMMENDED for the IPv4 side under these circumstances.
   Per-interface mappings help the NAT scale, while dual-stack operation
   helps reduce the pressure on the NAT device by moving key types of
   traffic to IPv6, eliminating the need for NAT processing.




Arkko, et al.                Standards Track                    [Page 5]

RFC 6619                      Scalable NATs                    June 2012


   The second deployment model involves the use of address and protocol
   translation, such as the one defined in [RFC6146].  In this
   deployment model, there is no IPv4 in the internal network at all.
   This model is applicable only in situations where all relevant
   devices and applications are IPv6 capable.  In this situation,
   per-interface mappings could be employed as specified above, but they
   are generally unnecessary, as the IPv6 address space is large enough
   to provide a sufficient number of mappings.

6.  Security Considerations



   The practices outlined in this document do not affect the security
   properties of address translation.  The binding method specified in
   this document is not observable to a device that is on the outside of
   the NAT; i.e., a regular NAT and a NAT specified here cannot be
   distinguished.  However, the use of point-to-point links implies
   naturally that the devices behind the NAT cannot communicate with
   each other directly without going through the NAT (or a router).  The
   use of the same address space for different devices implies in
   addition that a NAT operation must occur between two devices in order
   for them to communicate.

   The security implications of address translation in general have been
   discussed in many previous documents, including [RFC2663], [RFC2993],
   [RFC4787], and [RFC5382].

7.  References



7.1.  Normative References



   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

7.2.  Informative References



   [L2NAT]    Miles, D., Ed., and M. Townsley, "Layer2-Aware NAT", Work
              in Progress, March 2009.

   [RFC1661]  Simpson, W., Ed., "The Point-to-Point Protocol (PPP)",
              STD 51, RFC 1661, July 1994.

   [RFC1918]  Rekhter, Y., Moskowitz, R., Karrenberg, D., de Groot, G.,
              and E. Lear, "Address Allocation for Private Internets",
              BCP 5, RFC 1918, February 1996.

   [RFC2516]  Mamakos, L., Lidl, K., Evarts, J., Carrel, D., Simone, D.,
              and R. Wheeler, "A Method for Transmitting PPP Over
              Ethernet (PPPoE)", RFC 2516, February 1999.



Arkko, et al.                Standards Track                    [Page 6]

RFC 6619                      Scalable NATs                    June 2012


   [RFC2663]  Srisuresh, P. and M. Holdrege, "IP Network Address
              Translator (NAT) Terminology and Considerations",
              RFC 2663, August 1999.

   [RFC2993]  Hain, T., "Architectural Implications of NAT", RFC 2993,
              November 2000.

   [RFC4213]  Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms
              for IPv6 Hosts and Routers", RFC 4213, October 2005.

   [RFC4787]  Audet, F., Ed., and C. Jennings, "Network Address
              Translation (NAT) Behavioral Requirements for Unicast
              UDP", BCP 127, RFC 4787, January 2007.

   [RFC5382]  Guha, S., Ed., Biswas, K., Ford, B., Sivakumar, S., and P.
              Srisuresh, "NAT Behavioral Requirements for TCP", BCP 142,
              RFC 5382, October 2008.

   [RFC6127]  Arkko, J. and M. Townsley, "IPv4 Run-Out and IPv4-IPv6
              Co-Existence Scenarios", RFC 6127, May 2011.

   [RFC6146]  Bagnulo, M., Matthews, P., and I. van Beijnum, "Stateful
              NAT64: Network Address and Protocol Translation from IPv6
              Clients to IPv4 Servers", RFC 6146, April 2011.

   [RFC6333]  Durand, A., Droms, R., Woodyatt, J., and Y. Lee, "Dual-
              Stack Lite Broadband Deployments Following IPv4
              Exhaustion", RFC 6333, August 2011.

   [TRILOGY]  "Trilogy Project", <http://www.trilogy-project.org/>.





















Arkko, et al.                Standards Track                    [Page 7]

RFC 6619                      Scalable NATs                    June 2012


Appendix A.  Contributors



   The ideas in this document were first presented in [RFC6333].  This
   document is also indebted to [RFC6127] and [L2NAT].  However, all of
   these documents focused on additional components, such as tunneling
   protocols or the allocation of special IP address ranges.  We wanted
   to publish a specification that just focuses on the core
   functionality of per-interface NAT mappings.  However, David Miles
   and Alain Durand should be credited with coming up with the ideas
   discussed in this memo.

Appendix B.  Acknowledgments



   The authors would also like to thank Randy Bush, Fredrik Garneij, Dan
   Wing, Christian Vogt, Marcelo Braun, Joel Halpern, Wassim Haddad,
   Alan Kavanaugh, and others for interesting discussions in this
   problem space.

   Lars Eggert is partly funded by the Trilogy Project [TRILOGY], a
   research project supported by the European Commission under its
   Seventh Framework Program.






























Arkko, et al.                Standards Track                    [Page 8]

RFC 6619                      Scalable NATs                    June 2012


Authors' Addresses



   Jari Arkko
   Ericsson
   Jorvas  02420
   Finland

   EMail: jari.arkko@piuha.net


   Lars Eggert
   NetApp
   Sonnenallee 1
   85551 Kirchheim
   Germany

   Phone: +49 151 12055791
   EMail: lars@netapp.com
   URI:   http://eggert.org/


   Mark Townsley
   Cisco
   Paris  75006
   France

   EMail: townsley@cisco.com
























Arkko, et al.                Standards Track                    [Page 9]