Network Working Group
Request for Comments: 2766
Category: Standards Track

G. Tsirtsis
BT
P. Srisuresh
Campio Communications
February 2000

Network Address Translation - Protocol Translation (NAT-PT)

Status of this Memo

This document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited.

Copyright Notice

Abstract

This document specifies an IPv4-to-IPv6 transition mechanism, in addition to those already specified in [TRANS]. This solution attempts to provide transparent routing, as defined in [NAT-TERM], to end-nodes in V6 realm trying to communicate with end-nodes in V4 realm and vice versa. This is achieved using a combination of Network Address Translation and Protocol Translation. The scheme described does not mandate dual-stacks (i.e., IPv4 as well as V6 protocol support) or special purpose routing requirements (such as requiring tunneling support) on end nodes. This scheme is based on a combination of address translation theme as described in [NAT-TERM] and V6/V4 protocol translation theme as described in [SIIT].

Acknowledgements

Special thanks to Pedro Marques for reviewing an earlier version of this memo. Also, many thanks to Alan O'Neill and Martin Tatham, as the mechanism described in this document was initially developed through discussions with them.

   1. Introduction..................................................  2
   2. Terminology...................................................  3
      2.1 Network Address Translation (NAT).........................  4
      2.2 NAT-PT flavors............................................  4
         2.2.1 Traditional-NAT-PT...................................  4
         2.2.2 Bi-directional-NAT-PT................................  5
      2.3 Protocol Translation (PT).................................  5
      2.4 Application Level Gateway (ALG)...........................  5
      2.5 Requirements..............................................  5
   3. Traditional-NAT-PT operation (V6 to V4).......................  6
      3.1 NAT-PT Outgoing Sessions..................................  6
      3.2 NAPT-PT Outgoing Sessions.................................  7
   4. Use of DNS-ALG for Address assignment.........................  8
      4.1 V4 Address Assignment for Incoming Connections (V4 to V6).  9
      4.2 V4 Address Assignment for Outgoing Connections (V6 to V4). 11
   5. Protocol Translation Details.................................. 12
      5.1 Translating IPv4 Headers to IPv6 Headers.................. 13
      5.2 Translating IPv6 Headers to IPv4 Headers.................. 13
      5.3 TCP/UDP/ICMP Checksum Update.............................. 13
   6. FTP Application Level Gateway (FTP-ALG) Support............... 14
      6.1 Payload modifications for V4 originated FTP sessions...... 15
      6.2 Payload modifications for V6 originated FTP sessions...... 16
      6.3 Header updates for FTP control packets.................... 16
   7. NAT-PT Limitations and Future Work............................ 17
      7.1 Topology Limitations...................................... 17
      7.2 Protocol Translation Limitations.......................... 17
      7.3 Impact of Address Translation............................. 18
      7.4 Lack of End-to-End Security............................... 18
      7.5 DNS Translation and DNSSEC................................ 18
   8. Applicability Statement....................................... 18
   9. Security Considerations....................................... 19
   10. References................................................... 19
   Authors' Addresses............................................... 20
   Full Copyright Statement......................................... 21

1. Introduction

IPv6 is a new version of the IP protocol designed to modernize IPv4 which was designed in the 1970s. IPv6 has a number of advantages over IPv4 that will allow for future Internet growth and will simplify IP configuration and administration. IPv6 has a larger address space than IPv4, an addressing model that promotes aggressive route aggregation and a powerful autoconfiguration mechanism. In time, it is expected that Internet growth and a need for a plug-and-play solution will result in widespread adoption of IPv6.

There is expected to be a long transition period during which it will be necessary for IPv4 and IPv6 nodes to coexist and communicate. A strong, flexible set of IPv4-to-IPv6 transition and coexistence mechanisms will be required during this transition period.

The SIIT proposal [SIIT] describes a protocol translation mechanism that allows communication between IPv6-only and IPv4-only nodes via protocol independent translation of IPv4 and IPv6 datagrams, requiring no state information for the session. The SIIT proposal assumes that V6 nodes are assigned a V4 address for communicating with V4 nodes, and does not specify a mechanism for the assignment of these addresses.

   NAT-PT uses a pool of V4 addresses for assignment to V6 nodes on a
   dynamic basis as sessions are initiated across V4-V6  boundaries. The
   V4 addresses are assumed to be globally unique. NAT-PT with private
   V4 addresses is outside the scope of this document and for further
   study.  NAT-PT binds addresses in V6 network with addresses in V4
   network and vice versa to provide transparent routing [NAT-TERM] for
   the datagrams traversing between address realms. This requires no
   changes to end nodes and IP packet routing is completely transparent
   [NAT-TERM] to end nodes. It does, however, require NAT-PT to track
   the sessions it supports and mandates that inbound and outbound
   datagrams pertaining to a session traverse the same NAT-PT router.
   You will note that the topology restrictions on NAT-PT are the same
   with those described for V4 NATs in [NAT-TERM]. Protocol translation
   details specified in [SIIT] would be used to extend address
   translation with protocol syntax/semantics translation. A detailed
   applicability statement for NAT-PT may be found at the end of this
   document in section 7.

By combining SIIT protocol translation with the dynamic address translation capabilities of NAT and appropriate ALGs, NAT-PT provides a complete solution that would allow a large number of commonly used applications to interoperate between IPv6-only nodes and IPv4-only

A fundamental assumption for NAT-PT is only to be use when no other native IPv6 or IPv6 over IPv4 tunneled means of communication is possible. In other words the aim is to only use translation between IPv6 only nodes and IPv4 only nodes, while translation between IPv6 only nodes and the IPv4 part of a dual stack node should be avoided over other alternatives.

2. Terminology

The majority of terms used in this document are borrowed almost as is from [NAT-TERM]. The following lists terms specific to this document.

2.1 Network Address Translation (NAT)

The term NAT in this document is very similar to the IPv4 NAT described in [NAT-TERM], but is not identical. IPv4 NAT translates one IPv4 address into another IPv4 address. In this document, NAT refers to translation of an IPv4 address into an IPv6 address and vice versa.

While the V4 NAT [NAT-TERM] provides routing between private V4 and external V4 address realms, NAT in this document provides routing between a V6 address realm and an external V4 address realm.

2.2 NAT-PT flavors

Just as there are various flavors identified with V4 NAT in [NAT- TERM], the following NAT-PT variations may be identified in this document.

2.2.1 Traditional NAT-PT

Traditional-NAT-PT would allow hosts within a V6 network to access hosts in the V4 network. In a traditional-NAT-PT, sessions are uni- directional, outbound from the V6 network. This is in contrast with Bi-directional-NAT-PT, which permits sessions in both inbound and outbound directions.

   Just as with V4 traditional-NAT, there are two variations to
   traditional-NAT-PT, namely Basic-NAT-PT and NAPT-PT.

With Basic-NAT-PT, a block of V4 addresses are set aside for translating addresses of V6 hosts as they originate sessions to the V4 hosts in external domain. For packets outbound from the V6 domain, the source IP address and related fields such as IP, TCP, UDP and ICMP header checksums are translated. For inbound packets, the destination IP address and the checksums as listed above are translated.

NAPT-PT extends the notion of translation one step further by also translating transport identifier (e.g., TCP and UDP port numbers, ICMP query identifiers). This allows the transport identifiers of a number of V6 hosts to be multiplexed into the transport identifiers of a single assigned V4 address. NAPT-PT allows a set of V6 hosts to share a single V4 address. Note that NAPT-PT can be combined with Basic-NAT-PT so that a pool of external addresses are used in conjunction with port translation.

For packets outbound from the V6 network, NAPT-PT would translate the source IP address, source transport identifier and related fields such as IP, TCP, UDP and ICMP header checksums. Transport identifier can be one of TCP/UDP port or ICMP query ID. For inbound packets, the destination IP address, destination transport identifier and the IP and transport header checksums are translated.

2.2.2 Bi-Directional-NAT-PT

With Bi-directional-NAT-PT, sessions can be initiated from hosts in V4 network as well as the V6 network. V6 network addresses are bound to V4 addresses, statically or dynamically as connections are established in either direction. The name space (i.e., their Fully Qualified Domain Names) between hosts in V4 and V6 networks is assumed to be end-to-end unique. Hosts in V4 realm access V6-realm hosts by using DNS for address resolution. A DNS-ALG [DNS-ALG] must be employed in conjunction with Bi-Directional-NAT-PT to facilitate name to address mapping. Specifically, the DNS-ALG must be capable of translating V6 addresses in DNS Queries and responses into their V4-address bindings, and vice versa, as DNS packets traverse between V6 and V4 realms.

2.3 Protocol Translation (PT)

PT in this document refers to the translation of an IPv4 packet into a semantically equivalent IPv6 packet and vice versa. Protocol translation details are described in [SIIT].

2.4 Application Level Gateway (ALG)

Application Level Gateway (ALG) [NAT-TERM] is an application specific agent that allows a V6 node to communicate with a V4 node and vice versa. Some applications carry network addresses in payloads. NAT-PT is application unaware and does not snoop the payload. ALG could work in conjunction with NAT-PT to provide support for many such applications.

2.5 Requirements

The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this document, are to be interpreted as described in [KEYWORDS].

3. Traditional-NAT-PT Operation (V6 to V4)

NAT-PT offers a straight forward solution based on transparent routing [NAT-TERM] and address/protocol translation, allowing a large number of applications in V6 and V4 realms to inter-operate without requiring any changes to these applications.

In the following paragraphs we describe the operation of traditional-NAT-PT and the way that connections can be initiated from a host in IPv6 domain to a host in IPv4 domain through a traditional-NAT-PT

3.1 Basic-NAT-PT Operation

          [IPv6-B]-+
                   |                  +==============+
          [IPv6-A]-+-[NAT-PT]---------| IPv4 network |--[IPv4-C]
                        |             +==============+
                 (pool of v4 addresses)
          
                     Figure 1: IPv6 to IPv4 communication
           Node IPv6-A has an IPv6 address -> FEDC:BA98::7654:3210
           Node IPv6-B has an IPv6 address -> FEDC:BA98::7654:3211
              Node IPv4-C has an IPv4 address -> 132.146.243.30

NAT-PT has a pool of addresses including the IPv4 subnet 120.130.26/24

The V4 addresses in the address pool could be allocated one-to-one to the V6 addresses of the V6 end nodes in which case one needs as many V4 addresses as V6 end points. In this document we assume that the V6 network has less V4 addresses than V6 end nodes and thus dynamic address allocation is required for at least some of them.

Say the IPv6 Node A wants to communicate with the IPv4 Node C. Node A creates a packet with:

      Source Address, SA=FEDC:BA98::7654:3210 and Destination
      Address, DA = PREFIX::132.146.243.30

NOTE: The prefix PREFIX::/96 is advertised in the stub domain by the NAT-PT, and packets addressed to this PREFIX will be routed to the NAT-PT. The pre-configured PREFIX only needs to be routable within the IPv6 stub domain and as such it can be any routable prefix that the network administrator chooses.

The packet is routed via the NAT-PT gateway, where it is translated to IPv4.

If the outgoing packet is not a session initialisation packet, the NAT-PT SHOULD already have stored some state about the related session, including assigned IPv4 address and other parameters for the translation. If this state does not exist, the packet SHOULD be silently discarded.

   If the packet is a session initialisation packet, the NAT-PT locally
   allocates an address (e.g: 120.130.26.10)  from  its pool of
   addresses and the packet is translated to IPv4. The translation
   parameters are cached for the duration of the session and the IPv6 to
   IPv4 mapping is retained by NAT-PT.
   
   The resulting IPv4 packet has SA=120.130.26.10 and DA=132.146.243.30.
   Any returning traffic will be recognised as belonging to the same
   session by NAT-PT. NAT-PT will use the state information to translate
   the packet, and the resulting  addresses will be
   SA=PREFIX::132.146.243.30, DA=FEDC:BA98::7654:3210.  Note that this
   packet can now be routed inside the IPv6-only stub network as normal.

3.2 NAPT-PT Operation

NAPT-PT, which stands for "Network Address Port Translation + Protocol Translation", would allow V6 nodes to communicate with the V4 nodes transparently using a single V4 address. The TCP/UDP ports of the V6 nodes are translated into TCP/UDP ports of the registered V4 address.

While NAT-PT support is limited to TCP, UDP and other port multiplexing type of applications, NAPT-PT solves a problem that is inherent with NAT-PT. That is, NAT-PT would fall flat when the pool of V4 addresses assigned for translation purposes is exhausted. Once the address pool is exhausted, newer V6 nodes cannot establish sessions with the outside world anymore. NAPT-PT, on the other hand, will allow for a maximum of 63K TCP and 63K UDP sessions per IPv4 address before having no TCP and UDP ports left to assign.

To modify the example sited in figure 1, we could have NAPT-PT on the border router (instead of NAT-PT) and all V6 addresses could be mapped to a single v4 address 120.130.26.10.

IPv6 Node A would establish a TCP session with the IPv4 Node C as follows:

Node A creates a packet with:

Source Address, SA=FEDC:BA98::7654:3210 , source TCP port = 3017 and Destination Address, DA = PREFIX::132.146.243.30, destination TCP port = 23.

When the packet reaches the NAPT-PT box, NAPT-PT would assign one of the TCP ports from the assigned V4 address to translate the tuple of (Source Address, Source TCP port) as follows:

      SA=120.130.26.10, source TCP port = 1025  and
      DA=132.146.243.30, destination TCP port = 23.

The returning traffic from 132.146.243.30, TCP port 23 will be recognised as belonging to the same session and will be translated back to V6 as follows:

      SA = PREFIX::132.146.243.30, source TCP port = 23;
      DA = FEDC:BA98::7654:3210 , destination TCP port = 3017

Inbound NAPT-PT sessions are restricted to one server per service, assigned via static TCP/UDP port mapping. For example, the Node [IPv6-A] in figure 1 may be the only HTTP server (port 80) in the V6 domain. Node [IPv4-C] sends a packet:

      SA=132.146.243.30, source TCP port = 1025  and
      DA=120.130.26.10, destination TCP port = 80

NAPT-PT will translate this packet to:

      SA=PREFIX::132.146.243.30, source TCP port = 1025
      DA=FEDC:BA98::7654:3210, destination TCP port = 80

In the above example, note that all sessions which reach NAPT-PT with a destination port of 80 will be redirected to the same node [IPv6- A].

4. Use of DNS-ALG for Address Assignment

An IPv4 address is assigned by NAT-PT to a V6 node when NAT-PT identifies the start of session, inbound or outbound. Identification of the start of a new inbound session is performed differently than for outbound sessions. However, the same V4 address pool is used for assignment to V6 nodes, irrespective of whether a session is initiated outbound from a V6 node or initiated inbound from a V4 node.

Policies determining what type of sessions are allowed and in which direction and from/to which nodes is out of the scope of this document.

IPv4 name to address mappings are held in the DNS with "A" records. IPv6 name to address mappings are at the moment held in the DNS with "AAAA" records. "A6" records have also been defined but at the time of writing they are neither fully standardized nor deployed.

In any case, the DNS-ALG's principle of operation described in this section is the same with either "AAAA" or "A6" records. The only difference is that a name resolution using "A6" records may require more than one query - reply pairs. The DNS-ALG SHOULD, in that case, track all the replies in the transaction before translating an "A6" record to an "A" record.

One of the aims of NAT-PT design is to only use translation when there is no other means of communication, such as native IPv6 or some form of tunneling. For the following discussion NAT-PT, in addition to the IPv4 connectivity that it has it may also have a native IPv6 and/or a tunneled IPv6 connection.

4.1 V4 Address assignment for incoming connections (V4 to V6)

        [DNS]--+
               |              [DNS]------[DNS]-------[DNS]
      [IPv6-B]-+                           |           |
               |                  +==============+     |
      [IPv6-A]-+----[NAT-PT]------| IPv4 network |--[IPv4-C]
                       |          +==============+
                 (pool of v4 addresses)
      
                     Figure 2: IPv4 to IPv6 communication
           Node IPv6-A has an IPv6 address -> FEDC:BA98::7654:3210
           Node IPv6-B has an IPv6 address -> FEDC:BA98::7654:3211
              Node IPv4-C has an IPv4 address -> 132.146.243.30
   
   NAT-PT  has a pool of addresses including the IPv4 subnet
   120.130.26/24

In figure 2 above, when Node C's name resolver sends a name look up request for Node A, the lookup query is directed to the DNS server on the V6 network. Considering that NAT-PT is residing on the border router between V4 and V6 networks, this request datagram would traverse through the NAT-PT router. The DNS-ALG on the NAT-PT device would modify DNS Queries for A records going into the V6 domain as follows: (Note that a TCP/UDP DNS packet is recognised by the fact that its source or destination port number is 53)

      a) For Node Name to Node Address Query requests:  Change the Query
         type from "A" to "AAAA" or "A6".
      
      b) For Node address to Node name query requests:  Replace the
         string "IN-ADDR.ARPA" with the string "IP6.INT".  Replace the
         V4 address octets (in reverse order) preceding the string "IN-
         ADDR.ARPA" with the corresponding V6 address (if there exists a
         map) octets in reverse order.

In the opposite direction, when a DNS response traverses from the DNS server on the V6 network to the V4 node, the DNS-ALG once again intercepts the DNS packet and would:

      a) Translate DNS responses for "AAAA" or "A6" records into "A"
         records, (only translate "A6" records when the name has
         completely been resolved)
      b) Replace the V6 address resolved by the V6 DNS with the V4
         address internally assigned by the NAT-PT router.

If a V4 address is not previously assigned to this V6 node, NAT-PT would assign one at this time. As an example say IPv4-C attempts to initialise a session with node IPv6-A by making a name lookup ("A" record) for Node-A . The name query goes to the local DNS and from there it is propagated to the DNS server of the IPv6 network. The DNS-ALG intercepts and translates the "A" query to "AAAA" or "A6" query and then forwards it to the DNS server in the IPv6 network which replies as follows: (The example uses AAAA records for convenience)

      Node-A    AAAA     FEDC:BA98::7654:3210,

this is returned by the DNS server and gets intercepted and translated by the DNS-ALG to:

      Node-A     A      120.130.26.1
   
   The DNS-ALG also holds the mapping between FEDC:BA98::7654:3210 and
   120.130.26.1 in NAT-PT. The "A" record is then returned to Node-C.
   Node-C can now  initiate a session as follows:
   
      SA=132.146.243.30, source TCP port = 1025  and
      DA=120.130.26.1, destination TCP port = 80
   
   the packet will be routed to NAT-PT, which since it already holds a
   mapping between  FEDC:BA98::7654:3210 and 120.130.26.1 can translate
   the packet to:
   
      SA=PREFIX::132.146.243.30, source TCP port = 1025
      DA=FEDC:BA98::7654:3210, destination TCP port = 80

the communication can now proceed as normal.

The TTL values on all DNS resource records (RRs) passing through NAT-PT SHOULD be set to 0 so that DNS servers/clients do not cache temporarily assigned RRs. Note, however, that due to some buggy DNS client implementations a value of 1 might in some cases work better. The TTL values should be left unchanged for statically mapped addresses.

Address mappings for incoming sessions, as described above, are subject to denial of service attacks since one can make multiple queries for nodes residing in the V6 network causing the DNS-ALG to map all V4 addresses in NAT-PT and thus block legitimate incoming sessions. Thus, address mappings for incoming sessions should time out to minimise the effect of denial of service attacks. Additionally, one IPv4 address (using NAPT-PT, see 3.2) could be reserved for outgoing sessions only to minimise the effect of such attacks to outgoing sessions.

4.2 V4 Address assignment for outgoing connections (V6 to V4)

V6 nodes learn the address of V4 nodes from the DNS server in the V4 domain or from the DNS server internal to the V6 network. We recommend that DNS servers internal to V6 domains maintain a mapping of names to IPv6 addresses for internal nodes and possibly cache mappings for some external nodes. In the case where the DNS server in the v6 domain contains the mapping for external V4 nodes, the DNS queries will not cross the V6 domain and that would obviate the need for DNS-ALG intervention. Otherwise, the queries will cross the V6 domain and are subject to DNS-ALG intervention. We recommend external DNS servers in the V4 domain cache name mapping for external nodes (i.e., V4 nodes) only. Zone transfers across IPv4 - IPv6 boundaries are strongly discouraged.

In the case of NAPT-PT, a TCP/UDP source port is assigned from the registered V4 address upon detection of each new outbound session.

We saw that a V6 node that needs to communicate with a V4 node needs to use a specific prefix (PREFIX::/96) in front of the IPv4 address of the V4 node. The above technique allows the use of this PREFIX without any configuration in the nodes.

To create another example from Figure 2 say Node-A wants to set up a session with Node-C. For this Node-A starts by making a name look-up ("AAAA" or "A6" record) for Node-C.

Since Node-C may have IPv6 and/or IPv4 addresses, the DNS-ALG on the NAT-PT device forwards the original AAAA/A6 query to the external DNS system unchanged, as well as an A query for the same node. If an AAAA/A6 record exists for the destination, this will be returned to

NAT-PT which will forward it, also unchanged, to the originating host.

If there is an A record for Node-C the reply also returns to the NAT-PT. The DNS-ALG then, translates the reply adding the appropriate PREFIX and forwards it to the originating device with any IPv6 addresses that might have learned. So, if the reply is

      NodeC    A     132.146.243.30, it is translated to
      NodeC   AAAA   PREFIX::132.146.243.30 or to
      NodeC    A6    PREFIX::132.146.243.30

Now Node A can use this address like any other IPv6 address and the V6 DNS server can even cache it as long as the PREFIX does not change.

An issue here is how the V6 DNS server in the V6 stub domain talks to the V4 domain outside the V6 stub domain. Remember that there are no dual stack nodes here. The external V4 DNS server needs to point to a V4 address, part of the V4 pool of addresses, available to NAT-PT. NAT-PT keeps a one-to-one mapping between this V4 address and the V6 address of the internal V6 DNS server. In the other direction, the V6 DNS server points to a V6 address formed by the IPv4 address of the external V4 DNS servers and the prefix (PREFIX::/96) that indicates non IPv6 nodes. This mechanism can easily be extended to accommodate secondary DNS servers.

Note that the scheme described in this section impacts DNSSEC. See section 7.5 of this document for details.

5. Protocol Translation Details

The IPv4 and ICMPv4 headers are similar to their V6 counterparts but a number of field are either missing, have different meaning or different length. NAT-PT SHOULD translate all IP/ICMP headers from v4 to v6 and vice versa in order to make end-to-end IPv6 to IPv4 communication possible. Due to the address translation function and possible port multiplexing, NAT-PT SHOULD also make appropriate adjustments to the upper layer protocol (TCP/UDP) headers. A separate section on FTP-ALG describes the changes FTP-ALG would make to FTP payload as an FTP packet traverses from V4 to V6 realm or vice versa.

Protocol Translation details are described in [SIIT], but there are some modifications required to SIIT because of the fact that NAT-PT also performs Network Address Translation.

5.1 Translating IPv4 headers to IPv6 headers

This is done exactly the same as in SIIT apart from the following fields:

Source Address:: The low-order 32 bits is the IPv4 source address. The high- order 96 bits is the designated PREFIX for all v4 communications. Addresses using this PREFIX will be routed to the NAT-PT gateway (PREFIX::/96)

Destination Address:: NAT-PT retains a mapping between the IPv4 destination address and the IPv6 address of the destination node. The IPv4 destination address is replaced by the IPv6 address retained in that mapping.

5.2 Translating IPv6 headers to IPv4 headers

This is done exactly the same as in SIIT apart from the Source Address which should be determined as follows:

Source Address:: The NAT-PT retains a mapping between the IPv6 source address and an IPv4 address from the pool of IPv4 addresses available. The IPv6 source address is replaced by the IPv4 address retained in that mapping.

Destination Address:: IPv6 packets that are translated have a destination address of the form PREFIX::IPv4/96. Thus the low-order 32 bits of the IPv6 destination address is copied to the IPv4 destination address.

5.3 TCP/UDP/ICMP Checksum Update

NAT-PT retains mapping between IPv6 address and an IPv4 address from the pool of IPv4 addresses available. This mapping is used in the translation of packets that go through NAT-PT.

The following sub-sections describe TCP/UDP/ICMP checksum update procedure in NAT-PT, as packets are translated from V4 to V6 and vice versa.

5.3.1 TCP/UDP/ICMP Checksum Update from IPv4 to IPv6

UDP checksums, when set to a non-zero value, and TCP checksum SHOULD be recalculated to reflect the address change from v4 to v6. The incremental checksum adjustment algorithm may be borrowed from [NAT]. In the case of NAPT-PT, TCP/UDP checksum should be adjusted to account for the address and TCP/UDP port changes, going from V4 to V6 address.

When the checksum of a V4 UDP packet is set to zero, NAT-PT MUST evaluate the checksum in its entirety for the V6-translated UDP packet. If a V4 UDP packet with a checksum of zero arrives in fragments, NAT-PT MUST await all the fragments until they can be assembled into a single non-fragmented packet and evaluate the checksum prior to forwarding the translated V6 UDP packet.

ICMPv6, unlike ICMPv4, uses a pseudo-header, just like UDP and TCP during checksum computation. As a result, when the ICMPv6 header checksum is computed [SIIT], the checksum needs to be adjusted to account for the additional pseudo-header. Note, there may also be adjustments required to the checksum due to changes in the source and destination addresses (and changes in TCP/UDP/ICMP identifiers in the case of NAPT-PT) of the payload carried within ICMP.

5.3.2 TCP/UDP/ICMP Checksum Update from IPv6 to IPv4

TCP and UDP checksums SHOULD be recalculated to reflect the address change from v6 to v4. The incremental checksum adjustment algorithm may be borrowed from [NAT]. In the case of NAPT-PT, TCP/UDP checksums should be adjusted to account for the address and TCP/UDP port changes, going from V6 to V4 addresses. For UDP packets, optionally, the checksum may simply be changed to zero.

The checksum calculation for a V4 ICMP header needs to be derived from the V6 ICMP header by running the checksum adjustment algorithm [NAT] to remove the V6 pseudo header from the computation. Note, the adjustment must additionally take into account changes to the checksum as a result of updates to the source and destination addresses (and transport ports in the case of NAPT-PT) made to the payload carried within ICMP.

6. FTP Application Level Gateway (FTP-ALG) Support

Because an FTP control session carries, in its payload, the IP address and TCP port information for the data session, an FTP-ALG is required to provide application level transparency for this popular Internet application.

In the FTP application running on a legacy V4 node, arguments to the FTP PORT command and arguments in PASV response(successful) include an IP V4 address and a TCP port, both represented in ASCII as h1,h2,h3,h4,p1,p2. However, [FTP-IPV6] suggests EPRT and EPSV command extensions to FTP, with an intent to eventually retire the use of PORT and PASV commands. These extensions may be used on a V4 or V6 node. FTP-ALG, facilitating transparent FTP between V4 and V6 nodes, works as follows.

6.1 Payload modifications for V4 originated FTP sessions

A V4 host may or may not have the EPRT and EPSV command extensions implemented in its FTP application. If a V4 host originates the FTP session and uses PORT or PASV command, the FTP-ALG will translate these commands into EPRT and EPSV commands respectively prior to forwarding to the V6 node. Likewise, EPSV response from V6 nodes will be translated into PASV response prior to forwarding to V4 nodes. The format of EPRT and EPSV commands and EPSV response may be specified as follows[FTP-IPV6].

      EPRT<space><d><net-prt><d><net-addr><d><tcp-port><d>
      EPSV<space><net-prt>
            (or)
      EPSV<space>ALL

Format of EPSV response(Positive): 229 <text indicating extended passive mode> (<d><d><d><tcp-port><d>)

PORT command from a V4 node is translated into EPRT command, by setting the protocol <net-prt> field to AF #2 (IPV6) and translating the V4 host Address (represented as h1,h2,h3,h4) into its NAT-PT assigned V6 address in string notation, as defined in [V6ADDR] in the <net-addr> field. TCP port represented by p1,p2 in PORT command must be specified as a decimal <tcp-port> in the EPRT command. Further, <tcp-port> translation may also be required in the case of NAPT-PT. PASV command from a V4 node is be translated into a EPSV command with the <net-prt> argument set to AF #2. EPSV response from a V6 node is translated into PASV response prior to forwarding to the target V4 host.

If a V4 host originated the FTP session and was using EPRT and EPSV commands, the FTP-ALG will simply translate the parameters to these commands, without altering the commands themselves. The protocol Number <net-prt> field will be translated from AF #1 to AF #2. <net-addr> will be translated from the V4 address in ASCII to its NAT-PT assigned V6 address in string notation as defined in [V6ADDR]. <tcp-port> argument in EPSV response requires translation only in the case of NAPT-PT.

6.2 Payload modifications for V6 originated FTP sessions

If a V6 host originates the FTP session, however, the FTP-ALG has two approaches to pursue. In the first approach, the FTP-ALG will leave the command strings "EPRT" and "EPSV" unaltered and simply translate the <net-prt>, <net-addr> and <tcp-port> arguments from V6 to its NAT-PT (or NAPT-PT) assigned V4 information. <tcp-port> is translated only in the case of NAPT-PT. Same goes for EPSV response from V4 node. This is the approach we recommend to ensure forward support for RFC 2428. However, with this approach, the V4 hosts are mandated to have their FTP application upgraded to support EPRT and EPSV extensions to allow access to V4 and V6 hosts, alike.

In the second approach, the FTP-ALG will translate the command strings "EPRT" and "EPSV" and their parameters from the V6 node into their equivalent NAT-PT assigned V4 node info and attach to "PORT" and "PASV" commands prior to forwarding to V4 node. Likewise, PASV response from V4 nodes is translated into EPSV response prior to forwarding to the target V6 nodes. However, the FTP-ALG would be unable to translate the command "EPSV<space>ALL" issued by V6 nodes. In such a case, the V4 host, which receives the command, may return an error code indicating unsupported function. This error response may cause many RFC 2428 compliant FTP applications to simply fail, because EPSV support is mandated by RFC 2428. The benefit of this approach, however, is that is does not impose any FTP upgrade requirements on V4 hosts.

6.3 Header updates for FTP control packets

All the payload translations considered in the previous sections are based on ASCII encoded data. As a result, these translations may result in a change in the size of packet.

If the new size is the same as the previous, only the TCP checksum needs adjustment as a result of the payload translation. If the new size is different from the previous, TCP sequence numbers should also be changed to reflect the change in the length of the FTP control session payload. The IP packet length field in the V4 header or the IP payload length field in the V6 header should also be changed to reflect the new payload size. A table is used by the FTP-ALG to correct the TCP sequence and acknowledgement numbers in the TCP header for control packets in both directions.

The table entries should have the source address, source data port, destination address and destination data port for V4 and V6 portions of the session, sequence number delta for outbound control packets and sequence number delta for inbound control packets.

The sequence number for an outbound control packet is increased by the outbound sequence number delta, and the acknowledgement number for the same outbound packet is decreased by the inbound sequence number delta. Likewise, the sequence number for an inbound packet is increased by the inbound sequence number delta and the acknowledgement number for the same inbound packet is decreased by the outbound sequence number delta.

7. NAT-PT Limitations and Future Work

All limitations associated to NAT [NAT-TERM] are also associated to NAT-PT. Here are the most important of them in detail, as well as some unique to NAT-PT.

7.1 Topology limitations

There are limitations to using the NAT-PT translation method. It is mandatory that all requests and responses pertaining to a session be routed via the same NAT-PT router. One way to guarantee this would be to have NAT-PT based on a border router that is unique to a stub domain, where all IP packets are either originated from the domain or destined to the domain. This is a generic problem with NAT and it is fully described in [NAT-TERM].

Note, this limitation does not apply to packets originating from or directed to dual-stack nodes that do not require packet translation. This is because in a dual-stack set-up, IPv4 addresses implied in a V6 address can be identified from the address format PREFIX::x.y.z.w and a dual-stack router can accordingly route a packet between v4 and dual-stack nodes without tracking state information.

This should also not affect IPv6 to IPv6 communication and in fact only actually use translation when no other means of communication is possible. For example NAT-PT may also have a native IPv6 connection and/or some kind of tunneled IPv6 connection. Both of the above connections should be preferred over translation when possible. The above makes sure that NAT-PT is a tool only to be used to assist transition to native IPv6 to IPv6 communication.

7.2 Protocol Translation Limitations

A number of IPv4 fields have changed meaning in IPv6 and translation is not straightforward. For example, the option headers semantics and syntax have changed significantly in IPv6. Details of IPv4 to IPv6 Protocol Translation can be found in [SIIT].

7.3 Impact of Address Translation

Since NAT-PT performs address translation, applications that carry the IP address in the higher layers will not work. In this case Application Layer Gateways (ALG) need to be incorporated to provide support for those applications. This is a generic problem with NAT and it is fully described in [NAT-TERM].

7.4 Lack of end-to-end security

One of the most important limitations of the NAT-PT proposal is the fact that end-to-end network layer security is not possible. Also transport and application layer security may not be possible for applications that carry IP addresses to the application layer. This is an inherent limitation of the Network Address Translation function.

Independent of NAT-PT, end-to-end IPSec security is not possible across different address realms. The two end-nodes that seek IPSec network level security must both support one of IPv4 or IPv6.

7.5 DNS Translation and DNSSEC

The scheme described in section 4.2 involves translation of DNS messages. It is clear that this scheme can not be deployed in combination with secure DNS. I.e., an authoritative DNS name server in the V6 domain cannot sign replies to queries that originate from the V4 world. As a result, an V4 end-node that demands DNS replies to be signed will reject replies that have been tampered with by NAT-PT.

The good news, however, is that only servers in V6 domain that need to be accessible from the V4 world pay the price for the above limitation, as V4 end-nodes may not access V6 servers due to DNS replies not being signed.

Also note that zone transfers between DNS-SEC servers within the same V6 network are not impacted.

Clearly, with DNS SEC deployment in DNS servers and end-host resolvers, the scheme suggested in this document would not work.

8. Applicability Statement

NAT-PT can be a valuable transition tool at the border of a stub network that has been deployed as an IPv6 only network when it is connected to an Internet that is either V4-only or a combination of V4 and V6.

   NAT-PT, in its simplest form, without the support of DNS-ALG,
   provides one way connectivity between an IPv6 stub domain and the
   IPv4  world meaning  that only sessions initialised by IPv6 nodes
   internal to the IPv6 stub domain can be translated, while sessions
   initiated by  IPv4 nodes  are dropped. This makes NAT-PT a useful
   tool to IPv6 only stub networks that need to be able to maintain
   connectivity with the  IPv4 world without the need to deploy servers
   visible to the IPv4 world.
   
   NAT-PT  combined  with a DNS-ALG provides bi-directional connectivity
   between the IPv6 stub domain and the IPv4 world allowing sessions  to
   be  initialised  by  IPv4  nodes  outside the IPv6 stub domain.  This
   makes NAT-PT useful for IPv6 only stub  networks that need to  deploy
   servers visible to the IPv4 world.

Some applications count on a certain degree of address stability for their operation. Dynamic address reuse by NAT-PT might not be agreeable for these applications. For hosts running such address critical applications, NAT-PT may be configured to provide static address mapping between the host's V6 address and a specific V4 address. This will ensure that address related changes by NAT-PT do not become a significant source of operational failure.

9. Security Considerations

Section 7.4 of this document states that end-to-end network and transport layer security are not possible when a session is intercepted by a NAT-PT. Also application layer security may not be possible for applications that carry IP addresses in the application layer.

Section 7.5 of this document states that the DNS-ALG can not be deployed in combination with secure DNS.

Finally, all of the security considerations described in [NAT-TERM] are applicable to this document as well.

10. REFERENCES

   [DNS-ALG]  Srisuresh, P., Tsirtsis, G., Akkiraju, P. and A.
              Heffernan, "DNS extensions to Network Address Translators
              (DNS_ALG)", RFC 2694, September 1999.
   
   [DNSSEC]   Eastlake, D., "Domain Name System Security Extensions",
              RFC 2065, March 1999.
   
   [FTP-IPV6] Allman, M., Ostermann, S. and C. Metz, "FTP Extensions for
              IPv6 and NATs", RFC 2428, September 1998.
   
   [KEYWORDS] Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.
   
   [NAT]      Egevang, K. and P. Francis, "The IP Network Address
              Translator (NAT)", RFC 1631, May 1994.
   
   [NAT-TERM] Srisuresh, P. and M. Holdrege, "IP Network Address
              Translator (NAT) Terminology and Considerations", RFC
              2663, August 1999.
   
   [SIIT]     Nordmark, E., "Stateless IP/ICMP Translator (SIIT)", RFC
              2765, February 2000.
   
   [TRANS]    Gilligan, R. and  E. Nordmark, "Transition Mechanisms for
              IPv6 Hosts and Routers", RFC 1933, April 1996.
   
   [V6ADDR]   Hinden, R. and S. Deering, "IP Version 6 Addressing
              Architecture", RFC 2373, July 1998.

Authors' Addresses

George Tsirtsis
Internet Futures
B29 Room 129
BT Adastral Park
IPSWICH IP5 3RE
England

   Phone: +44 181 8260073
   Fax:   +44 181 8260073
   EMail: george.tsirtsis@bt.com
   EMail (alternative): gtsirt@hotmail.com

Pyda Srisuresh
630 Alder Drive
Milpitas, CA 95035
U.S.A.

   Phone: (408) 519-3849
   EMail: srisuresh@yahoo.com

Full Copyright Statement

This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English.

The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns.

This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Acknowledgement

Funding for the RFC Editor function is currently provided by the Internet Society.