IP Security Protocol Working Group (IPSEC)                    T. Kivinen
INTERNET-DRAFT                               SSH Communications Security
draft-ietf-ipsec-nat-t-ike-03.txt                            A. Huttunen
Expires: 24 December 2002                          F- Secure Corporation
                                                              B. Swander
                                                                V. Volpe
                                                           Cisco Systems
                                                            24 June 2002

                Negotiation of NAT-Traversal in the IKE

Status of This Memo

This document is a submission to the IETF IP Security Protocol
(IPSEC) Working Group.  Comments are solicited and should be
addressed to the working group mailing list (ipsec@lists.tislabs.com)
or to the editor.

This document is an Internet-Draft and is in full conformance
with all provisions of Section 10 of RFC2026.

Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups.  Note that
other groups may also distribute working documents as

Internet-Drafts are draft documents valid for a maximum of six
months and may be updated, replaced, or obsoleted by other
documents at any time.  It is inappropriate to use Internet-
Drafts as reference material or to cite them other than as
"work in progress."

The list of current Internet-Drafts can be accessed at

The list of Internet-Draft Shadow Directories can be accessed at


This document describes how to detect one or more NATs between IPsec
hosts, and how to negotiate the use of UDP encapsulation of the IPsec
packets through the NAT boxes in IKE.

T. Kivinen, et. al.                                             [page 1]

INTERNET-DRAFT                                              24 June 2002

Table of Contents

1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . .  2
2.  Specification of Requirements   . . . . . . . . . . . . . . . . .  2
3.  Phase 1   . . . . . . . . . . . . . . . . . . . . . . . . . . . .  2
  3.1.  Detecting support of Nat-Traversal  . . . . . . . . . . . . .  3
  3.2.  Detecting presence of NAT   . . . . . . . . . . . . . . . . .  3
4.  Floating to the new ports   . . . . . . . . . . . . . . . . . . .  5
5.  Quick Mode  . . . . . . . . . . . . . . . . . . . . . . . . . . .  7
  5.1.  Negotiation of the NAT-Traversal encapsulation  . . . . . . .  7
  5.2.  Sending the original source address   . . . . . . . . . . . .  8
6.  Initial contact notifications   . . . . . . . . . . . . . . . . .  8
7.  Recovering from the expiring NAT mappings   . . . . . . . . . . .  9
8.  Security Considerations   . . . . . . . . . . . . . . . . . . . .  9
9.  Intellectual property rights  . . . . . . . . . . . . . . . . . . 10
10.  Acknowledgments  . . . . . . . . . . . . . . . . . . . . . . . . 10
11.  References   . . . . . . . . . . . . . . . . . . . . . . . . . . 10
12.  Authors' Addresses   . . . . . . . . . . . . . . . . . . . . . . 11

1.  Introduction

This document is split in two parts. The first part describes what is
needed in the IKE phase 1 for the NAT-Traversal support. This includes
detecting if the other end supports NAT-Traversal, and detecting if
there is one or more NAT along the path from host to host.

The second part describes how to negotiate the use of UDP encapsulated
IPsec packets in the IKE Quick Mode. It also describes how to transmit
the original source address to the other end if needed. The original
source address can be used to incrementally update the TCP/IP checksums
so that they will match after the NAT transform (The NAT cannot do this,
because the TCP/IP checksum is inside the UDP encapsulated IPsec

The document [Hutt02] describes the details of the UDP encapsulation and
the document [Dixon01] provides background information and motivation of
the NAT-Traversal in general.

2.  Specification of Requirements

This document shall use the keywords "MUST", "MUST NOT", "REQUIRED",
"OPTIONAL" to describe requirements. They are to be interpreted as
described in [RFC-2119] document.

3.  Phase 1

The detection of the support for the NAT-Traversal and detection of the
NAT along the path happens in the IKE [RFC-2409] phase 1.

The NAT is supposed to float the IKE UDP port, and recipients MUST be

T. Kivinen, et. al.                                             [page 2]

INTERNET-DRAFT                                              24 June 2002

able to process IKE packets whose source port is different than 500.
There are cases where the NAT does not have to float the source port
(only one (IPsec) host behind the NAT or for the first IPsec host it can
keep the port 500, and float only the following IPsec hosts).

Recipients MUST reply back to the source address from the packet. This
also means that when the original responder is doing rekeying, or
sending notifications etc. to the original initiator it MUST send the
packets from the same set of port and IP numbers that was used when the
IKE SA was last time used (i.e the source and destination port and IP
numbers must be same).

For example the initiator sends packet having source and destination
port 500, the NAT changes that to such packet which have source port
12312 and destination port 500, the responder must be able to process
the packet whose source address is that 12312 and it must reply back
with packet whose source address is 500 and destination address 12312,
the NAT will then translate this packet to have source address 500 and
destination address 500.

3.1.  Detecting support of Nat-Traversal

The NAT-Traversal capability of the remote host is determined by an
exchange of vendor strings; in Phase 1 two first messages, the vendor id
payload for this specification of NAT-Traversal (MD5 hash of "draft-
ietf-ipsec-nat-t-ike-03" - ["7d9419a6 5310ca6f 2c179d92 15529d56"]) MUST
be sent if supported (and it MUST be received by both sides) for the
NAT-Traversal probe to continue.

3.2.  Detecting presence of NAT

The purpose of the NAT-D payload is twofold, It not only detects the
presence of NAT between two IKE peers, it also detects where the NAT is.
The location of the NAT device is important in that the keepalives need
to initiate from the peer "behind" the NAT.

To detect the NAT between the two hosts, we need to detect if the IP
address or the port changes along the path. This is done by sending the
hashes of IP address and port of both source and destination addresses
from each end to another. When both ends calculate those hashes and get
same result they know there is no NAT between. If the hashes do not
match, somebody translated the address or port between, meaning we need
to do NAT-Traversal to get IPsec packet through.

If the sender of the packet does not know his own IP address (in case of
multiple interfaces, and implementation don't know which is used to
route the packet out), he can include multiple local hashes to the
packet (as separate NAT-D payloads). In this case the NAT is detected if
and only if none of the hashes match.

The hashes are sent as a series of NAT-D (NAT discovery) payloads.  Each
payload contains one hash, so in case of multiple hashes, multiple NAT-D
payloads are sent. In normal case there is only two NAT-D payloads.

T. Kivinen, et. al.                                             [page 3]

INTERNET-DRAFT                                              24 June 2002

The NAT-D payloads are included in the third and fourth packet in the
main mode and second and third packet in the aggressive mode.

The format of the NAT-D packet is

      1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
     | Next Payload  |    RESERVED   |        Payload length         |
     ~               HASH of the address and port                    ~

The payload type for the NAT discovery payload is 130 (XXX CHANGE).

The HASH is calculated as follows:

        HASH = HASH(CKY-I | CKY-R | IP | Port)

using the negotiated HASH algorithm. All data inside the HASH is in the
network byte-order. The IP is 4 octets for the IPv4 address and 16
octets for the IPv6 address. The port number is encoded as 2 octet
number in network byte-order. The first NAT-D payload contains the
remote ends IP address and port (i.e the destination address of the UDP
packet). The rest of the NAT-D payloads contain possible local end IP
addresses and ports (i.e all possible source addresses of the UDP

If there is no NAT between then the first NAT-D payload should match one
of the local NAT-D packet (i.e the local NAT-D payloads this host is
sending out), and the one of the other NAT-D payloads must match the
remote ends IP address and port. If the first check fails (i.e first
NAT-D payload does not match any of the local IP addresses and ports),
then it means that there is dynamic NAT between, and this end should
start sending keepalives as defined in the [Hutt02].

The CKY-I and CKY-R are the initiator and responder cookies, and they
are added to the hash to make precomputation attacks for the IP address
and port impossible.

An example of phase 1 exchange using NAT-Traversal in main mode
(authentication with signatures) is:

         Initiator                       Responder
        ------------                    ------------
        HDR, SA, VID                 -->
                                     <-- HDR, SA, VID
        HDR, KE, Ni, NAT-D, NAT-D    -->
                                     <-- HDR, KE, Nr, NAT-D, NAT-D
        HDR*#, IDii, [CERT, ] SIG_I   -->
                                     <-- HDR*#, IDir, [ CERT, ], SIG_R

An example of phase 1 exchange using NAT-Traversal in aggressive mode
(authentication with signatures) is:

T. Kivinen, et. al.                                             [page 4]

INTERNET-DRAFT                                              24 June 2002

         Initiator                       Responder
        ------------                    ------------
        HDR, SA, KE, Ni, IDii, VID   -->
                                     <-- HDR, SA, KE, Nr, IDir,
                                         [CERT, ], VID, NAT-D,
                                         NAT-D, SIG_R
        HDR*#, [CERT, ], NAT-D, NAT-D,
          SIG_I                      -->

The '#' sign identifies that those packets are sent to the floated port
if NAT is detected.

4.  Floating to the new ports

IPsec-aware NATs can cause problems. Some NATs will not float IKE source
port 500 even if there are multiple clients behind the NAT.  They can
also map IKE cookies to demultiplex traffic instead of using the source
port. Both of these are problematic for generic NAT transparency since
it is difficult for IKE to discover the capabilities of the NAT. The
best approach is to simply move the IKE traffic off port 500 as soon as
possible to avoid any IPsec-aware NAT special casing. Moving IKE from
port 500 to port 4500 is known as port floating.

Take the common case of the initiator behind the NAT. The initiator must
quickly float to 4500 once the NAT has been detected to minimize the
window of IPsec-aware NAT problems.

In main mode, the initiator MUST float on the ID payload if there is NAT
between the hosts. The initiator MUST set both UDP source and
destination ports to 4500. All subsequent packets sent to this peer
(including informationals) MUST be sent on 4500. In addition, the IKE
data MUST be prepended with a non-ESP marker allowing for demultiplexing
of traffic as defined in [Hutt02].

Thus, the IKE packet now looks like:

        IP UDP(4500,4500) <non-ESP marker> HDR*, IDii, [CERT, ] SIG_I

assuming authentication using signatures. The 4 bytes of non-ESP marker
is defined in the [Hutt02].

When the responder gets this packet he performs the usual decryption and
processing of the various payloads. If this is successful, he MUST
update local state so that all subsequent packets (including
informationals) to the peer use the new port, and possibly new IP
address obtained from the incoming valid packet. The port will generally
be different since the NAT will map UDP(500,500) to UDP(X,500), and
UDP(4500,4500) to UDP(Y,4500). The IP address will seldom be different
from the pre-float IP address. The responder MUST respond with all
subsequent IKE packets to this peer using UDP(4500,Y).

Similarly, if the responder needs to rekey the phase 1 SA, then he MUST
start the negotiation using UDP(4500,Y). Any implementation that

T. Kivinen, et. al.                                             [page 5]

INTERNET-DRAFT                                              24 June 2002

supports NAT traversal, MUST support negotiations that begin on port
4500. If a negotiation starts on 4500, then it doesn't need to float
anywhere else in the exchange.

Once floating has occurred, if a packet is received on 500, that packet
is old. If the packet is an informational, it MAY be processed if local
policy allows. If the packet is a main mode or aggressive mode packet,
it SHOULD be discarded.

Here is an example of phase 1 exchange using NAT-Traversal in main mode
(authentication with signatures) with port floating:

         Initiator                       Responder
        ------------                    ------------
        UDP(500,500) HDR, SA, VID    -->
                                     <-- UDP(500,X) HDR, SA, VID
        UDP(500,500) HDR, KE, Ni,
                     NAT-D, NAT-D    -->
                                     <-- UDP(500,X) HDR, KE, Nr,
                                                    NAT-D, NAT-D
        UDP(4500,4500) HDR*#, IDii,
                      [CERT, ]SIG_I  -->
                                     <-- UDP(4500,Y) HDR*#, IDir,
                                                   [ CERT, ], SIG_R

The floating algorithm for aggressive mode is very similar. After the
NAT has been detected, the initiator sends: IP UDP(4500,4500) <4 bytes
of non-ESP marker> HDR*, [CERT, ], NAT-D, NAT-D, SIG_I The responder
does similar processing to the above, and if successful, MUST update his
internal IKE ports. The responder MUST respond with all subsequent IKE
packets to this peer using UDP(4500,Y).

         Initiator                              Responder
        ------------                          ------------

        UDP(500,500) HDR, SA, KE,
                     Ni, IDii, VID   -->
                                     <-- UDP(500,X) HDR, SA, KE,
                                                    Nr, IDir, [CERT, ],
                                                    VID, NAT-D, NAT-D,
        UDP(4500,4500) HDR*#, [CERT, ],
                       NAT-D, NAT-D,
                       SIG_I         -->

                                             <-- UDP(4500, Y) HDR*#, ...

While floating, the port in the ID payload in Main Mode/Aggressive Mode
MUST be 0.

The most common case for the responder behind the NAT is if the NAT is
simply doing 1-1 address translation.  In this case, the initiator still
floats both ports to 4500.  The responder uses the identical algorithm

T. Kivinen, et. al.                                             [page 6]

INTERNET-DRAFT                                              24 June 2002

as above, although in this case, Y will equal 4500, since no port
translation is happening.

A different floating case involves out-of-band discovery of the ports to
use. For instance, if the responder is behind a port translating NAT,
and the initiator needs to contact it first, then the initiator will
need to determine which ports to use, usually by contacting some other
server. Once the initiator knows which ports to use to traverse the NAT,
generally something like UDP(Z,4500), he initiates using these ports.
This is similar to the responder rekey case above in that the ports to
use are already known upfront, and no additional floating need take

Also the first keepalive timer starts after floating to new port, no
keepalives are sent to the port 500 mapping.

5.  Quick Mode

After the Phase 1 both ends know if there is a NAT present between.  The
final decision of using the NAT-Traversal is left to the quick mode. The
use of NAT-Traversal is negotiated inside the SA payloads of the quick
mode. In the quick mode both ends can also send the original source
addresses of the IPsec packets (in case of the transport mode) to the
other, end so the other end has possibility to fix the TCP/IP checksum
field after the NAT transform.

This sending of the original source address is optional, and it is not
useful in the UDP-Encapsulated-Tunnel mode, as there is going to be
proper IP header inside the UDP-Encapsulated packet. In case of only
UDP-Encapsulated-Tunnel mode is negotiation then both ends SHOULD NOT
send original source address.
It also might be unnecessary in the transport mode if the other end can
turn off TCP/IP checksum verification. If the sending end knows (for
example from the vendor id payload) that the other end can turn off
TCP/IP checksum verification, he MAY leave the original source address
payload away. Otherwise he SHOULD send the original source address.

5.1.  Negotiation of the NAT-Traversal encapsulation

The negotiation of the NAT-Traversal happens by adding two new
encapsulation modes. These encapsulation modes are:

UDP-Encapsulated-Tunnel         61443 (XXX CHANGE)
UDP-Encapsulated-Transport      61444 (XXX CHANGE)

It is not normally useful to propose both normal tunnel or transport
mode and UDP-Encapsulated modes. If there is a NAT box between normal
tunnel or transport encapsulations may not work, and if there is no NAT
box between, there is no point of wasting bandwidth by adding UDP
encapsulation of packets. Because of this initiator SHOULD NOT include
both normal tunnel or transport mode and UDP-Encapsulated-Tunnel or UDP-
Encapsulated-Transport in its proposals.

T. Kivinen, et. al.                                             [page 7]

INTERNET-DRAFT                                              24 June 2002

5.2.  Sending the original source address

In case of transport mode both ends SHOULD send the original source
address to the other end. For the tunnel mode both ends SHOULD NOT send
original source address to the other end.

The original source address of packets put to this transport mode IPsec
SA is sent to other end using NAT-OA (NAT Original Address) payload.

The NAT-OA payloads are sent inside the first and second packets of the
quick mode. The initiator SHOULD send the payload if it proposes any
UDP-Encapsulated-Transport mode and the responder SHOULD send the
payload only if it selected UDP-Encapsulated-Transport mode. I.e it is
possible that initiator send the NAT-OA payload, but proposes both UDP-
Encapsulated transport and tunnel mode, and then the responder selects
the UDP-Encapsulated tunnel mode and do not send NAT-OA payload back.

A peer MUST NOT fail a negotiation if it does not receive a NAT-OA
payload if the NAT-OA payload only would contain redundant information.
I.e. only the machine(s) that are actually behind the NAT need to send
the NAT-OA payload. A machine with a public, non-changing IP address
doesn't need to send the NAT-OA payload.

The format of the NAT-OA packet is

      1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
     | Next Payload  |    RESERVED   |        Payload length         |
     |    ID Type    |    RESERVED   |           RESERVED            |
     |         IPv4 (4 octets) or IPv6 address (16 octets)           |

The payload type for the NAT original address payload is 131 (XXX

The ID type is defined in the [RFC-2407]. Only ID_IPV4_ADDR and
ID_IPV6_ADDR types are allowed. The two reserved fields after the ID
Type must be zero.

An example of quick mode using NAT-OA payloads is:

         Initiator                       Responder
        ------------                    ------------
        HDR*, HASH(1), SA, Ni, [, KE]
          [, IDci, IDcr ] [, NAT-OA] -->
                                     <-- HDR*, HASH(2), SA, Nr, [, KE]
                                          [, IDci, IDcr ] [, NAT-OA]
        HDR*, HASH(3)

6.  Initial contact notifications

T. Kivinen, et. al.                                             [page 8]

INTERNET-DRAFT                                              24 June 2002

The source IP and port address of the INITIAL-CONTACT notification for
the host behind NAT are not meaningful, so the IP and port numbers MUST
NOT be used for the determine which IKE/IPsec SAs to remove. The ID
payload sent from the other SHOULD be used instead. I.e when INITIAL-
CONTACT notification is received from the other end, the receiving end
SHOULD remove all the SAs associated with the same ID payload.

7.  Recovering from the expiring NAT mappings

There are cases where NAT box decides to remove mappings that are still
alive (for example, the keepalive interval is too long, or the NAT box
is rebooted). To recover from those ends which are NOT behind NAT SHOULD
use the last valid authenticated packet from the other end to determine
which IP and port addresses the should be used. The host behind dynamic
NAT MUST NOT do this as otherwise it opens DoS attack possibility, and
there is no need for that, because the IP address or port of other host
will not change (it is not behind NAT).
Keepalives cannot be used for this purposes as they are not
authenticated, but any IKE authenticated IKE packet or ESP packet can be
used to notice that the IP address or the port has changed.

8.  Security Considerations

Whenever changes to some fundamental parts of a security protocol are
proposed, the examination of security implications cannot be skipped.
Therefore, here are some observations on the effects, and whether or not
these effects matter.

o  IKE probe reveals NAT-Traversal support to everyone. This should not
   be an issue.

o  The value of authentication mechanisms based on IP addresses
   disappears once NATs are in the picture. That is not necessarily a
   bad thing (for any real security, other authentication measures than
   IP addresses should be used). This means that pre-shared-keys
   authentication cannot be used with the main mode without group shared
   keys for everybody behind the NAT box, which is huge security risk.
   Use of group shared keys is NOT RECOMMENDED.

o  As the internal address space is only 32 bits, and it is usually very
   sparse, it might be possible for the attacker to find out the
   internal address used behind the NAT box by trying all possible IP-
   addresses and trying to find the matching hash. The port numbers are
   normally fixed to 500, and the cookies can be extracted from the
   packet. This limits the hash calculations down to 2^32. If educated
   guess of use of private address space is done, then the number of
   hash calculations needed to find out the internal IP address goes
   down to the 2^24 + 2 * (2^16).

o  Neither NAT-D payloads or Vendor ID payloads are authenticated at all
   in the main mode nor in the aggressive mode. This means that attacker
   can remove those payloads, modify them or add them. By removing or
   adding them the attacker can cause Denial Of Service attacks. By

T. Kivinen, et. al.                                             [page 9]

INTERNET-DRAFT                                              24 June 2002

   modifying the NAT-D packets the attacker can cause both ends to use
   UDP-Encapsulated modes instead of directly using tunnel or transport
   mode, thus wasting some bandwidth.

o  The sending of the original source address in the Quick Mode reveals
   the internal IP address behind the NAT to the other end. In this case
   we have already authenticated the other end, and sending of the
   original source address is only needed in transport mode.

o  Updating the IKE SA / ESP UDP encapsulation IP addresses and ports
   for each valid authenticated packet can cause DoS in case we have
   attacker who can listen all traffic in the network, and can change
   the order of the packet and inject new packets before the packet he
   has already seen. I.e attacker can take the authenticated packet from
   the host behind NAT, change the packet UDP source or destination
   ports or IP addresses and sent it out to the other end before the
   real packet reaches there. The host not behind the NAT will update
   its IP address and port mapping and sends further traffic to wrong
   host or port. This situation is fixed immediately when the attacker
   stops modifying the packets as the first real packet will fix the
   situation back to normal. Implementations MAY print out warning every
   time the mapping is changed, as in normal case it should not happen
   that often.

9.  Intellectual property rights

The IETF has been notified of intellectual property rights claimed in
regard to some or all of the specification contained in this document.
For more information consult the online list of claimed rights.

SSH Communications Security Corp has notified the working group of one
or more patents or patent applications that may be relevant to this
internet-draft. SSH Communications Security Corp has already given a
license for those patents to the IETF. For more information consult the
online list of claimed rights.

10.  Acknowledgments

Thanks to Markus Stenberg, Larry DiBurro and William Dixon who
contributed actively to this document.

Thanks to Tatu Ylonen, Santeri Paavolainen, and Joern Sierwald who
contributed to the drafts used as base for this document.

11.  References

[RFC-2409] Harkins D., Carrel D., "The Internet Key Exchange (IKE)",
November 1998

[RFC-2407] Piper D., "The Internet IP Security Domain Of Interpretation
for ISAKMP", November 1998

[RFC-2119] Bradner, S., "Key words for use in RFCs to indicate

T. Kivinen, et. al.                                            [page 10]

INTERNET-DRAFT                                              24 June 2002

Requirement Levels", March 1997

[Hutt02] Huttunen, A. et. al., "UDP Encapsulation of IPsec Packets",
draft-ietf-ipsec-udp-encaps-03.txt, June 2002

[Dixon01] Dixon, W. et. al., "IPSec over NAT Justification for UDP
Encapsulation", draft-ietf-ipsec-udp-encaps-justification-00.txt, June

12.  Authors' Addresses

    Tero Kivinen
    SSH Communications Security Corp
    Fredrikinkatu 42
    FIN-00100 HELSINKI
    E-mail: kivinen@ssh.fi

    Ari Huttunen
    F-Secure Corporation
    Tammasaarenkatu 7,
    FIN-00181 HELSINKI
    E-mail: Ari.Huttunen@F-Secure.com

    Brian Swander
    One Microsoft Way
    Redmond WA 98052
    E-mail: briansw@microsoft.com

    Victor Volpe
    Cisco Systems
    124 Grove Street
    Suite 205
    Franklin, MA 02038
    E-mail: vvolpe@cisco.com

T. Kivinen, et. al.                                            [page 11]

SSH Communications Security                  http://www.ssh.fi/
SSH IPSEC Toolkit                            http://www.ssh.fi/ipsec/