RFC 4430

Network Working Group                                          S. Sakane
Request for Comments: 4430                                     K. Kamada
Category: Standards Track                        Yokogawa Electric Corp.
                                                               M. Thomas
                                                             J. Vilhuber
                                                           Cisco Systems
                                                              March 2006

             Kerberized Internet Negotiation of Keys (KINK)

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

   Copyright (C) The Internet Society (2006).


   This document describes the Kerberized Internet Negotiation of Keys
   (KINK) protocol.  KINK defines a low-latency, computationally
   inexpensive, easily managed, and cryptographically sound protocol to
   establish and maintain security associations using the Kerberos
   authentication system.  KINK reuses the Quick Mode payloads of the
   Internet Key Exchange (IKE), which should lead to substantial reuse
   of existing IKE implementations.

Table of Contents

   1. Introduction ....................................................3
      1.1. Conventions Used in This Document ..........................3
   2. Protocol Overview ...............................................4
   3. Message Flows ...................................................4
      3.1. GETTGT Message Flow ........................................5
      3.2. CREATE Message Flow ........................................6
           3.2.1. CREATE Key Derivation Considerations ................7
      3.3. DELETE Message Flow ........................................8
      3.4. STATUS Message Flow ........................................9
      3.5. Reporting Errors ...........................................9
      3.6. Rekeying Security Associations ............................10
      3.7. Dead Peer Detection .......................................10
           3.7.1. Coping with Dead User-to-User Peers ................12

Sakane, et al.              Standards Track                     [Page 1]

RFC 4430                          KINK                        March 2006

   4. KINK Message Format ............................................13
      4.1. KINK Alignment Rules ......................................15
      4.2. KINK Payloads .............................................16
           4.2.1. KINK_AP_REQ Payload ................................17
           4.2.2. KINK_AP_REP Payload ................................18
           4.2.3. KINK_KRB_ERROR Payload .............................19
           4.2.4. KINK_TGT_REQ Payload ...............................20
           4.2.5. KINK_TGT_REP Payload ...............................21
           4.2.6. KINK_ISAKMP Payload ................................21
           4.2.7. KINK_ENCRYPT Payload ...............................22
           4.2.8. KINK_ERROR Payload .................................23
   5. Differences from IKE Quick Mode ................................25
      5.1. Security Association Payloads .............................26
      5.2. Proposal and Transform Payloads ...........................26
      5.3. Identification Payloads ...................................26
      5.4. Nonce Payloads ............................................26
      5.5. Notify Payloads ...........................................27
      5.6. Delete Payloads ...........................................28
      5.7. KE Payloads ...............................................28
   6. Message Construction and Constraints for IPsec DOI .............28
      6.1. REPLY Message .............................................28
      6.2. ACK Message ...............................................28
      6.3. CREATE Message ............................................29
      6.4. DELETE Message ............................................30
      6.5. STATUS Message ............................................31
      6.6. GETTGT Message ............................................32
   7. ISAKMP Key Derivation ..........................................32
   8. Key Usage Numbers for Kerberos Key Derivation ..................33
   9. Transport Considerations .......................................33
   10. Security Considerations .......................................34
   11. IANA Considerations ...........................................35
   12. Forward Compatibility Considerations ..........................35
      12.1. New Versions of Quick Mode ...............................36
      12.2. New DOI ..................................................36
   13. Related Work ..................................................36
   14. Acknowledgements ..............................................37
   15. References ....................................................37
      15.1. Normative References .....................................37
      15.2. Informative References ...................................38

Sakane, et al.              Standards Track                     [Page 2]

RFC 4430                          KINK                        March 2006

1.  Introduction

   KINK is designed to provide a secure, scalable mechanism for
   establishing keys between communicating entities within a centrally
   managed environment in which it is important to maintain consistent
   security policy.  The security goals of KINK are to provide privacy,
   authentication, and replay protection of key management messages and
   to avoid denial of service vulnerabilities whenever possible.  The
   performance goals of the protocol are to have a low computational
   cost, low latency, and a small footprint.  It is also to avoid or
   minimize the use of public key operations.  In particular, the
   protocol provides the capability to establish IPsec security
   associations (SAs) in two messages with minimal computational effort.
   These requirements are described in RFC 3129 [REQ4KINK].

   Kerberos [KERBEROS] provides an efficient authentication mechanism
   for clients and servers using a trusted third-party model.  Kerberos
   also provides a mechanism for cross-realm authentication natively.  A
   client obtains a ticket from an online authentication server, the Key
   Distribution Center (KDC).  The ticket is then used to construct a
   credential for authenticating the client to the server.  As a result
   of this authentication operation, the server will also share a secret
   key with the client.  KINK uses this property as the basis of
   distributing keys for IPsec.

   The central key management provided by Kerberos is efficient because
   it limits computational cost and limits complexity versus IKE's
   necessity of using public key cryptography [IKE].  Initial
   authentication to the KDC may be performed using either symmetric
   keys, or asymmetric keys using the Public Key Cryptography for
   Initial Authentication in Kerberos [PKINIT]; however, subsequent
   requests for tickets as well as authenticated exchanges between the
   client and servers always utilize symmetric cryptography.  Therefore,
   public key operations (if any) are limited and are amortized over the
   lifetime of the credentials acquired in the initial authentication
   operation to the KDC.  For example, a client may use a single public
   key exchange with the KDC to efficiently establish multiple SAs with
   many other servers in the realm of the KDC.  Kerberos also scales
   better than direct peer-to-peer keying when symmetric keys are used.
   The reason is that since the keys are stored in the KDC, the number
   of principal keys is O(n+m) rather than O(n*m), where "n" is the
   number of clients and "m" is the number of servers.

1.1.  Conventions Used in This Document

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in [RFC2119].

Sakane, et al.              Standards Track                     [Page 3]

RFC 4430                          KINK                        March 2006

   It is assumed that the readers are familiar with the terms and
   concepts described in Kerberos Version 5 [KERBEROS], IPsec [IPSEC],
   and IKE [IKE].

2.  Protocol Overview

   KINK is a command/response protocol that can create, delete, and
   maintain IPsec SAs.  Each command or response contains a common
   header along with a set of type-length-value payloads.  The type of a
   command or a response constrains the payloads sent in the messages of
   the exchange.  KINK itself is a stateless protocol in that each
   command or response does not require storage of hard state for KINK.
   This is in contrast to IKE, which uses Main Mode to first establish
   an Internet Security Association and Key Management Protocol (ISAKMP)
   SA followed by subsequent Quick Mode exchanges.

   KINK uses Kerberos mechanisms to provide mutual authentication and
   replay protection.  For establishing SAs, KINK provides
   confidentiality for the payloads that follow the Kerberos AP-REQ
   payload.  The design of KINK mitigates denial of service attacks by
   requiring authenticated exchanges before the use of any public key
   operations and the installation of any state.  KINK also provides a
   means of using Kerberos User-to-User mechanisms when there is not a
   key shared between the server and the KDC.  This is typically, but
   not limited to, the case with IPsec peers using PKINIT for initial

   KINK directly reuses Quick Mode payloads defined in section 5.5 of
   [IKE], with some minor changes and omissions.  In most cases, KINK
   exchanges are a single command and its response.  An optional third
   message is required when creating SAs, only if the responder rejects
   the first proposal from the initiator or wants to contribute the
   keying materials.  KINK also provides rekeying and dead peer

3.  Message Flows

   All KINK message flows follow the same pattern between the two peers:
   a command, a response, and an optional acknowledgement in a CREATE
   flow.  A command is a GETTGT, CREATE, DELETE, or STATUS message; a
   response is a REPLY message; and an acknowledgement is an ACK

   KINK uses Kerberos as the authentication mechanism; therefore, a KINK
   host needs to get a service ticket for each peer before actual key
   negotiations.  This is basically a pure Kerberos exchange and the
   actual KDC traffic here is for illustrative purposes only.  In
   practice, when a principal obtains various tickets is a subject of

Sakane, et al.              Standards Track                     [Page 4]

RFC 4430                          KINK                        March 2006

   Kerberos and local policy consideration.  As an exception, the GETTGT
   message flow of KINK (described in section 3.1) is used when a User-
   to-User authentication is required.  In this flow, we assume that
   both A and B have ticket-granting tickets (TGTs) from their KDCs.

   After a service ticket is obtained, KINK uses the CREATE message flow
   (section 3.2), DELETE message flow (section 3.3), and STATUS message
   flow (section 3.4) to manage SAs.  In these flows, we assume that A
   has a service ticket for B.

3.1.  GETTGT Message Flow

   This flow is used to retrieve a TGT from the remote peer in User-to-
   User authentication mode.

   If the initiator determines that it will not be able to get a normal
   (non-User-to-User) service ticket for the responder, it can try a
   User-to-User authentication.  In this case, it first fetches a TGT
   from the responder in order to get a User-to-User service ticket:

       A                        B                       KDC
     ------                  ------                     ---
    1  GETTGT+KINK_TGT_REQ------>

    2  <-------REPLY+KINK_TGT_REP

    3  TGS-REQ+TGT(B)------------------------------------>

    4  <-------------------------------------------TGS-REP

                       Figure 1: GETTGT Message Flow

   The initiator MAY support the following events as triggers to go to
   the User-to-User path.  Note that the two errors described below will
   not be authenticated, and how to act on them depends on the policy.

        o    The local policy says that the responder requires a User-
             to-User authentication.

        o    A KRB_AP_ERR_USER_TO_USER_REQUIRED error is returned from
             the responder.

        o    A KDC_ERR_MUST_USE_USER2USER error is returned from the

Sakane, et al.              Standards Track                     [Page 5]

RFC 4430                          KINK                        March 2006

3.2.  CREATE Message Flow

   This flow creates SAs.  The CREATE command takes an "optimistic"
   approach, where SAs are initially created on the expectation that the
   responder will choose the initial proposed payload.  The optimistic
   proposal is placed in the first transform payload(s) of the first
   proposal.  The initiator MUST check to see if the optimistic proposal
   was selected by comparing all transforms and attributes, which MUST
   be identical to those in the initiator's optimistic proposal with the
   exceptions of LIFE_KILOBYTES and LIFE_SECONDS.  Each of these
   attributes MAY be set to a lower value by the responder and still
   expect optimistic keying, but MUST NOT be set to a higher value that
   MUST generate a NO-PROPOSAL-CHOSEN error.  The initiator MUST use the
   shorter lifetime.

   When a CREATE command contains an existing Security Parameter Index
   (SPI), the responder MUST reject it and SHOULD return an ISAKMP
   notification with INVALID-SPI.

   When a key exchange (KE) payload is sent from the initiator but the
   responder does not support it, the responder MUST reject it with an
   ISAKMP notification of INVALID-PAYLOAD-TYPE containing a KE payload
   type as its notification data.  When the initiator receives this
   error, it MAY retry without a KE payload (as another transaction) if
   its policy allows that.

       A                        B                       KDC
     ------                  ------                     ---

       A creates an optimistic inbound SA (B->A) unless using a KE.

    1  CREATE+ISAKMP------------>

       B creates an inbound SA (A->B).
       B creates an outbound SA (B->A) if optimistic and not using a KE.

    2  <-------------REPLY+ISAKMP

       A creates an outbound SA (A->B).
       A replaces an inbound SA (B->A) if non-optimistic.
       A creates an inbound SA (B->A) if using a KE.

    3 [ ACK--------------------->                            ]

      [ B creates an outbound SA (B->A).                     ]

                       Figure 2: CREATE Message Flow

Sakane, et al.              Standards Track                     [Page 6]

RFC 4430                          KINK                        March 2006

   Creating SAs has two modes:  2-way handshake and 3-way handshake.
   The initiator usually begins a negotiation expecting a 2-way
   handshake.  When the optimistic proposal is not chosen by the
   responder, the negotiation is switched to a 3-way handshake.  When
   and only when the initiator uses a KE payload, 3-way handshake is
   expected from the beginning.

   A 2-way handshake is performed in the following steps:

      1) The host A creates an inbound SA (B->A) in its SA database
         using the optimistic proposal in the ISAKMP SA proposal.  It is
         then ready to receive any messages from B.
      2) A then sends the CREATE message to B.
      3) If B agrees to A's optimistic proposal, B creates an inbound SA
         (A->B) and an outbound SA (B->A) in its database.  If B does
         not choose the first proposal or wants to add a Nonce payload,
         switch to step 3 of the 3-way handshake described below.
      4) B then sends a REPLY to A without a Nonce payload and without
         requesting an ACK.
      5) Upon receipt of the REPLY, A creates an outbound SA (A->B).

   A 3-way handshake is performed in the following steps:

      1) The host A sends the CREATE message to B without creating any
      2) B chooses one proposal according to its policy.
      3) B creates an inbound SA (A->B) and sends the actual choice in
         the REPLY.  It SHOULD send the optional Nonce payload (as it
         does not increase message count and generally increases entropy
         sources) and MUST request that the REPLY be acknowledged.
      4) Upon receipt of the REPLY, A creates the inbound SA (B->A) (or
         modifies it as necessary, if switched from 2-way), and the
         outbound SA (A->B).
      5) A now sends the ACK message.
      6) Upon receipt of the ACK, B installs the final outbound SA

   If B does not choose the first proposal, adds a nonce, or accepts the
   KE exchange, then it MUST request an ACK (i.e., set the ACKREQ bit)
   so that it can install the final outbound SA.  The initiator MUST
   always generate an ACK if the ACKREQ bit is set in the KINK header,
   even if it believes that the responder was in error.

3.2.1.  CREATE Key Derivation Considerations

   The CREATE command's optimistic approach allows an SA to be created
   in two messages rather than three.  The implication of a two-message
   exchange is that B will not contribute to the key since A must set up

Sakane, et al.              Standards Track                     [Page 7]

RFC 4430                          KINK                        March 2006

   the inbound SA before it receives any additional keying material from
   B.  This may be suspect under normal circumstances; however, KINK
   takes advantage of the fact that the KDC provides a reliable source
   of randomness which is used in key derivation.  In many cases, this
   will provide an adequate session key so that B will not require an
   acknowledgement.  Since B is always at liberty to contribute to the
   keying material, this is strictly a trade-off between the key
   strength versus the number of messages, which KINK implementations
   may decide as a matter of policy.

3.3.  DELETE Message Flow

   The DELETE command deletes existing SAs.  The domain of
   interpretation (DOI)-specific payloads describe the actual SA to be
   deleted.  For the IPsec DOI, those payloads will include an ISAKMP
   payload containing the list of the SPIs to be deleted.

       A                        B                       KDC
     ------                  ------                     ---

       A deletes outbound SA to B.

    1  DELETE+ISAKMP------------>

       B deletes inbound and outbound SA to A.

    2  <-------------REPLY+ISAKMP

       A deletes inbound SA to B.

                       Figure 3: DELETE Message Flow

   The DELETE command takes a "pessimistic" approach, which does not
   delete inbound SAs until it receives acknowledgement that the other
   host has received the DELETE.  The exception to the pessimistic
   approach is if the initiator wants to immediately cease all activity
   on an inbound SA.  In this case, it MAY delete the inbound SA as well
   in step 1, above.

   The ISAKMP payload contains ISAKMP Delete payload(s) that indicate
   the inbound SA(s) for the initiator of this flow.  KINK does not
   allow half-open SAs; thus, when the responder receives a DELETE
   command, it MUST delete SAs of both directions, and MUST reply with
   ISAKMP Delete payload(s) that indicate the inbound SA(s) for the
   responder of this flow.  If the responder cannot find an appropriate
   SPI to be deleted, it MUST return an ISAKMP notification with
   INVALID_SPI, which also serves to inform the initiator that it can
   delete the inbound SA.

Sakane, et al.              Standards Track                     [Page 8]

RFC 4430                          KINK                        March 2006

   A race condition with the DELETE flow exists.  Due to network
   reordering, etc., packets in flight while the DELETE operation is
   taking place may arrive after the diagrams above, which recommend
   deleting the inbound SA.  A KINK implementation SHOULD implement a
   grace timer that SHOULD be set to a period of at least two times the
   average round-trip time, or to a configurable value.  A KINK
   implementation MAY choose to set the grace period to zero at
   appropriate times to delete an SA ungracefully.  The behavior
   described here is referred from the behavior of the TCP [RFC793]
   flags FIN and RST.

3.4.  STATUS Message Flow

   This flow is used to send any information to a peer or to elicit any
   information from a peer.  An initiator may send a STATUS command to
   the responder at any time, optionally with DOI-specific ISAKMP
   payloads.  In the case of the IPsec DOI, these are generally in the
   form of ISAKMP Notification payloads.  A STATUS command is also used
   as a means of dead peer detection described in section 3.7.

       A                        B                       KDC
     ------                  ------                     ---

    1  STATUS[+ISAKMP]---------->

    2  <-----------REPLY[+ISAKMP]

                       Figure 4: STATUS Message Flow

3.5.  Reporting Errors

   When the responder detects an error in a received command, it can
   send a DOI-specific payload to indicate the error in a REPLY message.
   There are three types of payloads that can indicate errors:
   KINK_KRB_ERROR payloads for Kerberos errors, KINK_ERROR payloads for
   KINK errors, and KINK_ISAKMP payloads for ISAKMP errors.  Details are
   described in sections 4.2.3, 4.2.8, and 4.2.6, respectively.

   If the initiator detects an error in a received reply, there is no
   means to report it back to the responder.  The initiator SHOULD log
   the event and MAY take a remedial action by reinitiating the initial

   If the server clock and the client clock are off by more than the
   policy-determined clock skew limit (usually 5 minutes), the server
   MUST return a KRB_AP_ERR_SKEW.  The optional client's time in the
   KRB-ERROR SHOULD be filled out.  If the server protects the error by
   adding the Cksum field and returning the correct client's time, the

Sakane, et al.              Standards Track                     [Page 9]

RFC 4430                          KINK                        March 2006

   client SHOULD compute the difference (in seconds) between the two
   clocks based upon the client and server time contained in the
   KRB-ERROR message.  The client SHOULD store this clock difference and
   use it to adjust its clock in subsequent messages.  If the error is
   not protected, the client MUST NOT use the difference to adjust
   subsequent messages, because doing so would allow an attacker to
   construct authenticators that can be used to mount replay attacks.

3.6.  Rekeying Security Associations

   KINK expects the initiator of an SA to be responsible for rekeying
   the SA for two reasons.  The first reason is to prevent needless
   duplication of SAs as the result of collisions due to an initiator
   and responder both trying to renew an existing SA.  The second reason
   is due to the client/server nature of Kerberos exchanges, which
   expects the client to get and maintain tickets.  While KINK expects
   that a KINK host is able to get and maintain tickets, in practice it
   is often advantageous for servers to wait for clients to initiate
   sessions so that they do not need to maintain a large ticket cache.

   There are no special semantics for rekeying SAs in KINK.  That is, in
   order to rekey an existing SA, the initiator must CREATE a new SA
   followed by either deleting the old SA with the DELETE flow or
   letting it time out.  When identical flow selectors are available on
   different SAs, KINK implementations SHOULD choose the SA most
   recently created.  It should be noted that KINK avoids most of the
   problems of [IKE] rekeying by having a reliable delete mechanism.

   Normally, a KINK implementation that rekeys existing SAs will try to
   rekey the SA ahead of an SA termination, which may include the hard
   lifetime in time/bytecount or the overflow of the sequence number
   counter.  We call this time "soft lifetime".  The soft lifetime MUST
   be randomized to avoid synchronization with similar implementations.
   In the case of the lifetime in time, one reasonable approach to
   determine the soft lifetime is picking a random time between T-rekey
   and T-retrans and subtracting it from the hard lifetime.  Here,
   T-rekey is the reasonable maximum rekeying margin, and T-retrans is
   the amount of time it would take to go through a full retransmission
   cycle.  T-rekey SHOULD be at least twice as high as T-retrans.

3.7.  Dead Peer Detection

   In order to determine that a KINK peer has lost its security database
   information, KINK peers MUST record the current epoch for which they
   have valid SA information for a peer and reflect that epoch in each
   AP-REQ and AP-REP message.  When a KINK peer creates state for a
   given SA, it MUST also record the principal's epoch.  If it discovers

Sakane, et al.              Standards Track                    [Page 10]

RFC 4430                          KINK                        March 2006

   on a subsequent message that the principal's epoch has changed, it
   MUST consider all SAs created by that principal as invalid, and take
   some action such as tearing those SAs down.

   While a KINK peer SHOULD use feedback from routing (in the form of
   ICMP messages) as a trigger to check whether or not the peer is still
   alive, a KINK peer MUST NOT conclude the peer is dead simply based on
   unprotected routing information (said ICMP messages).

   If there is suspicion that a peer may be dead (based on any
   information available to the KINK peer, including lack of IPsec
   traffic, etc.), the KINK STATUS message SHOULD be used to coerce an
   acknowledgement out of the peer.  Since nothing is negotiated about
   dead peer detection in KINK, each peer can decide its own metric for
   "suspicion" and also what timeouts to use before declaring a peer
   dead due to lack of response to the STATUS message.  This is
   desirable, and does not break interoperability.

   The STATUS message has a twofold effect.  First, it elicits a
   cryptographically secured (and replay-protected) response from the
   peer, which tells us whether or not the peer is reachable/alive.
   Second, it carries the epoch number of the peer, so we know whether
   or not the peer has rebooted and lost all state.  This is crucial to
   the KINK protocol: In IKE, if a peer reboots, we lose all
   cryptographic context, and no cryptographically secure communication
   is possible without renegotiating keys.  In KINK, due to Kerberos
   tickets, we can communicate securely with a peer, even if the peer
   rebooted, as the shared cryptographic key used is carried in the
   Kerberos ticket.  Thus, active cryptographic communication is not an
   indication that the peer has not rebooted and lost all state, and the
   epoch is needed.

   Assume a Peer A sending a STATUS and a peer B sending the REPLY (see
   section 3.4).  Peer B MAY assume that the sender is alive, and the
   epoch in the STATUS message will indicate whether or not the peer A
   has lost state.  Peer B MUST acknowledge the STATUS message with a
   REPLY message, as described in section 3.4.

   The REPLY message will indicate to peer A that the peer is alive, and
   the epoch in the REPLY will indicate whether peer B has lost its
   state or not.  If peer A does not receive a REPLY message from peer B
   in a suitable timeout, peer A MAY send another STATUS message.  It is
   up to peer A to decide how aggressively to declare peer B dead.  The
   level of aggressiveness may depend on many factors such as rapid fail
   over versus number of messages sent by nodes with large numbers of

Sakane, et al.              Standards Track                    [Page 11]

RFC 4430                          KINK                        March 2006

   Note that peer B MUST NOT make any inferences about a lack of STATUS
   message from peer A.  Peer B MAY use a STATUS message from peer A as
   an indication of A's aliveness, but peer B MUST NOT expect another
   STATUS message at any time (i.e., dead peer detection is not periodic

   Strategies for sending STATUS messages are the following: Peer A may
   decide to send a STATUS message only after a prolonged period where
   no traffic was sent in either direction over the IPsec SAs with the
   peer.  Once there is traffic, peer A may want to know if the traffic
   is going into a black hole, and send a STATUS message.
   Alternatively, peer A may use an idle timer to detect lack of traffic
   with the peer, and send STATUS messages in the quiet phase to make
   sure the peer is still alive for when traffic needs to finally be

3.7.1.  Coping with Dead User-to-User Peers

   When an initiator uses a User-to-User ticket and a responder has lost
   its previous TGT, the usual dead peer detection (DPD) mechanism does
   not work, because the responder cannot decrypt the ticket with its
   new TGT.  In this case, the following actions are taken.

     o    When the responder receives a KINK command with a User-to-User
          ticket that cannot be decrypted with its TGT, it returns a
          REPLY with a KINK_TGT_REP payload containing the TGT.

     o    When the initiator receives a KINK_TGT_REP, it retrieves a new
          service ticket with the TGT and retries the command.

   This does not directly define a method to detect a dead User-to-User
   peer, but to recover from the situation that the responder does not
   have an appropriate TGT to decrypt a service ticket sent from the
   initiator.  After recovery, they can exchange their epochs, and usual
   DPD mechanism will detect a dead peer if it really has been dead.

   The initiator MUST NOT think the peer has been dead on the receipt of
   a KINK_TGT_REP because of two reasons.  One is that the message is
   not authenticated, and the other is that losing a TGT does not
   necessarily mean losing the SA database information.  The initiator
   SHOULD NOT forget the previous service ticket until the new one is
   successfully obtained in order to reduce the cost when a forged
   KINK_TGT_REP is received.

Sakane, et al.              Standards Track                    [Page 12]

RFC 4430                          KINK                        March 2006

4.  KINK Message Format

   All values in KINK are formatted in network byte order (most
   significant byte first).  The RESERVED fields MUST be set to zero (0)
   when a packet is sent.  The receiver MUST ignore these fields.

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    |   Type        | MjVer |RESRVED|            Length             |
    |                 Domain of Interpretation (DOI)                |
    |                      Transaction ID (XID)                     |
    |  NextPayload  |A|  RESERVED2  |           CksumLen            |
    |                                                               |
    ~                      A series of payloads                     ~
    |                                                               |
    |                                                               |
    ~                       Cksum (variable)                        ~
    |                                                               |

                    Figure 5:  Format of a KINK Message


     o    Type (1 octet) -- The type of this message.

              Type              Value
              -----             -----
              RESERVED            0
              CREATE              1
              DELETE              2
              REPLY               3
              GETTGT              4
              ACK                 5
              STATUS              6
              RESERVED TO IANA    7 - 127
              Private Use       128 - 255

     o    MjVer (4 bits) -- Major protocol version number.  This MUST be
          set to 1.

Sakane, et al.              Standards Track                    [Page 13]

RFC 4430                          KINK                        March 2006

     o    RESRVED (4 bits) -- Reserved and MUST be zero when sent, MUST
          be ignored when received.

     o    Length (2 octets) -- Length of the message in octets.  It is
          not forbidden in KINK that there are unnecessary data after
          the message, but the Length field MUST represent the actual
          length of the message.

     o    DOI (4 octets) -- The domain of interpretation.  All DOIs must
          be registered with the IANA in the ISAKMP Domain of
          Interpretation section of the isakmp-registry [ISAKMP-REG].
          The IANA Assigned Number for the Internet IP Security DOI
          [IPDOI] is one (1).  This field defines the context of all
          sub-payloads in this message.  If sub-payloads have a DOI
          field (e.g., Security Association Payload), then the DOI in
          that sub-payload MUST be checked against the DOI in this
          header, and the values MUST be the same.

     o    XID (4 octets) -- The transaction ID.  A KINK transaction is
          bound together by a transaction ID, which is created by the
          command initiator and replicated in subsequent messages in the
          transaction.  A transaction is defined as a command, a reply,
          and an optional acknowledgement.  Transaction IDs are used by
          the initiator to discriminate between multiple outstanding
          requests to a responder.  It is not used for replay protection
          because that functionality is provided by Kerberos.  The value
          of XID is chosen by the initiator and MUST be unique with all
          outstanding transactions.  XIDs MAY be constructed by using a
          monotonic counter or random number generator.

     o    NextPayload (1 octet) -- Indicates the type of the first
          payload after the message header.

     o    A, or ACKREQ (1 bit) -- ACK Request.  Set to one if the
          responder requires an explicit acknowledgement that a REPLY
          was received.  An initiator MUST NOT set this flag, nor should
          a responder except for a REPLY to a CREATE when the optimistic
          proposal is chosen.

     o    RESERVED2 (7 bits) -- Reserved and MUST be zero on send, MUST
          be ignored by a receiver.

     o    CksumLen (2 octets) -- CksumLen is the length in octets of the
          cryptographic checksum of the message.  A CksumLen of zero
          implies that the message is unauthenticated.

Sakane, et al.              Standards Track                    [Page 14]

RFC 4430                          KINK                        March 2006

     o    Cksum (variable) -- Kerberos keyed checksum over the entire
          message excluding the Cksum field itself.  When any padding
          bytes are required between the last payload and the Cksum
          field, they MUST be included in the calculation.  This field
          MUST always be present whenever a key is available via an
          AP-REQ or AP-REP payload.  The key used MUST be the session
          key in the ticket.  When a key is not available, this field is
          not present, and the CksumLen field is set to zero.  The
          content of this field is the output of the Kerberos 5 get_mic
          function [KCRYPTO].  The get_mic function used is specified by
          a checksum type, which is a "required checksum mechanism" of
          the etype for the Kerberos session key in the Kerberos ticket.
          If the checksum type is not a keyed algorithm, the message
          MUST be rejected.

          To compute the checksum, the CksumLen field is zeroed out and
          the Length field is filled with the total packet length
          without the checksum.  Then, the packet is passed to the
          get_mic function and its output is appended to the packet.
          Any KINK padding after the Cksum field is not allowed, except
          the Kerberos internal one, which may be included in the output
          of the get_mic function.  Finally, the CksumLen field is
          filled with the checksum length and the Length field is filled
          with the total packet length including the checksum.

          To verify the checksum, a length-without-checksum is
          calculated from the value of Length field, subtracting the
          CksumLen.  The Length field is filled with the length-
          without-checksum value and the CksumLen field is zeroed out.
          Then, the packet without checksum (offset from 0 to length-
          without-checksum minus 1 of the received packet) and the
          checksum (offset from length-without-checksum to the last) are
          passed to the verify_mic function.  If verification fails, the
          message MUST be dropped.

   The KINK header is followed immediately by a series of
   Type/Length/Value fields, defined in section 4.2.

4.1.  KINK Alignment Rules

   KINK has the following rules regarding alignment and padding:

     o    All length fields MUST reflect the actual number of octets in
          the structure; i.e., they do not account for padding bytes
          required by KINK alignments.

     o    KINK headers, payloads, and the Cksum field MUST be aligned on
          4-octet boundaries.

Sakane, et al.              Standards Track                    [Page 15]

RFC 4430                          KINK                        March 2006

     o    Variable length fields (except the Cksum field) MUST always
          start immediately after the last octet of the previous field.
          That is, they are not aligned to 4-octet boundaries.

4.2.  KINK Payloads

   Immediately following the header, there is a list of
   Type/Length/Value (TLV) payloads.  There can be any number of
   payloads following the header.  Each payload MUST begin with a
   payload header.  Each payload header is built on the generic payload
   header.  Any data immediately follows the generic header.  Payloads
   are all implicitly aligned to 4-octet boundaries, though the payload
   length field MUST accurately reflect the actual number of octets in
   the payload.

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                      value (variable)                         |

                    Figure 6:  Format of a KINK Payload


     o    Next Payload (1 octet) -- The type of the next payload.

              NextPayload       Value
              ----              -----
              KINK_DONE           0
              KINK_AP_REQ         1
              KINK_AP_REP         2
              KINK_KRB_ERROR      3
              KINK_TGT_REQ        4
              KINK_TGT_REP        5
              KINK_ISAKMP         6
              KINK_ENCRYPT        7
              KINK_ERROR          8
              RESERVED TO IANA    9 - 127
              Private Use       128 - 255

          Next Payload type KINK_DONE denotes that the current payload
          is the final payload in the message.

     o    RESERVED (1 octet) -- Reserved and MUST be set to zero by a
          sender, MUST be ignored by a receiver.

Sakane, et al.              Standards Track                    [Page 16]

RFC 4430                          KINK                        March 2006

     o    Payload Length (2 octets) -- The length of this payload,
          including the type and length fields.

     o    Value (variable) -- This value of this field depends on the

4.2.1.  KINK_AP_REQ Payload

   The KINK_AP_REQ payload relays a Kerberos AP-REQ to the responder.
   The AP-REQ MUST request mutual authentication.

   This document does not specify how to generate the principal name.
   That is, complete principal names may be stored in local policy,
   Fully Qualified Domain Names (FQDNs) may be converted to principal
   names, IP addresses may be converted to principal names by secure
   name services, etc., but see the first paragraph of the Security
   Considerations section.

   If the peer's principal name for the KINK service is generated from
   an FQDN, the principal name, which the initiator starts from, will be
   "kink/fqdn@REALM"; where "kink" is a literal string for the KINK
   IPsec service, "fqdn" is the fully qualified domain name of the
   service host, and "REALM" is the Kerberos realm of the service.  A
   principal name is case sensitive, and "fqdn" part MUST be lowercase
   as described in [KERBEROS].

   The value field of this payload has the following format:

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                         EPOCH                                 |
    |                                                               |
    ~                        AP-REQ                                 ~
    |                                                               |

                      Figure 7:  KINK_AP_REQ Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

Sakane, et al.              Standards Track                    [Page 17]

RFC 4430                          KINK                        March 2006

     o    EPOCH -- The absolute time at which the creator of the AP-REQ
          has valid SA information.  Typically, this is when the KINK
          keying daemon started if it does not retain SA information
          across restarts.  The value in this field is the least
          significant 4 octets of so-called POSIX time, which is the
          elapsed seconds (but without counting leap seconds) from
          1970-01-01T00:00:00 UTC.  For example, 2038-01-19T03:14:07 UTC
          is represented as 0x7fffffff.

     o    AP-REQ -- The value field of this payload contains a raw
          Kerberos AP-REQ.

4.2.2.  KINK_AP_REP Payload

   The KINK_AP_REP payload relays a Kerberos AP-REP to the initiator.
   The AP-REP MUST be checked for freshness as described in [KERBEROS].

   The value field of this payload has the following format:

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                         EPOCH                                 |
    |                                                               |
    ~                        AP-REP                                 ~
    |                                                               |

                      Figure 8:  KINK_AP_REP Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    EPOCH -- The absolute time at which the creator of the AP-REP
          has valid SA information.  Typically, this is when the KINK
          keying daemon started if it does not retain SA information
          across restarts.  The value in this field is the least
          significant 4 octets of so-called POSIX time, which is the
          elapsed seconds (but without counting leap seconds) from
          1970-01-01T00:00:00 UTC.  For example, 2038-01-19T03:14:07 UTC
          is represented as 0x7fffffff.

Sakane, et al.              Standards Track                    [Page 18]

RFC 4430                          KINK                        March 2006

     o    AP-REP -- The value field of this payload contains a raw
          Kerberos AP-REP.

4.2.3.  KINK_KRB_ERROR Payload

   The KINK_KRB_ERROR payload relays Kerberos type errors back to the
   initiator.  The initiator MUST be prepared to receive any valid
   Kerberos error type [KERBEROS].

   KINK implementations SHOULD make use of a KINK Cksum field when
   returning KINK_KRB_ERROR and the appropriate service key is
   available.  Especially in the case of clock skew errors, protecting
   the error at the server creates a better user experience because it
   does not require clocks to be synchronized.  However, many Kerberos
   implementations do not make it easy to obtain the session key in
   order to protect error packets.  For unauthenticated Kerberos errors,
   the initiator MAY choose to act on them, but SHOULD take precautions
   against make-work kinds of attacks.

   Note that KINK does not make use of the text or e_data field of the
   Kerberos error message, though a compliant KINK implementation MUST
   be prepared to receive them and MAY log them.

   The value field of this payload has the following format:

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                                                               |
    ~                      KRB-ERROR                                ~
    |                                                               |

                     Figure 9:  KINK_KRB_ERROR Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    KRB-ERROR -- The value field of this payload contains a raw
          Kerberos KRB-ERROR.

Sakane, et al.              Standards Track                    [Page 19]

RFC 4430                          KINK                        March 2006

4.2.4.  KINK_TGT_REQ Payload

   The KINK_TGT_REQ payload provides a means to get a TGT from the peer
   in order to obtain a User-to-User service ticket from the KDC.

   The value field of this payload has the following format:

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                                                               |
    ~                     PrincName (variable)                      ~
    |                                                               |

                     Figure 10:  KINK_TGT_REQ Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    PrincName -- The name of the principal that the initiator
          wants to communicate with.  It is assumed that the initiator
          knows the responder's principal name (including the realm
          name) in the same way as the non-User-to-User case.  The TGT
          returned MUST NOT be an inter-realm TGT and its cname and
          crealm MUST match the requested principal name, so that the
          initiator can rendezvous with the responder at the responder's

          PrincName values are octet string representations of a
          principal and realm name formatted just like the octet string
          used in the "NAME" component of Generic Security Service
          Application Program Interface (GSS-API) [RFC2743] exported
          name token for the Kerberos V5 GSS-API mechanism [RFC1964].
          See RFC 1964, section 2.1.3.

   If the responder is not the requested principal and is unable to get
   a TGT for the name, it MAY return a KRB_AP_ERR_NOT_US.  If the
   administrative policy prohibits returning a TGT, it MAY return a

Sakane, et al.              Standards Track                    [Page 20]

RFC 4430                          KINK                        March 2006

4.2.5.  KINK_TGT_REP Payload

   The value field of this payload contains the TGT requested in a
   previous KINK_TGT_REQ payload of a GETTGT command.

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                                                               |
    ~                        TGT (variable)                         ~
    |                                                               |

                     Figure 11:  KINK_TGT_REP Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    TGT -- The Distinguished Encoding Rules (DER)-encoded TGT of
          the responder.

4.2.6.  KINK_ISAKMP Payload

   The value field of this payload has the following format:

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    | InnerNextPload| QMMaj | QMMin |            RESERVED           |
    |                Quick Mode Payloads (variable)                 |

                      Figure 12:  KINK_ISAKMP Payload


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    InnerNextPload -- First payload type of the inner series of
          ISAKMP payloads.

Sakane, et al.              Standards Track                    [Page 21]

RFC 4430                          KINK                        March 2006

     o    QMMaj -- The major version of the inner payloads.  MUST be set
          to 1.

     o    QMMin -- The minor version of the inner payloads.  MUST be set
          to 0.

   The KINK_ISAKMP payload encapsulates the IKE Quick Mode (phase 2)
   payloads to take the appropriate action dependent on the KINK
   command.  There may be any number of KINK_ISAKMP payloads within a
   single KINK message.  While [IKE] is somewhat fuzzy about whether
   multiple different SAs may be created within a single IKE message,
   KINK explicitly requires that a new ISAKMP header be used for each
   discrete SA operation.  In other words, a KINK implementation MUST
send multiple Quick Mode transactions within a single KINK_ISAKMP

   The purpose of the Quick Mode version is to allow backward
   compatibility with IKE and ISAKMP if there are subsequent revisions.
   At the present time, the Quick Mode major and minor versions are set
   to one and zero (1.0), respectively.  These versions do not
   correspond to the ISAKMP version in the ISAKMP header.  A compliant
   KINK implementation MUST support receipt of 1.0 payloads.  It MAY
   support subsequent versions (both sending and receiving), and SHOULD
   provide a means to resort back to Quick Mode version 1.0 if the KINK
   peer is unable to process future versions.  A compliant KINK
   implementation MUST NOT mix Quick Mode versions in any given

4.2.7.  KINK_ENCRYPT Payload

   The KINK_ENCRYPT payload encapsulates other KINK payloads and is
   encrypted using the session key and the algorithm specified by its
   etype.  This payload MUST be the final one in the outer payload chain
   of the message.  The KINK_ENCRYPT payload MUST be encrypted before
   the final KINK checksum is applied.

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    | InnerNextPload|                   RESERVED2                   |
    |                         Payload (variable)                    |

                     Figure 13:  KINK_ENCRYPT Payload

Sakane, et al.              Standards Track                    [Page 22]

RFC 4430                          KINK                        March 2006


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.  This payload is the last one in a
          message, and accordingly, the Next Payload field must be
          KINK_DONE (0).

     o    InnerNextPload -- First payload type of the inner series of
          encrypted KINK payloads.

     o    RESERVED2 -- Reserved and MUST be zero when sent, MUST be
          ignored when received.

   The coverage of the encrypted data begins at InnerNextPload so that
   the first payload's type is kept confidential.  Thus, the number of
   encrypted octets is PayloadLength - 4.

   The format of the encryption payload follows the normal Kerberos
   semantics.  Its content is the output of an encrypt function defined
   in the Encryption Algorithm Profile section of [KCRYPTO].  Parameters
   such as encrypt function itself, specific-key, and initial state are
   defined with the etype.  The encrypt function may have padding in
   itself and there may be some garbage data at the end of the decrypted
   plaintext.  A KINK implementation MUST be prepared to ignore such
   padding after the last sub-payload inside the KINK_ENCRYPT payload.
   Note that each encrypt function has its own integrity protection
   mechanism.  It is redundant with the checksum in the KINK header, but
   this is unavoidable because it is not always possible to remove the
   integrity protection part from the encrypt function.

4.2.8.  KINK_ERROR Payload

   The KINK_ERROR payload type provides a protocol-level mechanism of
   returning an error condition.  This payload should not be used for
   either Kerberos-generated errors or DOI-specific errors that have
   their own payloads defined.  The error code is in network order.

     0                   1                   2                   3
     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
    | Next Payload  |   RESERVED    |         Payload Length        |
    |                           ErrorCode                           |

                      Figure 14:  KINK_ERROR Payload

Sakane, et al.              Standards Track                    [Page 23]

RFC 4430                          KINK                        March 2006


     o    Next Payload, RESERVED, Payload Length -- Defined in the
          beginning of this section.

     o    ErrorCode -- One of the following values in the network byte

          ErrorCode          Value             Purpose
          ---------          -----       -------------------
          KINK_OK              0         No error detected
          KINK_PROTOERR        1         The message was malformed
          KINK_INVDOI          2         Invalid DOI
          KINK_INVMAJ          3         Invalid Major Version
          RESERVED             4
          KINK_INTERR          5         An unrecoverable internal error
          KINK_BADQMVERS       6         Unsupported Quick Mode Version
          KINK_U2UDENIED       7         Returning a TGT is prohibited
          RESERVED TO IANA     8 - 8191
          Private Use       8192 - 16383
          RESERVED         16384 -

   The responder MUST NOT return KINK_OK.  When received, the initiator
   MAY act as if the specific KINK_ERROR payload were not present.  If
   the initiator supports multiple Quick Mode versions or DOIs,
   KINK_BADQMVERS or KINK_INVDOI is received, and the Cksum is verified,
   then it MAY retry with another version or DOI.  A responder SHOULD
   return a KINK error with KINK_INVMAJ, when it receives an unsupported
   KINK version number in the header.  When KINK_U2UDENIED is received,
   the initiator MAY retry with the non-User-to-User mode (if it has not
   yet been tried).

   In general, the responder MAY choose to return these errors in reply
   to unauthenticated commands, but SHOULD take care to avoid being
   involved in denial of service attacks.  Similarly, the initiator MAY
   choose to act on unauthenticated errors, but SHOULD take care to
   avoid denial of service attacks.

Sakane, et al.              Standards Track                    [Page 24]

RFC 4430                          KINK                        March 2006

5.  Differences from IKE Quick Mode

   KINK directly uses ISAKMP payloads to negotiate SAs.  In particular,
   KINK uses IKE phase 2 payload types (aka Quick Mode).  In general,
   there should be very few changes necessary to an IKE implementation
   to establish the SAs, and unless there is a note to the contrary in
   the memo, all capabilities and requirements in [IKE] MUST be
   supported.  IKE phase 1 payloads MUST NOT be sent.

   Unlike IKE, KINK defines specific commands for creation, deletion,
   and status of SAs, mainly to facilitate predictable SA
   creation/deletion (see sections 3.2 and 3.3).  As such, KINK places
   certain restrictions on what payloads may be sent with which
   commands, and some additional restrictions and semantics of some of
   the payloads.  Implementors should refer to [IKE] and [ISAKMP] for
   the actual format and semantics.  If a particular IKE phase 2 payload
   is not mentioned here, it means that there are no differences in its

     o   The Security Association Payload header for IP is defined in
         section 4.6.1 of [IPDOI].  For this memo, the Domain of
         Interpretation MUST be set to 1 (IPsec) and the Situation
         bitmap MUST be set to 1 (SIT_IDENTITY_ONLY).  All other fields
         are omitted (because SIT_IDENTITY_ONLY is set).

     o   KINK also expands the semantics of IKE in that it defines an
         optimistic proposal for CREATE commands to allow SA creation to
         complete in two messages.

     o   IKE Quick Mode (phase 2) uses the hash algorithm used in main
         mode (phase 1) to generate the keying material.  For this
         purpose, KINK MUST use a pseudo-random function determined by
         the etype of the session key.

     o   KINK does not use the HASH payload at all.

     o   KINK allows the Nonce payload Nr to be optional to facilitate
         optimistic keying.

Sakane, et al.              Standards Track                    [Page 25]

RFC 4430                          KINK                        March 2006

5.1.  Security Association Payloads

   KINK supports the following SA attributes from [IPDOI]:

   class                     value           type
   SA Life Type                1               B
   SA Life Duration            2               V
   Encapsulation Mode          4               B
   Authentication Algorithm    5               B
   Key Length                  6               B
   Key Rounds                  7               B

   Refer to [IPDOI] for the actual definitions of these attributes.

5.2.  Proposal and Transform Payloads

   KINK directly uses the Proposal and Transform payloads with no
   differences.  KINK, however, places additional relevance to the first
   proposal and first transform of each conjugate for optimistic keying.

5.3.  Identification Payloads

   The Identification payload carries information that is used to
   identify the traffic that is to be protected by the SA that will be
   established.  KINK restricts the ID types, which are defined in
   section of [IPDOI], to the following values:

      ID Type                  Value
      -------                  -----
      ID_IPV4_ADDR               1
      ID_IPV4_ADDR_SUBNET        4
      ID_IPV6_ADDR               5
      ID_IPV6_ADDR_SUBNET        6
      ID_IPV4_ADDR_RANGE         7
      ID_IPV6_ADDR_RANGE         8

5.4.  Nonce Payloads

   The Nonce payload contains random data that MUST be used in key
   generation.  It MUST be sent by the initiating KINK peer, and MAY be
   sent by the responding KINK peer.  See section 7 for the discussion
   of its use in key generation.

Sakane, et al.              Standards Track                    [Page 26]

RFC 4430                          KINK                        March 2006

5.5.  Notify Payloads

   Notify payloads are used to transmit several informational data, such
   as error conditions and state transitions to a peer.  For example,
   notification information transmit can be error messages specifying
   why an SA could not be established.  It can also be status data that
   a process managing an SA database wishes to communicate with a peer

   Types in the range 0 - 16383 are intended for reporting errors
   [ISAKMP].  An implementation receiving a type in this range that it
   does not recognize in a response MUST assume that the corresponding
   request has failed entirely.  Unrecognized error types in a request
   and status types in a request or response MUST be ignored, and they
   SHOULD be logged.  Notify payloads with status types MAY be added to
   any message and MUST be ignored if not recognized.  They are intended
   to indicate capabilities, and as part of SA negotiation are used to
   negotiate non-cryptographic parameters.

   The table below lists the Notification messages and their
   corresponding values.  PAYLOAD-MALFORMED denotes some error types
   defined by [ISAKMP].  Hence INVALID-PROTOCOL-ID, for example, is not
   used in this document.  INVALID-MAJOR-VERSION and INVALID-MINOR-
   VERSION are not used because KINK_BADQMVERS is used to tell the
   initiator that the version of IKE is not supported.

   -----------------------------           -----
   INVALID-PAYLOAD-TYPE                      1

      Sent if the ISAKMP payload type is not recognized.  It is also
      sent when the KE payload is not supported by the responder.
      Notification Data MUST contains the one-octet payload type.

   INVALID-SPI                              11

      Sent if the responder has an SPI indicated by the initiator in
      case of CREATE flow, or if the responder does not have an SPI
      indicated by the initiator in case of DELETE flow.

   NO-PROPOSAL-CHOSEN                       14

      Sent if none of the proposals in the SA payload was

Sakane, et al.              Standards Track                    [Page 27]

RFC 4430                          KINK                        March 2006

   PAYLOAD-MALFORMED                        16

      Sent if the KINK_ISAKMP payload received was invalid because
      some type, length, or value was out of range.  It is also sent
      when the request was rejected for reason that was not matched
      with other error types.

5.6.  Delete Payloads

   KINK directly uses ISAKMP Delete payloads with no changes.

5.7.  KE Payloads

   IKE requires that perfect forward secrecy (PFS) be supported through
   the use of the KE payload.  KINK retains the ability to use PFS, but
   relaxes the requirement from must implement to SHOULD implement.  The
   reasons are described in the Security Considerations section.

6.  Message Construction and Constraints for IPsec DOI

   All commands, responses, and acknowledgements are bound together by
   the XID field of the message header.  The XID is normally a
   monotonically incrementing field, and is used by the initiator to
   differentiate between outstanding requests to a responder.  The XID
   field does not provide replay protection as that functionality is
   provided by the Kerberos mechanisms.  In addition, commands and
   responses MUST use a cryptographic checksum over the entire message
   if the two peers share a key via a ticket exchange.

   In all cases in this section, if a message contains a KINK_AP_REQ or
   KINK_AP_REP payload, other KINK payloads MAY be encapsulated in a
   KINK_ENCRYPT payload.

6.1.  REPLY Message

   The REPLY message is a generic reply that MUST contain either a
   messages MAY contain additional DOI-specific payloads such as ISAKMP
   payloads that are defined in the following sections.

6.2.  ACK Message

   ACKs are sent only when the ACKREQ bit is set in a REPLY message.  An
   ACK message MUST contain an AP-REQ payload and no other payload.

Sakane, et al.              Standards Track                    [Page 28]

RFC 4430                          KINK                        March 2006

6.3.  CREATE Message

   This message initiates an establishment of new security
   association(s).  The CREATE message must contain an AP-REQ payload
   and any DOI-specific payloads.

        KINK_ISAKMP payloads
            SA Payload
                 Proposal Payloads
                      Transform Payloads
            Nonce Payload (Ni)
            [IDci, IDcr]
            [Notification Payloads]

   Replies are of the following forms:

   REPLY KINK Header
        KINK_ISAKMP payloads
            SA Payload
                 Proposal Payloads
                      Transform Payload
            [Nonce Payload (Nr)]
            [IDci, IDcr]
            [Notification Payloads]

   Note that there MUST be at least a single proposal payload and a
   single transform payload in REPLY messages.  There will be multiple
   proposal payloads only when an SA bundle is negotiated.  Also: unlike
   IKE, the Nonce payload Nr is not required, and if it exists, an
   acknowledgement must be requested to indicate that the initiator's
   outgoing SAs must be modified.  If any of the first proposals are not
   chosen by the recipient, it SHOULD include the Nonce payload.

   KINK, like IKE, allows the creation of many SAs in one create
   command.  If any of the optimistic proposals are not chosen by the
   responder, it MUST request an ACK.

   If an IPsec DOI-specific error is encountered, the responder must
   reply with a Notify payload describing the error:

Sakane, et al.              Standards Track                    [Page 29]

RFC 4430                          KINK                        March 2006

   REPLY KINK Header
        KINK_ISAKMP payloads
            [Notification Payloads]

   If the responder finds a Kerberos error for which it can produce a
   valid authenticator, the REPLY takes the following form:

   REPLY KINK Header

   Finally, if the responder finds a Kerberos or KINK type of error for
   which it cannot create an AP-REP, it MUST reply with a lone

   REPLY KINK Header

6.4.  DELETE Message

   This message indicates that the sending peer has deleted or will
   shortly delete Security Association(s) with the other peer.

        KINK_ISAKMP payloads
            Delete Payloads
            [Notification Payloads]

   There are three forms of replies for a DELETE.  The normal form is:

   REPLY KINK Header
        KINK_ISAKMP payloads
            Delete Payloads
            [Notification Payloads]

   If an IPsec DOI-specific error is encountered, the responder must
   reply with a Notify payload describing the error:

Sakane, et al.              Standards Track                    [Page 30]

RFC 4430                          KINK                        March 2006

   REPLY KINK Header
        KINK_ISAKMP payloads
            [Notification Payloads]

   If the responder finds a Kerberos error for which it can produce a
   valid authenticator, the REPLY takes the following form:

   REPLY KINK Header

   If the responder finds a KINK or Kerberos type of error, it MUST
   reply with a lone KINK_KRB_ERROR or KINK_ERROR payload:

   REPLY KINK Header

6.5.  STATUS Message

   The STATUS command is used in two ways:

     1)  As a means to relay an ISAKMP Notification message.

     2)  As a means of probing a peer whether its epoch has changed for
         dead peer detection.

   STATUS contains the following payloads:
     KINK Header
        KINK_ISAKMP payload
            [Notification Payloads]]

   There are three forms of replies for a STATUS.  The normal form is:

   REPLY KINK Header
        KINK_ISAKMP payload
            [Notification Payloads]]

Sakane, et al.              Standards Track                    [Page 31]

RFC 4430                          KINK                        March 2006

   If the responder finds a Kerberos error for which it can produce a
   valid authenticator, the REPLY takes the following form:

   REPLY KINK Header

   If the responder finds a KINK or Kerberos type of error, it MUST
   reply with a lone KINK_KRB_ERROR or KINK_ERROR payload:

   REPLY KINK Header

6.6.  GETTGT Message

   A GETTGT command is only used to carry a Kerberos TGT and is not
   related to SA management; therefore, it contains only KINK_TGT_REQ
   payload and does not contain any DOI-specific payload.

   There are two forms of replies for a GETTGT.  In the normal form,
   where the responder is allowed to return its TGT, the REPLY contains
   KINK_TGT_REP payload.  If the responder is not allowed to return its
   TGT, it MUST reply with a KINK_ERROR payload.

7.  ISAKMP Key Derivation

   KINK uses the same key derivation mechanisms defined in section 5.5
   of [IKE], which is:

   KEYMAT = prf(SKEYID_d, [g(qm)^xy |] protocol | SPI | Ni_b [| Nr_b])

   The following differences apply:

     o    prf is the pseudo-random function corresponding to the session
          key's etype.  They are defined in [KCRYPTO].

     o    SKEYID_d is the session key in the Kerberos service ticket
          from the AP-REQ.  Note that subkeys are not used in KINK and
          MUST be ignored if received.

     o    Both Ni_b and Nr_b are the part of the Nonce payloads (Ni and
          Nr, respectively) as described in section 3.2 of [IKE].  Nr_b
          is optional, which means that Nr_b is treated as if a zero
          length value was supplied when the responder's nonce (Nr) does
          not exist.  When Nr exists, Nr_b MUST be included in the

Sakane, et al.              Standards Track                    [Page 32]

RFC 4430                          KINK                        March 2006

   Note that g(qm)^xy refers to the keying material generated when KE
   payloads are supplied using Diffie-Hellman key agreement.  This is
   explained in section 5.5 of [IKE].

   The rest of the key derivation (e.g., how to expand KEYMAT) follows
   IKE.  How to use derived keying materials is up to each service
   (e.g., section 4.5.2 of [IPSEC]).

8.  Key Usage Numbers for Kerberos Key Derivation

   Kerberos encrypt/decrypt functions and get_mic/verify_mic functions
   require "key usage numbers".  They are used to generate specific keys
   for cryptographic operations so that different keys are used for
   different purposes/objects.  KINK uses two usage numbers, listed

      Purpose                                   Usage number
      -------                                   ------------
      KINK_ENCRYPT payload (for encryption)      39
      Cksum field (for checksum)                 40

9.  Transport Considerations

   KINK uses UDP on port 910 to transport its messages.  There is one
   timer T which SHOULD take into consideration round-trip
   considerations and MUST implement a truncated exponential back-off
   mechanism.  The state machine is simple: any message that expects a
   response MUST retransmit the request using timer T.  Since Kerberos
   requires that messages be retransmitted with new times for replay
   protection, the message MUST be re-created each time including the
   checksum of the message.  Both commands and replies with the ACKREQ
   bit set are kept on retransmit timers.  When a KINK initiator
   receives a REPLY with the ACKREQ bit set, it MUST retain the ability
   to regenerate the ACK message for the transaction for a minimum of
   its full retransmission timeout cycle or until it notices that
   packets have arrived on the newly constructed SA, whichever comes

   When a KINK peer retransmits a message, it MUST create a new Kerberos
   authenticator for the AP-REQ so that the peer can differentiate
   between replays and dropped packets.  This results in a potential
   race condition when a retransmission occurs before an in-flight reply
   is received/processed.  To counter this race condition, the
   retransmitting party SHOULD keep a list of valid authenticators that
   are outstanding for any particular transaction.

Sakane, et al.              Standards Track                    [Page 33]

RFC 4430                          KINK                        March 2006

   When a KINK peer retransmits a command, it MUST use the same ticket
   within the retransmissions.  This is to avoid race conditions on
   using different keys, which result in different KEYMATs between an
   initiator and a responder.  For this reason, (1) an initiator MUST
   obtain a ticket whose lifetime is greater than the initiator's
   maximum transaction time including timeouts, or (2) it MUST continue
   to use the same ticket within a set of retransmissions, and iff it
   receives an error (most likely KRB_AP_ERR_TKT_EXPIRED) from the
   responder, it starts a new transaction with a new ticket.

10.  Security Considerations

   The principal names are the identities of the KINK services, but the
   traffic protected by SAs are identified by DOI-specific selectors (IP
   addresses, port numbers, etc.).  This may lead to a breakaway of
   SA-protected data from authentication.  For example, if two different
   hosts claim that they have the same IP address, it may be impossible
   to predict which principal's key protects the data.  Thus, an
   implementation must take care for the binding between principal names
   and the SA selectors.

   Sending errors without cryptographic protection must be handled very
   carefully.  There is a trade-off between wanting to be helpful in
   diagnosing a problem and wanting to avoid being a dupe in a denial of
   service attack.

   KINK cobbles together and reuses many parts of both Kerberos and IKE,
   the latter which in turn is cobbled together from many other memos.
   As such, KINK inherits many of the weaknesses and considerations of
   each of its components.  However, KINK uses only IKE phase 2 payloads
   to create and delete SAs; the security considerations which pertain
   to IKE phase 1 may be safely ignored.  However, being able to ignore
   IKE's authentication phase necessarily means that KINK inherits all
   of the security considerations of Kerberos authentication as outlined
   in [KERBEROS].  For one, a KDC, like an Authentication,
   Authorization, and Accounting (AAA) server, is a point of attack and
   all that implies.  Much has been written about various shortcomings
   and mitigations of Kerberos, and they should be evaluated for any

   KINK's use of Kerberos presents a couple of considerations.  First,
   KINK explicitly expects that the KDC will provide adequate entropy
   when it generates session keys.  Second, Kerberos is used as a user
   authentication protocol with the possibility of dictionary attacks on
   user passwords.  This memo does not describe a particular method to
   avoid these pitfalls, but recommends that suitable randomly generated

Sakane, et al.              Standards Track                    [Page 34]

RFC 4430                          KINK                        March 2006

   keys should be used for the service principals such as using the
   -randomkey option with MIT's "kadmin addprinc" command as well as for
   clients when that is practical.

   Kerberos does not currently provide perfect forward secrecy in
   general.  KINK with the KE payload can provide PFS for a service key
   from a Kerberos key, but the KE is not mandatory because of the
   computational cost.  This is a trade-off and operators can choose the
   PFS over the cost, and vice versa.  KINK itself should be secure from
   offline analysis from compromised principal passphrases if PFS is
   used, but from an overall system's standpoint, the existence of other
   Kerberized services that do not provide PFS makes this a less than
   optimal situation.

11.  IANA Considerations

   The IANA has assigned a well-known port number for KINK.

   The IANA has created a new registry for KINK parameters, and has
   registered the following identifiers.

      KINK Message Types (section 4)
      KINK Next Payload Types (section 4.2)
      KINK Error Codes (section 4.2.8)

   Changes and additions to this registry follow the policies described
   below.  Their meanings are described in [BCP26].

     o    Using the numbers in the "Private Use" range is Private Use.

     o    Assignment from the "RESERVED TO IANA" range needs Standards
          Action, or non-standards-track RFCs with Expert Review.
          (Though the full specification may be a public and permanent
          document of a standards body other than IETF, an RFC referring
          it is needed.)

     o    Other change requires Standards Action.

12.  Forward Compatibility Considerations

   KINK can accommodate future versions of Quick Mode through the use of
   the version field in the ISAKMP payload as well as new domains of
   interpretation.  In this memo, the only supported Quick Mode version
   is 1.0, which corresponds to [IKE].  Likewise, the only DOI supported
   is the IPsec domain of interpretation [IPDOI].  New Quick Mode
   versions and DOIs MUST be described in subsequent memos.

Sakane, et al.              Standards Track                    [Page 35]

RFC 4430                          KINK                        March 2006

   KINK implementations MUST reject ISAKMP versions that are greater
   than the highest currently supported version with a KINK_BADQMVERS
   error type.  A KINK implementation that receives a KINK_BADQMVERS
   message SHOULD be capable of reverting back to version 1.0.

12.1.  New Versions of Quick Mode

   The IPsec working group is defining the next-generation IKE protocol
   [IKEv2], which does not use Quick Mode, but it is similar to the one
   in IKEv1.  The difference between the two is summarized in Appendix A
   of [IKEv2].  Each of them must be considered in order to use IKEv2
   with KINK.

12.2.  New DOI

   The KINK message header contains a field called "Domain of
   Interpretation (DOI)" to allow other domains of interpretation to use
   KINK as a secure transport mechanism for keying.

   As one example of a new DOI, the MSEC working group defined the Group
   Domain of Interpretation [GDOI], which defines a few new messages,
   which look like ISAKMP messages, but are not defined in ISAKMP.

   In order to carry GDOI messages in KINK, the DOI field in the KINK
   header would indicate that GDOI is being used, instead of IPSEC-DOI,
   and the KINK_ISAKMP payload would contain the payloads defined in the
   GDOI document rather than the payloads used by [IKE] Quick Mode.  The
   version number in the KINK_ISAKMP header is related to the DOI in the
   KINK header, so a maj.min version 1.0 under DOI GDOI is different
   from a maj.min version 1.0 under DOI IPSEC-DOI.

13.  Related Work

   The IPsec working group has defined a number of protocols that
   provide the ability to create and maintain cryptographically secure
   SAs at layer three (i.e., the IP layer).  This effort has produced
   two distinct protocols:

     o    a mechanism for encrypting and authenticating IP datagram
          payloads that assumes a shared secret between the sender and

     o    a mechanism for IPsec peers to perform mutual authentication
          and exchange keying material

   The IPsec working group has defined a peer-to-peer authentication and
   keying mechanism, IKE (RFC 2409).  One of the drawbacks of a peer-
   to-peer protocol is that each peer must know and implement a site's

Sakane, et al.              Standards Track                    [Page 36]

RFC 4430                          KINK                        March 2006

   security policy, which in practice can be quite complex.  In
   addition, the peer-to-peer nature of IKE requires the use of Diffie-
   Hellman (DH) to establish a shared secret.  DH, unfortunately, is
   computationally quite expensive and prone to denial of service
   attacks.  IKE also relies on X.509 certificates to realize scalable
   authentication of peers.  Digital signatures are also computationally
   expensive, and certificate-based trust models are difficult to deploy
   in practice.  While IKE does allow for a pre-shared key, key
   distribution is required between all peers -- an O(n^2) problem --
   which is problematic for large deployments.

14.  Acknowledgements

   Many have contributed to the KINK effort, including our working group
   chairs Derek Atkins and Jonathan Trostle.  The original inspiration
   came from CableLab's PacketCable effort, which defined a simplified
   version of Kerberized IPsec, including Sasha Medvinsky, Mike Froh,
   and Matt Hur and David McGrew.  The inspiration for wholly reusing
   IKE phase 2 is the result of Tero Kivinen's document suggesting
   grafting Kerberos authentication onto Quick Mode.

15.  References

15.1.  Normative References

   [BCP26]       Narten, T. and H. Alvestrand, "Guidelines for Writing
                 an IANA Considerations Section in RFCs", BCP 26, RFC
                 2434, October 1998.

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

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

   [IPSEC]       Kent, S. and K. Seo, "Security Architecture for the
                 Internet Protocol", RFC 4301, December 2005.

   [ISAKMP]      Maughan,  D., Schertler, M., Schneider, M., and J.
                 Turner, "Internet Security Association and Key
                 Management Protocol (ISAKMP)", RFC 2408, November 1998.

   [ISAKMP-REG]  IANA, "Internet Security Association and Key Management
                 Protocol (ISAKMP) Identifiers",

   [KCRYPTO]     Raeburn, K., "Encryption and Checksum Specifications
                 for Kerberos 5", RFC 3961, February 2005.

Sakane, et al.              Standards Track                    [Page 37]

RFC 4430                          KINK                        March 2006

   [KERBEROS]    Neuman, C., Yu, T., Hartman, S., and K. Raeburn, "The
                 Kerberos Network Authentication Service (V5)", RFC
                 4120, July 2005.

   [RFC1964]     Linn, J., "The Kerberos Version 5 GSS-API Mechanism",
                 RFC 1964, June 1996.

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

15.2.  Informative References

   [GDOI]        Baugher, M., Weis, B., Hardjono, T., and H. Harney,
                 "The Group Domain of Interpretation", RFC 3547, July

   [IKEv2]       Kaufman, C., "Internet Key Exchange (IKEv2) Protocol",
                 RFC 4306, December 2005.

   [PKINIT]      Zhu, L. and B. Tung, "Public Key Cryptography for
                 Initial Authentication in Kerberos", Work in Progress,
                 February 2006.

   [REQ4KINK]    Thomas, M., "Requirements for Kerberized Internet
                 Negotiation of Keys", RFC 3129, June 2001.

   [RFC793]      Postel, J., "Transmission Control Protocol", STD 7, RFC
                 793, September 1981.

   [RFC2743]     Linn, J., "Generic Security Service Application Program
                 Interface Version 2, Update 1", RFC 2743, January 2000.

Sakane, et al.              Standards Track                    [Page 38]

RFC 4430                          KINK                        March 2006

Authors' Addresses

   Shoichi Sakane
   Yokogawa Electric Corporation
   2-9-32 Nakacho, Musashino-shi,
   Tokyo 180-8750 Japan

   EMail: Shouichi.Sakane@jp.yokogawa.com

   Ken'ichi Kamada
   Yokogawa Electric Corporation
   2-9-32 Nakacho, Musashino-shi,
   Tokyo 180-8750 Japan

   EMail: Ken-ichi.Kamada@jp.yokogawa.com

   Michael Thomas
   Cisco Systems
   170 West Tasman Drive
   San Jose, CA 95134

   EMail: mat@cisco.com

   Jan Vilhuber
   Cisco Systems
   170 West Tasman Drive
   San Jose, CA 95134

   EMail: vilhuber@cisco.com

Sakane, et al.              Standards Track                    [Page 39]

RFC 4430                          KINK                        March 2006

Full Copyright Statement

   Copyright (C) The Internet Society (2006).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an

Intellectual Property

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at


   Funding for the RFC Editor function is provided by the IETF
   Administrative Support Activity (IASA).

Sakane, et al.              Standards Track                    [Page 40]