RFC 8784

Internet Engineering Task Force (IETF)                        S. Fluhrer
Request for Comments: 8784                                 P. Kampanakis
Category: Standards Track                                      D. McGrew
ISSN: 2070-1721                                            Cisco Systems
                                                              V. Smyslov
                                                               June 2020

Mixing Preshared Keys in the Internet Key Exchange Protocol Version 2
                   (IKEv2) for Post-quantum Security


   The possibility of quantum computers poses a serious challenge to
   cryptographic algorithms deployed widely today.  The Internet Key
   Exchange Protocol Version 2 (IKEv2) is one example of a cryptosystem
   that could be broken; someone storing VPN communications today could
   decrypt them at a later time when a quantum computer is available.
   It is anticipated that IKEv2 will be extended to support quantum-
   secure key exchange algorithms; however, that is not likely to happen
   in the near term.  To address this problem before then, this document
   describes an extension of IKEv2 to allow it to be resistant to a
   quantum computer by using preshared keys.

Status of This Memo

   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at

Copyright Notice

   Copyright (c) 2020 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1.  Introduction
     1.1.  Requirements Language
   2.  Assumptions
   3.  Exchanges
   4.  Upgrade Procedure
   5.  PPK
     5.1.  PPK_ID Format
     5.2.  Operational Considerations
       5.2.1.  PPK Distribution
       5.2.2.  Group PPK
       5.2.3.  PPK-Only Authentication
   6.  Security Considerations
   7.  IANA Considerations
   8.  References
     8.1.  Normative References
     8.2.  Informative References
   Appendix A.  Discussion and Rationale

   Authors' Addresses

1.  Introduction

   Recent achievements in developing quantum computers demonstrate that
   it is probably feasible to build one that is cryptographically
   significant.  If such a computer is implemented, many of the
   cryptographic algorithms and protocols currently in use would be
   insecure.  A quantum computer would be able to solve Diffie-Hellman
   (DH) and Elliptic Curve Diffie-Hellman (ECDH) problems in polynomial
   time [C2PQ], and this would imply that the security of existing IKEv2
   [RFC7296] systems would be compromised.  IKEv1 [RFC2409], when used
   with strong preshared keys, is not vulnerable to quantum attacks
   because those keys are one of the inputs to the key derivation
   function.  If the preshared key has sufficient entropy and the
   Pseudorandom Function (PRF), encryption, and authentication
   transforms are quantum secure, then the resulting system is believed
   to be quantum secure -- that is, secure against classical attackers
   of today or future attackers with a quantum computer.

   This document describes a way to extend IKEv2 to have a similar
   property; assuming that the two end systems share a long secret key,
   then the resulting exchange is quantum secure.  By bringing post-
   quantum security to IKEv2, this document removes the need to use an
   obsolete version of IKE in order to achieve that security goal.

   The general idea is that we add an additional secret that is shared
   between the initiator and the responder; this secret is in addition
   to the authentication method that is already provided within IKEv2.
   We stir this secret into the SK_d value, which is used to generate
   the key material (KEYMAT) for the Child Security Associations (SAs)
   and the SKEYSEED for the IKE SAs created as a result of the initial
   IKE SA rekey.  This secret provides quantum resistance to the IPsec
   SAs and any subsequent IKE SAs.  We also stir the secret into the
   SK_pi and SK_pr values; this allows both sides to detect a secret
   mismatch cleanly.

   It was considered important to minimize the changes to IKEv2.  The
   existing mechanisms to perform authentication and key exchange remain
   in place (that is, we continue to perform (EC)DH and potentially PKI
   authentication if configured).  This document does not replace the
   authentication checks that the protocol does; instead, they are
   strengthened by using an additional secret key.

1.1.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

2.  Assumptions

   We assume that each IKE peer has a list of Post-quantum Preshared
   Keys (PPKs) along with their identifiers (PPK_ID), and any potential
   IKE initiator selects which PPK to use with any specific responder.
   In addition, implementations have a configurable flag that determines
   whether this PPK is mandatory.  This PPK is independent of the
   preshared key (if any) that the IKEv2 protocol uses to perform
   authentication (because the preshared key in IKEv2 is not used for
   any key derivation and thus doesn't protect against quantum
   computers).  The PPK-specific configuration that is assumed to be on
   each node consists of the following tuple:

   Peer, PPK, PPK_ID, mandatory_or_not

   We assume the reader is familiar with the payload notation defined in
   Section 1.2 of [RFC7296].

3.  Exchanges

   If the initiator is configured to use a PPK with the responder
   (whether or not the use of the PPK is mandatory), then it MUST
   include a notification USE_PPK in the IKE_SA_INIT request message as

   Initiator                       Responder
   HDR, SAi1, KEi, Ni, N(USE_PPK)  --->

   N(USE_PPK) is a status notification payload with the type 16435; it
   has a protocol ID of 0, no Security Parameter Index (SPI), and no
   notification data associated with it.

   If the initiator needs to resend this initial message with a COOKIE
   notification, then the resend would include the USE_PPK notification
   if the original message did (see Section 2.6 of [RFC7296]).

   If the responder does not support this specification or does not have
   any PPK configured, then it ignores the received notification (as
   defined in [RFC7296] for unknown status notifications) and continues
   with the IKEv2 protocol as normal.  Otherwise, the responder replies
   with the IKE_SA_INIT message, including a USE_PPK notification in the

   Initiator                       Responder
                   <--- HDR, SAr1, KEr, Nr, [CERTREQ,] N(USE_PPK)

   When the initiator receives this reply, it checks whether the
   responder included the USE_PPK notification.  If the responder did
   not include the USE_PPK notification and the flag mandatory_or_not
   indicates that using PPKs is mandatory for communication with this
   responder, then the initiator MUST abort the exchange.  This
   situation may happen in case of misconfiguration, i.e., when the
   initiator believes it has a mandatory-to-use PPK for the responder
   and the responder either doesn't support PPKs at all or doesn't have
   any PPK configured for the initiator.  See Section 6 for discussion
   of the possible impacts of this situation.

   If the responder did not include the USE_PPK notification and using a
   PPK for this particular responder is optional, then the initiator
   continues with the IKEv2 protocol as normal, without using PPKs.

   If the responder did include the USE_PPK notification, then the
   initiator selects a PPK, along with its identifier PPK_ID.  Then, it
   computes this modification of the standard IKEv2 key derivation from
   Section 2.14 of [RFC7296]:

    SKEYSEED = prf(Ni | Nr, g^ir)
    {SK_d' | SK_ai | SK_ar | SK_ei | SK_er | SK_pi' | SK_pr'}
                    = prf+ (SKEYSEED, Ni | Nr | SPIi | SPIr)

    SK_d  = prf+ (PPK, SK_d')
    SK_pi = prf+ (PPK, SK_pi')
    SK_pr = prf+ (PPK, SK_pr')

   That is, we use the standard IKEv2 key derivation process, except
   that the three resulting subkeys SK_d, SK_pi, and SK_pr (marked with
   primes in the formula above) are then run through the prf+ again,
   this time using the PPK as the key.  The result is the unprimed
   versions of these keys, which are then used as inputs to subsequent
   steps of the IKEv2 exchange.

   Using a prf+ construction ensures that it is always possible to get
   the resulting keys of the same size as the initial ones, even if the
   underlying PRF has an output size different from its key size.  Note
   that at the time of this writing, all PRFs defined for use in IKEv2
   (see the "Transform Type 2 - Pseudorandom Function Transform IDs"
   subregistry [IANA-IKEV2]) have an output size equal to the
   (preferred) key size.  For such PRFs, only the first iteration of
   prf+ is needed:

    SK_d  = prf (PPK, SK_d'  | 0x01)
    SK_pi = prf (PPK, SK_pi' | 0x01)
    SK_pr = prf (PPK, SK_pr' | 0x01)

   Note that the PPK is used in SK_d, SK_pi, and SK_pr calculations only
   during the initial IKE SA setup.  It MUST NOT be used when these
   subkeys are calculated as result of IKE SA rekey, resumption, or
   other similar operations.

   The initiator then sends the IKE_AUTH request message, including the
   PPK_ID value as follows:

   Initiator                       Responder
       [IDr,] AUTH, SAi2,
       TSi, TSr, N(PPK_IDENTITY, PPK_ID), [N(NO_PPK_AUTH)]}  --->

   PPK_IDENTITY is a status notification with the type 16436; it has a
   protocol ID of 0, no SPI, and notification data that consists of the
   identifier PPK_ID.

   A situation may happen when the responder has some PPKs but doesn't
   have a PPK with the PPK_ID received from the initiator.  In this
   case, the responder cannot continue with the PPK (in particular, it
   cannot authenticate the initiator), but the responder could be able
   to continue with the normal IKEv2 protocol if the initiator provided
   its authentication data computed as in the normal IKEv2 without using
   PPKs.  For this purpose, if using PPKs for communication with this
   responder is optional for the initiator (based on the
   mandatory_or_not flag), then the initiator MUST include a NO_PPK_AUTH
   notification in the above message.  This notification informs the
   responder that PPKs are optional and allows for authenticating the
   initiator without using PPKs.

   NO_PPK_AUTH is a status notification with the type 16437; it has a
   protocol ID of 0 and no SPI.  The Notification Data field contains
   the initiator's authentication data computed using SK_pi', which has
   been computed without using PPKs.  This is the same data that would
   normally be placed in the Authentication Data field of an AUTH
   payload.  Since the Auth Method field is not present in the
   notification, the authentication method used for computing the
   authentication data MUST be the same as the method indicated in the
   AUTH payload.  Note that if the initiator decides to include the
   NO_PPK_AUTH notification, the initiator needs to perform
   authentication data computation twice, which may consume computation
   power (e.g., if digital signatures are involved).

   When the responder receives this encrypted exchange, it first
   computes the values:

    SKEYSEED = prf(Ni | Nr, g^ir)
    {SK_d' | SK_ai | SK_ar | SK_ei | SK_er | SK_pi' | SK_pr'}
                    = prf+ (SKEYSEED, Ni | Nr | SPIi | SPIr)

   The responder then uses the SK_ei/SK_ai values to decrypt/check the
   message and then scans through the payloads for the PPK_ID attached
   to the PPK_IDENTITY notification.  If no PPK_IDENTITY notification is
   found and the peers successfully exchanged USE_PPK notifications in
   the IKE_SA_INIT exchange, then the responder MUST send back an
   AUTHENTICATION_FAILED notification and then fail the negotiation.

   If the PPK_IDENTITY notification contains a PPK_ID that is not known
   to the responder or is not configured for use for the identity from
   the IDi payload, then the responder checks whether using PPKs for
   this initiator is mandatory and whether the initiator included a
   NO_PPK_AUTH notification in the message.  If using PPKs is mandatory
   or no NO_PPK_AUTH notification is found, then the responder MUST send
   back an AUTHENTICATION_FAILED notification and then fail the
   negotiation.  Otherwise (when a PPK is optional and the initiator
   included a NO_PPK_AUTH notification), the responder MAY continue the
   regular IKEv2 protocol, except that it uses the data from the
   NO_PPK_AUTH notification as the authentication data (which usually
   resides in the AUTH payload) for the purpose of the initiator
   authentication.  Note that the authentication method is still
   indicated in the AUTH payload.

   Table 1 summarizes the above logic for the responder:

   | Received | Received    | Configured | PPK is    | Action         |
   | USE_PPK  | NO_PPK_AUTH | with PPK   | Mandatory |                |
   | No       | *           | No         | *         | Standard IKEv2 |
   |          |             |            |           | protocol       |
   | No       | *           | Yes        | No        | Standard IKEv2 |
   |          |             |            |           | protocol       |
   | No       | *           | Yes        | Yes       | Abort          |
   |          |             |            |           | negotiation    |
   | Yes      | No          | No         | *         | Abort          |
   |          |             |            |           | negotiation    |
   | Yes      | Yes         | No         | Yes       | Abort          |
   |          |             |            |           | negotiation    |
   | Yes      | Yes         | No         | No        | Standard IKEv2 |
   |          |             |            |           | protocol       |
   | Yes      | *           | Yes        | *         | Use PPK        |

                                 Table 1

   If a PPK is in use, then the responder extracts the corresponding PPK
   and computes the following values:

    SK_d  = prf+ (PPK, SK_d')
    SK_pi = prf+ (PPK, SK_pi')
    SK_pr = prf+ (PPK, SK_pr')

   The responder then continues with the IKE_AUTH exchange (validating
   the AUTH payload that the initiator included) as usual and sends back
   a response, which includes the PPK_IDENTITY notification with no data
   to indicate that the PPK is used in the exchange:

   Initiator                       Responder
                              <--  HDR, SK {IDr, [CERT,]
                                   AUTH, SAr2,
                                   TSi, TSr, N(PPK_IDENTITY)}

   When the initiator receives the response, it checks for the presence
   of the PPK_IDENTITY notification.  If it receives one, it marks the
   SA as using the configured PPK to generate SK_d, SK_pi, and SK_pr (as
   shown above); the content of the received PPK_IDENTITY (if any) MUST
   be ignored.  If the initiator does not receive the PPK_IDENTITY, it
   MUST either fail the IKE SA negotiation sending the
   AUTHENTICATION_FAILED notification in the INFORMATIONAL exchange (if
   the PPK was configured as mandatory) or continue without using the
   PPK (if the PPK was not configured as mandatory and the initiator
   included the NO_PPK_AUTH notification in the request).

   If the Extensible Authentication Protocol (EAP) is used in the
   IKE_AUTH exchange, then the initiator doesn't include the AUTH
   payload in the first request message; however, the responder sends
   back the AUTH payload in the first reply.  The peers then exchange
   AUTH payloads after EAP is successfully completed.  As a result, the
   responder sends the AUTH payload twice -- in the first and last
   IKE_AUTH reply message -- while the initiator sends the AUTH payload
   only in the last IKE_AUTH request.  See more details about EAP
   authentication in IKEv2 in Section 2.16 of [RFC7296].

   The general rule for using a PPK in the IKE_AUTH exchange, which
   covers the EAP authentication case too, is that the initiator
   includes a PPK_IDENTITY (and optionally a NO_PPK_AUTH) notification
   in the request message containing the AUTH payload.  Therefore, in
   case of EAP, the responder always computes the AUTH payload in the
   first IKE_AUTH reply message without using a PPK (by means of
   SK_pr'), since PPK_ID is not yet known to the responder.  Once the
   IKE_AUTH request message containing the PPK_IDENTITY notification is
   received, the responder follows the rules described above for the
   non-EAP authentication case.

      Initiator                         Responder
      HDR, SK {IDi, [CERTREQ,]
          [IDr,] SAi2,
          TSi, TSr}  -->
                                   <--  HDR, SK {IDr, [CERT,] AUTH,
      HDR, SK {EAP}  -->
                                   <--  HDR, SK {EAP (success)}
      HDR, SK {AUTH,
          [, N(NO_PPK_AUTH)]}  -->
                                   <--  HDR, SK {AUTH, SAr2, TSi, TSr
                                        [, N(PPK_IDENTITY)]}

   Note that the diagram above shows both the cases when the responder
   uses a PPK and when it chooses not to use it (provided the initiator
   has included the NO_PPK_AUTH notification); thus, the responder's
   PPK_IDENTITY notification is marked as optional.  Also, note that the
   IKE_SA_INIT exchange using a PPK is as described above (including
   exchange of the USE_PPK notifications), regardless of whether or not
   EAP is employed in the IKE_AUTH.

4.  Upgrade Procedure

   This algorithm was designed so that someone can introduce PPKs into
   an existing IKE network without causing network disruption.

   In the initial phase of the network upgrade, the network
   administrator would visit each IKE node and configure:

   *  The set of PPKs (and corresponding PPK_IDs) that this node would
      need to know.

   *  The PPK that will be used for each peer that this node would
      initiate to.

   *  The value "false" for the mandatory_or_not flag for each peer that
      this node would initiate to (thus indicating that the use of PPKs
      is not mandatory).

   With this configuration, the node will continue to operate with nodes
   that have not yet been upgraded.  This is due to the USE_PPK
   notification and the NO_PPK_AUTH notification; if the initiator has
   not been upgraded, it will not send the USE_PPK notification (and so
   the responder will know that the peers will not use a PPK).  If the
   responder has not been upgraded, it will not send the USE_PPK
   notification (and so the initiator will know to not use a PPK).  If
   both peers have been upgraded but the responder isn't yet configured
   with the PPK for the initiator, then the responder could continue
   with the standard IKEv2 protocol if the initiator sent a NO_PPK_AUTH
   notification.  If both the responder and initiator have been upgraded
   and properly configured, they will both realize it, and the Child SAs
   will be quantum secure.

   As an optional second step, after all nodes have been upgraded, the
   administrator should then go back through the nodes and mark the use
   of a PPK as mandatory.  This will not affect the strength against a
   passive attacker, but it would mean that an active attacker with a
   quantum computer (which is sufficiently fast to be able to break the
   (EC)DH in real time) would not be able to perform a downgrade attack.

5.  PPK

5.1.  PPK_ID Format

   This standard requires that both the initiator and the responder have
   a secret PPK value, with the responder selecting the PPK based on the
   PPK_ID that the initiator sends.  In this standard, both the
   initiator and the responder are configured with fixed PPK and PPK_ID
   values and perform the lookup based on the PPK_ID value.  It is
   anticipated that later specifications will extend this technique to
   allow dynamically changing PPK values.  To facilitate such an
   extension, we specify that the PPK_ID the initiator sends will have
   its first octet be the PPK_ID type value.  This document defines two
   values for the PPK_ID type:

   *  PPK_ID_OPAQUE (1) - For this type, the format of the PPK_ID (and
      the PPK itself) is not specified by this document; it is assumed
      to be mutually intelligible by both the initiator and the
      responder.  This PPK_ID type is intended for those implementations
      that choose not to disclose the type of PPK to active attackers.

   *  PPK_ID_FIXED (2) - In this case, the format of the PPK_ID and the
      PPK are fixed octet strings; the remaining bytes of the PPK_ID are
      a configured value.  We assume that there is a fixed mapping
      between PPK_ID and PPK, which is configured locally to both the
      initiator and the responder.  The responder can use the PPK_ID to
      look up the corresponding PPK value.  Not all implementations are
      able to configure arbitrary octet strings; to improve the
      potential interoperability, it is recommended that, in the
      PPK_ID_FIXED case, both the PPK and the PPK_ID strings be limited
      to the base64 character set [RFC4648].

5.2.  Operational Considerations

   The need to maintain several independent sets of security credentials
   can significantly complicate a security administrator's job and can
   potentially slow down widespread adoption of this specification.  It
   is anticipated that administrators will try to simplify their job by
   decreasing the number of credentials they need to maintain.  This
   section describes some of the considerations for PPK management.

5.2.1.  PPK Distribution

   PPK_IDs of the type PPK_ID_FIXED (and the corresponding PPKs) are
   assumed to be configured within the IKE device in an out-of-band
   fashion.  While the method of distribution is a local matter and is
   out of scope of this document or IKEv2, [RFC6030] describes a format
   for the transport and provisioning of symmetric keys.  That format
   could be reused using the PIN profile (defined in Section 10.2 of
   [RFC6030]) with the "Id" attribute of the <Key> element being the
   PPK_ID (without the PPK_ID type octet for a PPK_ID_FIXED) and the
   <Secret> element containing the PPK.

5.2.2.  Group PPK

   This document doesn't explicitly require that the PPK be unique for
   each pair of peers.  If this is the case, then this solution provides
   full peer authentication, but it also means that each host must have
   as many independent PPKs as peers it is going to communicate with.
   As the number of peers grows, the PPKs will not scale.

   It is possible to use a single PPK for a group of users.  Since each
   peer uses classical public key cryptography in addition to a PPK for
   key exchange and authentication, members of the group can neither
   impersonate each other nor read each other's traffic unless they use
   quantum computers to break public key operations.  However, group
   members can record any traffic they have access to that comes from
   other group members and decrypt it later, when they get access to a
   quantum computer.

   In addition, the fact that the PPK is known to a (potentially large)
   group of users makes it more susceptible to theft.  When an attacker
   equipped with a quantum computer gets access to a group PPK, all
   communications inside the group are revealed.

   For these reasons, using a group PPK is NOT RECOMMENDED.

5.2.3.  PPK-Only Authentication

   If quantum computers become a reality, classical public key
   cryptography will provide little security, so administrators may find
   it attractive not to use it at all for authentication.  This will
   reduce the number of credentials they need to maintain because they
   only need to maintain PPK credentials.  Combining group PPK and PPK-
   only authentication is NOT RECOMMENDED since, in this case, any
   member of the group can impersonate any other member, even without
   the help of quantum computers.

   PPK-only authentication can be achieved in IKEv2 if the NULL
   Authentication method [RFC7619] is employed.  Without PPK, the NULL
   Authentication method provides no authentication of the peers;
   however, since a PPK is stirred into the SK_pi and the SK_pr, the
   peers become authenticated if a PPK is in use.  Using PPKs MUST be
   mandatory for the peers if they advertise support for PPKs in
   IKE_SA_INIT and use NULL Authentication.  Additionally, since the
   peers are authenticated via PPKs, the ID Type in the IDi/IDr payloads
   SHOULD NOT be ID_NULL, despite using the NULL Authentication method.

6.  Security Considerations

   A critical consideration is how to ensure the randomness of this
   post-quantum preshared key.  Quantum computers are able to perform
   Grover's algorithm [GROVER]; that effectively halves the size of a
   symmetric key.  In addition, an adversary impersonating the server,
   even with a conventional computer, can perform a dictionary search
   over plausible post-quantum preshared key values.  The strongest
   practice is to ensure that any post-quantum preshared key contains at
   least 256 bits of entropy; this will provide 128 bits of post-quantum
   security, while providing security against conventional dictionary
   attacks.  That provides the security equivalent to Category 5 as
   defined in the NIST Post-Quantum Cryptography Call for Proposals
   [NISTPQCFP].  Deriving a post-quantum preshared key from a password,
   name, or other low-entropy source is not secure because of these
   known attacks.

   With this protocol, the computed SK_d is a function of the PPK.
   Assuming that the PPK has sufficient entropy (for example, at least
   2^(256) possible values), even if an attacker was able to recover the
   rest of the inputs to the PRF function, it would be infeasible to use
   Grover's algorithm with a quantum computer to recover the SK_d value.
   Similarly, all keys that are a function of SK_d, which include all
   Child SA keys and all keys for subsequent IKE SAs (created when the
   initial IKE SA is rekeyed), are also quantum secure (assuming that
   the PPK was of high enough entropy and that all the subkeys are
   sufficiently long).

   An attacker with a quantum computer that can decrypt the initial IKE
   SA has access to all the information exchanged over it, such as
   identities of the peers, configuration parameters, and all negotiated
   IPsec SA information (including traffic selectors), with the
   exception of the cryptographic keys used by the IPsec SAs, which are
   protected by the PPK.

   Deployments that treat this information as sensitive or that send
   other sensitive data (like cryptographic keys) over IKE SAs MUST
   rekey the IKE SA before the sensitive information is sent to ensure
   this information is protected by the PPK.  It is possible to create a
   childless IKE SA as specified in [RFC6023].  This prevents Child SA
   configuration information from being transmitted in the original IKE
   SA that is not protected by a PPK.  Some information related to IKE
   SA that is sent in the IKE_AUTH exchange, such as peer identities,
   feature notifications, vendor IDs, etc., cannot be hidden from the
   attack described above, even if the additional IKE SA rekey is

   In addition, the policy SHOULD be set to negotiate only quantum-
   secure symmetric algorithms; while this RFC doesn't claim to give
   advice as to what algorithms are secure (as that may change based on
   future cryptographical results), below is a list of defined IKEv2 and
   IPsec algorithms that should not be used, as they are known to
   provide less than 128 bits of post-quantum security:

   *  Any IKEv2 encryption algorithm, PRF, or integrity algorithm with a
      key size less than 256 bits.

   *  Any ESP transform with a key size less than 256 bits.

   *  PRF_AES128_XCBC and PRF_AES128_CBC: even though they can use as
      input a key of arbitrary size, such input keys are converted into
      a 128-bit key for internal use.

   Section 3 requires the initiator to abort the initial exchange if
   using PPKs is mandatory for it but the responder does not include the
   USE_PPK notification in the response.  In this situation, when the
   initiator aborts the negotiation, it leaves a half-open IKE SA on the
   responder (because IKE_SA_INIT completes successfully from the
   responder's point of view).  This half-open SA will eventually expire
   and be deleted, but if the initiator continues its attempts to create
   IKE SA with a high enough rate, then the responder may consider it a
   denial-of-service (DoS) attack and take protective measures (see
   [RFC8019] for more details).  In this situation, it is RECOMMENDED
   that the initiator cache the negative result of the negotiation and
   not attempt to create it again for some time.  This period of time
   may vary, but it is believed that waiting for at least few minutes
   will not cause the responder to treat it as a DoS attack.  Note that
   this situation would most likely be a result of misconfiguration, and
   some reconfiguration of the peers would probably be needed.

   If using PPKs is optional for both peers and they authenticate
   themselves using digital signatures, then an attacker in between,
   equipped with a quantum computer capable of breaking public key
   operations in real time, is able to mount a downgrade attack by
   removing the USE_PPK notification from the IKE_SA_INIT and forging
   digital signatures in the subsequent exchange.  If using PPKs is
   mandatory for at least one of the peers or if a preshared key mode is
   used for authentication, then the attack will be detected and the SA
   won't be created.

   If using PPKs is mandatory for the initiator, then an attacker able
   to eavesdrop and inject packets into the network can prevent creation
   of an IKE SA by mounting the following attack.  The attacker
   intercepts the initial request containing the USE_PPK notification
   and injects a forged response containing no USE_PPK.  If the attacker
   manages to inject this packet before the responder sends a genuine
   response, then the initiator would abort the exchange.  To thwart
   this kind of attack, it is RECOMMENDED that, if using PPKs is
   mandatory for the initiator and the received response doesn't contain
   the USE_PPK notification, the initiator not abort the exchange
   immediately.  Instead, it waits for more response messages,
   retransmitting the request as if no responses were received at all,
   until either the received message contains the USE_PPK notification
   or the exchange times out (see Section 2.4 of [RFC7296] for more
   details about retransmission timers in IKEv2).  If none of the
   received responses contains USE_PPK, then the exchange is aborted.

   If using a PPK is optional for both peers, then in case of
   misconfiguration (e.g., mismatched PPK_ID), the IKE SA will be
   created without protection against quantum computers.  It is advised
   that if a PPK was configured but was not used for a particular IKE
   SA, then implementations SHOULD audit this event.

7.  IANA Considerations

   This document defines three new Notify Message Types in the "IKEv2
   Notify Message Types - Status Types" subregistry under the "Internet
   Key Exchange Version 2 (IKEv2) Parameters" registry [IANA-IKEV2]:

          | Value | NOTIFY MESSAGES - STATUS TYPES | Reference |
          | 16435 | USE_PPK                        | RFC 8784  |
          | 16436 | PPK_IDENTITY                   | RFC 8784  |
          | 16437 | NO_PPK_AUTH                    | RFC 8784  |

                                 Table 2

   Per this document, IANA has created a new subregistry titled "IKEv2
   Post-quantum Preshared Key ID Types" under the "Internet Key Exchange
   Version 2 (IKEv2) Parameters" registry [IANA-IKEV2].  This new
   subregistry is for the PPK_ID types used in the PPK_IDENTITY
   notification defined in this specification.  The initial contents of
   the new subregistry are as follows:

            | Value   | PPK_ID Type              | Reference |
            | 0       | Reserved                 | RFC 8784  |
            | 1       | PPK_ID_OPAQUE            | RFC 8784  |
            | 2       | PPK_ID_FIXED             | RFC 8784  |
            | 3-127   | Unassigned               | RFC 8784  |
            | 128-255 | Reserved for Private Use | RFC 8784  |

                                 Table 3

   The PPK_ID type value 0 is reserved; values 3-127 are to be assigned
   by IANA; and values 128-255 are for private use among mutually
   consenting parties.  To register new PPK_IDs in the Unassigned range,
   a type name, a value between 3 and 127, and a reference specification
   need to be defined.  Changes and additions to the Unassigned range of
   this registry are made using the Expert Review policy [RFC8126].
   Changes and additions to the Reserved for Private Use range of this
   registry are made using the Private Use policy [RFC8126].

8.  References

8.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,

   [RFC7296]  Kaufman, C., Hoffman, P., Nir, Y., Eronen, P., and T.
              Kivinen, "Internet Key Exchange Protocol Version 2
              (IKEv2)", STD 79, RFC 7296, DOI 10.17487/RFC7296, October
              2014, <https://www.rfc-editor.org/info/rfc7296>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

8.2.  Informative References

   [C2PQ]     Hoffman, P., "The Transition from Classical to Post-
              Quantum Cryptography", Work in Progress, Internet-Draft,
              draft-hoffman-c2pq-07, 26 May 2020,

   [GROVER]   Grover, L., "A Fast Quantum Mechanical Algorithm for
              Database Search", STOC '96: Proceedings of the Twenty-
              Eighth Annual ACM Symposium on the Theory of Computing,
              pp. 212-219", DOI 10.1145/237814.237866, July 1996,

              IANA, "Internet Key Exchange Version 2 (IKEv2)

              NIST, "Submission Requirements and Evaluation Criteria for
              the Post-Quantum Cryptography Standardization Process",
              December 2016, <https://csrc.nist.gov/CSRC/media/Projects/

   [RFC2409]  Harkins, D. and D. Carrel, "The Internet Key Exchange
              (IKE)", RFC 2409, DOI 10.17487/RFC2409, November 1998,

   [RFC4648]  Josefsson, S., "The Base16, Base32, and Base64 Data
              Encodings", RFC 4648, DOI 10.17487/RFC4648, October 2006,

   [RFC6023]  Nir, Y., Tschofenig, H., Deng, H., and R. Singh, "A
              Childless Initiation of the Internet Key Exchange Version
              2 (IKEv2) Security Association (SA)", RFC 6023,
              DOI 10.17487/RFC6023, October 2010,

   [RFC6030]  Hoyer, P., Pei, M., and S. Machani, "Portable Symmetric
              Key Container (PSKC)", RFC 6030, DOI 10.17487/RFC6030,
              October 2010, <https://www.rfc-editor.org/info/rfc6030>.

   [RFC7619]  Smyslov, V. and P. Wouters, "The NULL Authentication
              Method in the Internet Key Exchange Protocol Version 2
              (IKEv2)", RFC 7619, DOI 10.17487/RFC7619, August 2015,

   [RFC8019]  Nir, Y. and V. Smyslov, "Protecting Internet Key Exchange
              Protocol Version 2 (IKEv2) Implementations from
              Distributed Denial-of-Service Attacks", RFC 8019,
              DOI 10.17487/RFC8019, November 2016,

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,

Appendix A.  Discussion and Rationale

   The primary goal of this document is to augment the IKEv2 protocol to
   provide protection against quantum computers without requiring novel
   cryptographic algorithms.  The idea behind this document is that
   while a quantum computer can easily reconstruct the shared secret of
   an (EC)DH exchange, it cannot as easily recover a secret from a
   symmetric exchange.  This document makes the SK_d (and thus also the
   IPsec KEYMAT and any subsequent IKE SA's SKEYSEED) depend on both the
   symmetric PPK and the Diffie-Hellman exchange.  If we assume that the
   attacker knows everything except the PPK during the key exchange and
   there are 2^(n) plausible PPKs, then a quantum computer (using
   Grover's algorithm) would take O(2^(n/2)) time to recover the PPK.
   So, even if the (EC)DH can be trivially solved, the attacker still
   can't recover any key material (except for the SK_ei, SK_er, SK_ai,
   and SK_ar values for the initial IKE exchange) unless they can find
   the PPK, which is too difficult if the PPK has enough entropy (for
   example, 256 bits).  Note that we do allow an attacker with a quantum
   computer to rederive the keying material for the initial IKE SA; this
   was a compromise to allow the responder to select the correct PPK

   Another goal of this protocol is to minimize the number of changes
   within the IKEv2 protocol, in particular, within the cryptography of
   IKEv2.  By limiting our changes to notifications and only adjusting
   the SK_d, SK_pi, and SK_pr, it is hoped that this would be
   implementable, even on systems that perform most of the IKEv2
   processing in hardware.

   A third goal is to be friendly to incremental deployment in
   operational networks for which we might not want to have a global
   shared key or for which quantum-secure IKEv2 is rolled out
   incrementally.  This is why we specifically try to allow the PPK to
   be dependent on the peer and why we allow the PPK to be configured as

   A fourth goal is to avoid violating any of the security properties
   provided by IKEv2.


   We would like to thank Tero Kivinen, Paul Wouters, Graham Bartlett,
   Tommy Pauly, Quynh Dang, and the rest of the IPSECME Working Group
   for their feedback and suggestions for the scheme.

Authors' Addresses

   Scott Fluhrer
   Cisco Systems

   Email: sfluhrer@cisco.com

   Panos Kampanakis
   Cisco Systems

   Email: pkampana@cisco.com

   David McGrew
   Cisco Systems

   Email: mcgrew@cisco.com

   Valery Smyslov

   Phone: +7 495 276 0211
   Email: svan@elvis.ru