4 Network Working Group C. Kaufman
5 Internet-Draft Microsoft
6 Obsoletes: 4306, 4718 P. Hoffman
7 (if approved) VPN Consortium
8 Intended status: Standards Track Y. Nir
9 Expires: January 9, 2010 Check Point
15 Internet Key Exchange Protocol: IKEv2
16 draft-ietf-ipsecme-ikev2bis-04
20 This Internet-Draft is submitted to IETF in full conformance with the
21 provisions of BCP 78 and BCP 79. This document may contain material
22 from IETF Documents or IETF Contributions published or made publicly
23 available before November 10, 2008. The person(s) controlling the
24 copyright in some of this material may not have granted the IETF
25 Trust the right to allow modifications of such material outside the
26 IETF Standards Process. Without obtaining an adequate license from
27 the person(s) controlling the copyright in such materials, this
28 document may not be modified outside the IETF Standards Process, and
29 derivative works of it may not be created outside the IETF Standards
30 Process, except to format it for publication as an RFC or to
31 translate it into languages other than English.
33 Internet-Drafts are working documents of the Internet Engineering
34 Task Force (IETF), its areas, and its working groups. Note that
35 other groups may also distribute working documents as Internet-
38 Internet-Drafts are draft documents valid for a maximum of six months
39 and may be updated, replaced, or obsoleted by other documents at any
40 time. It is inappropriate to use Internet-Drafts as reference
41 material or to cite them other than as "work in progress."
43 The list of current Internet-Drafts can be accessed at
44 http://www.ietf.org/ietf/1id-abstracts.txt.
46 The list of Internet-Draft Shadow Directories can be accessed at
47 http://www.ietf.org/shadow.html.
49 This Internet-Draft will expire on January 9, 2010.
55 Kaufman, et al. Expires January 9, 2010 [Page 1]
57 Internet-Draft IKEv2bis July 2009
60 Copyright (c) 2009 IETF Trust and the persons identified as the
61 document authors. All rights reserved.
63 This document is subject to BCP 78 and the IETF Trust's Legal
64 Provisions Relating to IETF Documents in effect on the date of
65 publication of this document (http://trustee.ietf.org/license-info).
66 Please review these documents carefully, as they describe your rights
67 and restrictions with respect to this document.
71 This document describes version 2 of the Internet Key Exchange (IKE)
72 protocol. IKE is a component of IPsec used for performing mutual
73 authentication and establishing and maintaining security associations
74 (SAs). It replaces and updates RFC 4306, and includes all of the
75 clarifications from RFC 4718.
111 Kaufman, et al. Expires January 9, 2010 [Page 2]
113 Internet-Draft IKEv2bis July 2009
118 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 6
119 1.1. Usage Scenarios . . . . . . . . . . . . . . . . . . . . . 7
120 1.1.1. Security Gateway to Security Gateway Tunnel Mode . . 8
121 1.1.2. Endpoint-to-Endpoint Transport Mode . . . . . . . . . 8
122 1.1.3. Endpoint to Security Gateway Tunnel Mode . . . . . . 9
123 1.1.4. Other Scenarios . . . . . . . . . . . . . . . . . . . 10
124 1.2. The Initial Exchanges . . . . . . . . . . . . . . . . . . 10
125 1.3. The CREATE_CHILD_SA Exchange . . . . . . . . . . . . . . 13
126 1.3.1. Creating New Child SAs with the CREATE_CHILD_SA
127 Exchange . . . . . . . . . . . . . . . . . . . . . . 14
128 1.3.2. Rekeying IKE SAs with the CREATE_CHILD_SA Exchange . 15
129 1.3.3. Rekeying Child SAs with the CREATE_CHILD_SA
130 Exchange . . . . . . . . . . . . . . . . . . . . . . 16
131 1.4. The INFORMATIONAL Exchange . . . . . . . . . . . . . . . 17
132 1.4.1. Deleting an SA with INFORMATIONAL Exchanges . . . . . 17
133 1.5. Informational Messages outside of an IKE SA . . . . . . . 18
134 1.6. Requirements Terminology . . . . . . . . . . . . . . . . 19
135 1.7. Differences Between RFC 4306 and This Document . . . . . 19
136 2. IKE Protocol Details and Variations . . . . . . . . . . . . . 21
137 2.1. Use of Retransmission Timers . . . . . . . . . . . . . . 21
138 2.2. Use of Sequence Numbers for Message ID . . . . . . . . . 23
139 2.3. Window Size for Overlapping Requests . . . . . . . . . . 23
140 2.4. State Synchronization and Connection Timeouts . . . . . . 25
141 2.5. Version Numbers and Forward Compatibility . . . . . . . . 27
142 2.6. IKE SA SPIs and Cookies . . . . . . . . . . . . . . . . . 28
143 2.6.1. Interaction of COOKIE and INVALID_KE_PAYLOAD . . . . 31
144 2.7. Cryptographic Algorithm Negotiation . . . . . . . . . . . 32
145 2.8. Rekeying . . . . . . . . . . . . . . . . . . . . . . . . 33
146 2.8.1. Simultaneous Child SA rekeying . . . . . . . . . . . 35
147 2.8.2. Simultaneous IKE SA Rekeying . . . . . . . . . . . . 37
148 2.8.3. Rekeying the IKE SA Versus Reauthentication . . . . . 38
149 2.9. Traffic Selector Negotiation . . . . . . . . . . . . . . 39
150 2.9.1. Traffic Selectors Violating Own Policy . . . . . . . 41
151 2.10. Nonces . . . . . . . . . . . . . . . . . . . . . . . . . 42
152 2.11. Address and Port Agility . . . . . . . . . . . . . . . . 42
153 2.12. Reuse of Diffie-Hellman Exponentials . . . . . . . . . . 43
154 2.13. Generating Keying Material . . . . . . . . . . . . . . . 43
155 2.14. Generating Keying Material for the IKE SA . . . . . . . . 45
156 2.15. Authentication of the IKE SA . . . . . . . . . . . . . . 45
157 2.16. Extensible Authentication Protocol Methods . . . . . . . 47
158 2.17. Generating Keying Material for Child SAs . . . . . . . . 49
159 2.18. Rekeying IKE SAs Using a CREATE_CHILD_SA Exchange . . . . 50
160 2.19. Requesting an Internal Address on a Remote Network . . . 51
161 2.19.1. Configuration Payloads . . . . . . . . . . . . . . . 53
162 2.20. Requesting the Peer's Version . . . . . . . . . . . . . . 54
163 2.21. Error Handling . . . . . . . . . . . . . . . . . . . . . 55
167 Kaufman, et al. Expires January 9, 2010 [Page 3]
169 Internet-Draft IKEv2bis July 2009
172 2.22. IPComp . . . . . . . . . . . . . . . . . . . . . . . . . 56
173 2.23. NAT Traversal . . . . . . . . . . . . . . . . . . . . . . 57
174 2.24. Explicit Congestion Notification (ECN) . . . . . . . . . 61
175 3. Header and Payload Formats . . . . . . . . . . . . . . . . . 61
176 3.1. The IKE Header . . . . . . . . . . . . . . . . . . . . . 61
177 3.2. Generic Payload Header . . . . . . . . . . . . . . . . . 64
178 3.3. Security Association Payload . . . . . . . . . . . . . . 66
179 3.3.1. Proposal Substructure . . . . . . . . . . . . . . . . 68
180 3.3.2. Transform Substructure . . . . . . . . . . . . . . . 70
181 3.3.3. Valid Transform Types by Protocol . . . . . . . . . . 73
182 3.3.4. Mandatory Transform IDs . . . . . . . . . . . . . . . 73
183 3.3.5. Transform Attributes . . . . . . . . . . . . . . . . 74
184 3.3.6. Attribute Negotiation . . . . . . . . . . . . . . . . 76
185 3.4. Key Exchange Payload . . . . . . . . . . . . . . . . . . 77
186 3.5. Identification Payloads . . . . . . . . . . . . . . . . . 78
187 3.6. Certificate Payload . . . . . . . . . . . . . . . . . . . 80
188 3.7. Certificate Request Payload . . . . . . . . . . . . . . . 82
189 3.8. Authentication Payload . . . . . . . . . . . . . . . . . 84
190 3.9. Nonce Payload . . . . . . . . . . . . . . . . . . . . . . 85
191 3.10. Notify Payload . . . . . . . . . . . . . . . . . . . . . 86
192 3.10.1. Notify Message Types . . . . . . . . . . . . . . . . 87
193 3.11. Delete Payload . . . . . . . . . . . . . . . . . . . . . 90
194 3.12. Vendor ID Payload . . . . . . . . . . . . . . . . . . . . 92
195 3.13. Traffic Selector Payload . . . . . . . . . . . . . . . . 93
196 3.13.1. Traffic Selector . . . . . . . . . . . . . . . . . . 94
197 3.14. Encrypted Payload . . . . . . . . . . . . . . . . . . . . 96
198 3.15. Configuration Payload . . . . . . . . . . . . . . . . . . 98
199 3.15.1. Configuration Attributes . . . . . . . . . . . . . . 99
200 3.15.2. Meaning of INTERNAL_IP4_SUBNET/INTERNAL_IP6_SUBNET . 102
201 3.15.3. Configuration payloads for IPv6 . . . . . . . . . . . 104
202 3.15.4. Address Assignment Failures . . . . . . . . . . . . . 104
203 3.16. Extensible Authentication Protocol (EAP) Payload . . . . 105
204 4. Conformance Requirements . . . . . . . . . . . . . . . . . . 107
205 5. Security Considerations . . . . . . . . . . . . . . . . . . . 109
206 5.1. Traffic selector authorization . . . . . . . . . . . . . 112
207 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 113
208 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 113
209 8. References . . . . . . . . . . . . . . . . . . . . . . . . . 114
210 8.1. Normative References . . . . . . . . . . . . . . . . . . 114
211 8.2. Informative References . . . . . . . . . . . . . . . . . 115
212 Appendix A. Summary of changes from IKEv1 . . . . . . . . . . . 119
213 Appendix B. Diffie-Hellman Groups . . . . . . . . . . . . . . . 120
214 B.1. Group 1 - 768 Bit MODP . . . . . . . . . . . . . . . . . 120
215 B.2. Group 2 - 1024 Bit MODP . . . . . . . . . . . . . . . . . 121
216 Appendix C. Exchanges and Payloads . . . . . . . . . . . . . . . 121
217 C.1. IKE_SA_INIT Exchange . . . . . . . . . . . . . . . . . . 122
218 C.2. IKE_AUTH Exchange without EAP . . . . . . . . . . . . . . 123
219 C.3. IKE_AUTH Exchange with EAP . . . . . . . . . . . . . . . 124
223 Kaufman, et al. Expires January 9, 2010 [Page 4]
225 Internet-Draft IKEv2bis July 2009
228 C.4. CREATE_CHILD_SA Exchange for Creating or Rekeying
229 Child SAs . . . . . . . . . . . . . . . . . . . . . . . . 125
230 C.5. CREATE_CHILD_SA Exchange for Rekeying the IKE SA . . . . 125
231 C.6. INFORMATIONAL Exchange . . . . . . . . . . . . . . . . . 125
232 Appendix D. Significant Changes from RFC 4306 . . . . . . . . . 125
233 Appendix E. Changes Between Internet Draft Versions . . . . . . 126
234 E.1. Changes from IKEv2 to draft -00 . . . . . . . . . . . . . 126
235 E.2. Changes from draft -00 to draft -01 . . . . . . . . . . . 126
236 E.3. Changes from draft -00 to draft -01 . . . . . . . . . . . 128
237 E.4. Changes from draft -01 to draft -02 . . . . . . . . . . . 129
238 E.5. Changes from draft -02 to draft -03 . . . . . . . . . . . 130
239 E.6. Changes from draft -03 to
240 draft-ietf-ipsecme-ikev2bis-00 . . . . . . . . . . . . . 131
241 E.7. Changes from draft-ietf-ipsecme-ikev2bis-00 to
242 draft-ietf-ipsecme-ikev2bis-01 . . . . . . . . . . . . . 132
243 E.8. Changes from draft-ietf-ipsecme-ikev2bis-01 to
244 draft-ietf-ipsecme-ikev2bis-02 . . . . . . . . . . . . . 136
245 E.9. Changes from draft-ietf-ipsecme-ikev2bis-01 to
246 draft-ietf-ipsecme-ikev2bis-02 . . . . . . . . . . . . . 138
247 E.10. Changes from draft-ietf-ipsecme-ikev2bis-02 to
248 draft-ietf-ipsecme-ikev2bis-03 . . . . . . . . . . . . . 139
249 E.11. Changes from draft-ietf-ipsecme-ikev2bis-03 to
250 draft-ietf-ipsecme-ikev2bis-04 . . . . . . . . . . . . . 139
251 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 140
279 Kaufman, et al. Expires January 9, 2010 [Page 5]
281 Internet-Draft IKEv2bis July 2009
286 {{ An introduction to the differences between RFC 4306 [IKEV2] and
287 this document is given at the end of Section 1. It is put there
288 (instead of here) to preserve the section numbering of RFC 4306. }}
290 IP Security (IPsec) provides confidentiality, data integrity, access
291 control, and data source authentication to IP datagrams. These
292 services are provided by maintaining shared state between the source
293 and the sink of an IP datagram. This state defines, among other
294 things, the specific services provided to the datagram, which
295 cryptographic algorithms will be used to provide the services, and
296 the keys used as input to the cryptographic algorithms.
298 Establishing this shared state in a manual fashion does not scale
299 well. Therefore, a protocol to establish this state dynamically is
300 needed. This memo describes such a protocol -- the Internet Key
301 Exchange (IKE). Version 1 of IKE was defined in RFCs 2407 [DOI],
302 2408 [ISAKMP], and 2409 [IKEV1]. IKEv2 replaced all of those RFCs.
303 IKEv2 was defined in [IKEV2] (RFC 4306) and was clarified in [Clarif]
304 (RFC 4718). This document replaces and updates RFC 4306 and RFC
305 4718. IKEv2 was a change to the IKE protocol that was not backward
306 compatible. In contrast, the current document not only provides a
307 clarification of IKEv2, but makes minimum changes to the IKE
310 IKE performs mutual authentication between two parties and
311 establishes an IKE security association (SA) that includes shared
312 secret information that can be used to efficiently establish SAs for
313 Encapsulating Security Payload (ESP) [ESP] or Authentication Header
314 (AH) [AH] and a set of cryptographic algorithms to be used by the SAs
315 to protect the traffic that they carry. In this document, the term
316 "suite" or "cryptographic suite" refers to a complete set of
317 algorithms used to protect an SA. An initiator proposes one or more
318 suites by listing supported algorithms that can be combined into
319 suites in a mix-and-match fashion. IKE can also negotiate use of IP
320 Compression (IPComp) [IP-COMP] in connection with an ESP or AH SA.
321 The SAs for ESP or AH that get set up through that IKE SA we call
324 All IKE communications consist of pairs of messages: a request and a
325 response. The pair is called an "exchange". We call the first
326 messages establishing an IKE SA IKE_SA_INIT and IKE_AUTH exchanges
327 and subsequent IKE exchanges CREATE_CHILD_SA or INFORMATIONAL
328 exchanges. In the common case, there is a single IKE_SA_INIT
329 exchange and a single IKE_AUTH exchange (a total of four messages) to
330 establish the IKE SA and the first Child SA. In exceptional cases,
331 there may be more than one of each of these exchanges. In all cases,
335 Kaufman, et al. Expires January 9, 2010 [Page 6]
337 Internet-Draft IKEv2bis July 2009
340 all IKE_SA_INIT exchanges MUST complete before any other exchange
341 type, then all IKE_AUTH exchanges MUST complete, and following that
342 any number of CREATE_CHILD_SA and INFORMATIONAL exchanges may occur
343 in any order. In some scenarios, only a single Child SA is needed
344 between the IPsec endpoints, and therefore there would be no
345 additional exchanges. Subsequent exchanges MAY be used to establish
346 additional Child SAs between the same authenticated pair of endpoints
347 and to perform housekeeping functions.
349 IKE message flow always consists of a request followed by a response.
350 It is the responsibility of the requester to ensure reliability. If
351 the response is not received within a timeout interval, the requester
352 needs to retransmit the request (or abandon the connection).
354 The first request/response of an IKE session (IKE_SA_INIT) negotiates
355 security parameters for the IKE SA, sends nonces, and sends Diffie-
358 The second request/response (IKE_AUTH) transmits identities, proves
359 knowledge of the secrets corresponding to the two identities, and
360 sets up an SA for the first (and often only) AH or ESP Child SA
361 (unless there is failure setting up the AH or ESP Child SA, in which
362 case the IKE SA is still established without IPsec SA).
364 The types of subsequent exchanges are CREATE_CHILD_SA (which creates
365 a Child SA) and INFORMATIONAL (which deletes an SA, reports error
366 conditions, or does other housekeeping). Every request requires a
367 response. An INFORMATIONAL request with no payloads (other than the
368 empty Encrypted payload required by the syntax) is commonly used as a
369 check for liveness. These subsequent exchanges cannot be used until
370 the initial exchanges have completed.
372 In the description that follows, we assume that no errors occur.
373 Modifications to the flow should errors occur are described in
378 IKE is expected to be used to negotiate ESP or AH SAs in a number of
379 different scenarios, each with its own special requirements.
391 Kaufman, et al. Expires January 9, 2010 [Page 7]
393 Internet-Draft IKEv2bis July 2009
396 1.1.1. Security Gateway to Security Gateway Tunnel Mode
398 +-+-+-+-+-+ +-+-+-+-+-+
400 Protected |Tunnel | tunnel |Tunnel | Protected
401 Subnet <-->|Endpoint |<---------->|Endpoint |<--> Subnet
403 +-+-+-+-+-+ +-+-+-+-+-+
405 Figure 1: Security Gateway to Security Gateway Tunnel
407 In this scenario, neither endpoint of the IP connection implements
408 IPsec, but network nodes between them protect traffic for part of the
409 way. Protection is transparent to the endpoints, and depends on
410 ordinary routing to send packets through the tunnel endpoints for
411 processing. Each endpoint would announce the set of addresses
412 "behind" it, and packets would be sent in tunnel mode where the inner
413 IP header would contain the IP addresses of the actual endpoints.
415 1.1.2. Endpoint-to-Endpoint Transport Mode
417 +-+-+-+-+-+ +-+-+-+-+-+
418 | | IPsec transport | |
419 |Protected| or tunnel mode SA |Protected|
420 |Endpoint |<---------------------------------------->|Endpoint |
422 +-+-+-+-+-+ +-+-+-+-+-+
424 Figure 2: Endpoint to Endpoint
426 In this scenario, both endpoints of the IP connection implement
427 IPsec, as required of hosts in [IPSECARCH]. Transport mode will
428 commonly be used with no inner IP header. A single pair of addresses
429 will be negotiated for packets to be protected by this SA. These
430 endpoints MAY implement application layer access controls based on
431 the IPsec authenticated identities of the participants. This
432 scenario enables the end-to-end security that has been a guiding
433 principle for the Internet since [ARCHPRINC], [TRANSPARENCY], and a
434 method of limiting the inherent problems with complexity in networks
435 noted by [ARCHGUIDEPHIL]. Although this scenario may not be fully
436 applicable to the IPv4 Internet, it has been deployed successfully in
437 specific scenarios within intranets using IKEv1. It should be more
438 broadly enabled during the transition to IPv6 and with the adoption
441 It is possible in this scenario that one or both of the protected
442 endpoints will be behind a network address translation (NAT) node, in
443 which case the tunneled packets will have to be UDP encapsulated so
447 Kaufman, et al. Expires January 9, 2010 [Page 8]
449 Internet-Draft IKEv2bis July 2009
452 that port numbers in the UDP headers can be used to identify
453 individual endpoints "behind" the NAT (see Section 2.23).
455 1.1.3. Endpoint to Security Gateway Tunnel Mode
457 +-+-+-+-+-+ +-+-+-+-+-+
458 | | IPsec | | Protected
459 |Protected| tunnel |Tunnel | Subnet
460 |Endpoint |<------------------------>|Endpoint |<--- and/or
462 +-+-+-+-+-+ +-+-+-+-+-+
464 Figure 3: Endpoint to Security Gateway Tunnel
466 In this scenario, a protected endpoint (typically a portable roaming
467 computer) connects back to its corporate network through an IPsec-
468 protected tunnel. It might use this tunnel only to access
469 information on the corporate network, or it might tunnel all of its
470 traffic back through the corporate network in order to take advantage
471 of protection provided by a corporate firewall against Internet-based
472 attacks. In either case, the protected endpoint will want an IP
473 address associated with the security gateway so that packets returned
474 to it will go to the security gateway and be tunneled back. This IP
475 address may be static or may be dynamically allocated by the security
476 gateway. In support of the latter case, IKEv2 includes a mechanism
477 (namely, configuration payloads) for the initiator to request an IP
478 address owned by the security gateway for use for the duration of its
481 In this scenario, packets will use tunnel mode. On each packet from
482 the protected endpoint, the outer IP header will contain the source
483 IP address associated with its current location (i.e., the address
484 that will get traffic routed to the endpoint directly), while the
485 inner IP header will contain the source IP address assigned by the
486 security gateway (i.e., the address that will get traffic routed to
487 the security gateway for forwarding to the endpoint). The outer
488 destination address will always be that of the security gateway,
489 while the inner destination address will be the ultimate destination
492 In this scenario, it is possible that the protected endpoint will be
493 behind a NAT. In that case, the IP address as seen by the security
494 gateway will not be the same as the IP address sent by the protected
495 endpoint, and packets will have to be UDP encapsulated in order to be
503 Kaufman, et al. Expires January 9, 2010 [Page 9]
505 Internet-Draft IKEv2bis July 2009
508 1.1.4. Other Scenarios
510 Other scenarios are possible, as are nested combinations of the
511 above. One notable example combines aspects of 1.1.1 and 1.1.3. A
512 subnet may make all external accesses through a remote security
513 gateway using an IPsec tunnel, where the addresses on the subnet are
514 routed to the security gateway by the rest of the Internet. An
515 example would be someone's home network being virtually on the
516 Internet with static IP addresses even though connectivity is
517 provided by an ISP that assigns a single dynamically assigned IP
518 address to the user's security gateway (where the static IP addresses
519 and an IPsec relay are provided by a third party located elsewhere).
521 1.2. The Initial Exchanges
523 Communication using IKE always begins with IKE_SA_INIT and IKE_AUTH
524 exchanges (known in IKEv1 as Phase 1). These initial exchanges
525 normally consist of four messages, though in some scenarios that
526 number can grow. All communications using IKE consist of request/
527 response pairs. We'll describe the base exchange first, followed by
528 variations. The first pair of messages (IKE_SA_INIT) negotiate
529 cryptographic algorithms, exchange nonces, and do a Diffie-Hellman
532 The second pair of messages (IKE_AUTH) authenticate the previous
533 messages, exchange identities and certificates, and establish the
534 first Child SA. Parts of these messages are encrypted and integrity
535 protected with keys established through the IKE_SA_INIT exchange, so
536 the identities are hidden from eavesdroppers and all fields in all
537 the messages are authenticated. (See Section 2.14 for information on
538 how the encryption keys are generated.)
540 All messages following the initial exchange are cryptographically
541 protected using the cryptographic algorithms and keys negotiated in
542 the the IKE_SA_INIT exchange. These subsequent messages use the
543 syntax of the Encrypted Payload described in Section 3.14, encrypted
544 with keys that are derived as described in Section 2.14. All
545 subsequent messages include an Encrypted Payload, even if they are
546 referred to in the text as "empty". For the CREATE_CHILD_SA,
547 IKE_AUTH, or IKE_INFORMATIONAL exchanges, the message following the
548 header is encrypted and the message including the header is integrity
549 protected using the cryptographic algorithms negotiated for the IKE
552 In the following descriptions, the payloads contained in the message
553 are indicated by names as listed below.
559 Kaufman, et al. Expires January 9, 2010 [Page 10]
561 Internet-Draft IKEv2bis July 2009
565 -----------------------------------------
568 CERTREQ Certificate Request
572 EAP Extensible Authentication
574 IDi Identification - Initiator
575 IDr Identification - Responder
579 SA Security Association
580 TSi Traffic Selector - Initiator
581 TSr Traffic Selector - Responder
584 The details of the contents of each payload are described in section
585 3. Payloads that may optionally appear will be shown in brackets,
586 such as [CERTREQ], indicate that optionally a certificate request
587 payload can be included.
589 The initial exchanges are as follows:
592 -------------------------------------------------------------------
593 HDR, SAi1, KEi, Ni -->
595 HDR contains the Security Parameter Indexes (SPIs), version numbers,
596 and flags of various sorts. The SAi1 payload states the
597 cryptographic algorithms the initiator supports for the IKE SA. The
598 KE payload sends the initiator's Diffie-Hellman value. Ni is the
601 <-- HDR, SAr1, KEr, Nr, [CERTREQ]
603 The responder chooses a cryptographic suite from the initiator's
604 offered choices and expresses that choice in the SAr1 payload,
605 completes the Diffie-Hellman exchange with the KEr payload, and sends
606 its nonce in the Nr payload.
608 At this point in the negotiation, each party can generate SKEYSEED,
609 from which all keys are derived for that IKE SA. The messages that
610 follow are encrypted and integrity protected in their entirety, with
611 the exception of the message headers. The keys used for the
615 Kaufman, et al. Expires January 9, 2010 [Page 11]
617 Internet-Draft IKEv2bis July 2009
620 encryption and integrity protection are derived from SKEYSEED and are
621 known as SK_e (encryption) and SK_a (authentication, a.k.a. integrity
622 protection). A separate SK_e and SK_a is computed for each
623 direction. In addition to the keys SK_e and SK_a derived from the DH
624 value for protection of the IKE SA, another quantity SK_d is derived
625 and used for derivation of further keying material for Child SAs.
626 The notation SK { ... } indicates that these payloads are encrypted
627 and integrity protected using that direction's SK_e and SK_a.
629 HDR, SK {IDi, [CERT,] [CERTREQ,]
633 The initiator asserts its identity with the IDi payload, proves
634 knowledge of the secret corresponding to IDi and integrity protects
635 the contents of the first message using the AUTH payload (see
636 Section 2.15). It might also send its certificate(s) in CERT
637 payload(s) and a list of its trust anchors in CERTREQ payload(s). If
638 any CERT payloads are included, the first certificate provided MUST
639 contain the public key used to verify the AUTH field.
641 The optional payload IDr enables the initiator to specify which of
642 the responder's identities it wants to talk to. This is useful when
643 the machine on which the responder is running is hosting multiple
644 identities at the same IP address. If the IDr proposed by the
645 initiator is not acceptable to the responder, the responder might use
646 some other IDr to finish the exchange. If the initiator then does
647 not accept that fact that responder used different IDr than the one
648 that was requested, the initiator can close the SA after noticing the
651 The initiator begins negotiation of a Child SA using the SAi2
652 payload. The final fields (starting with SAi2) are described in the
653 description of the CREATE_CHILD_SA exchange.
655 <-- HDR, SK {IDr, [CERT,] AUTH,
658 The responder asserts its identity with the IDr payload, optionally
659 sends one or more certificates (again with the certificate containing
660 the public key used to verify AUTH listed first), authenticates its
661 identity and protects the integrity of the second message with the
662 AUTH payload, and completes negotiation of a Child SA with the
663 additional fields described below in the CREATE_CHILD_SA exchange.
665 The recipients of messages 3 and 4 MUST verify that all signatures
666 and MACs are computed correctly and that the names in the ID payloads
667 correspond to the keys used to generate the AUTH payload.
671 Kaufman, et al. Expires January 9, 2010 [Page 12]
673 Internet-Draft IKEv2bis July 2009
676 If creating the Child SA during the IKE_AUTH exchange fails for some
677 reason, the IKE SA is still created as usual. The list of responses
678 in the IKE_AUTH exchange that do not prevent an IKE SA from being set
679 up include at least the following: NO_PROPOSAL_CHOSEN,
680 TS_UNACCEPTABLE, SINGLE_PAIR_REQUIRED, INTERNAL_ADDRESS_FAILURE, and
683 Note that IKE_AUTH messages do not contain KEi/KEr or Ni/Nr payloads.
684 Thus, the SA payloads in the IKE_AUTH exchange cannot contain
685 Transform Type 4 (Diffie-Hellman Group) with any value other than
686 NONE. Implementations SHOULD omit the whole transform substructure
687 instead of sending value NONE.
689 1.3. The CREATE_CHILD_SA Exchange
691 The CREATE_CHILD_SA exchange is used to create new Child SAs and to
692 rekey both IKE SAs and Child SAs. This exchange consists of a single
693 request/response pair, and some of its function was referred to as a
694 phase 2 exchange in IKEv1. It MAY be initiated by either end of the
695 IKE SA after the initial exchanges are completed.
697 All messages following the initial exchange are cryptographically
698 protected using the cryptographic algorithms and keys negotiated in
699 the first two messages of the IKE exchange. These subsequent
700 messages use the syntax of the Encrypted Payload described in
701 Section 3.14, encrypted with keys that are derived as described in
702 Section 2.14. All subsequent messages include an Encrypted Payload,
703 even if they are referred to in the text as "empty". For both
704 messages in the CREATE_CHILD_SA, the message following the header is
705 encrypted and the message including the header is integrity protected
706 using the cryptographic algorithms negotiated for the IKE SA.
708 The CREATE_CHILD_SA is also used for rekeying IKE SAs and Child SAs.
709 An SA is rekeyed by creating a new SA and then deleting the old one.
710 This section describes the first part of rekeying, the creation of
711 new SAs; Section 2.8 covers the mechanics of rekeying, including
712 moving traffic from old to new SAs and the deletion of the old SAs.
713 The two sections must be read together to understand the entire
716 Either endpoint may initiate a CREATE_CHILD_SA exchange, so in this
717 section the term initiator refers to the endpoint initiating this
718 exchange. An implementation MAY refuse all CREATE_CHILD_SA requests
721 The CREATE_CHILD_SA request MAY optionally contain a KE payload for
722 an additional Diffie-Hellman exchange to enable stronger guarantees
723 of forward secrecy for the Child SA. The keying material for the
727 Kaufman, et al. Expires January 9, 2010 [Page 13]
729 Internet-Draft IKEv2bis July 2009
732 Child SA is a function of SK_d established during the establishment
733 of the IKE SA, the nonces exchanged during the CREATE_CHILD_SA
734 exchange, and the Diffie-Hellman value (if KE payloads are included
735 in the CREATE_CHILD_SA exchange).
737 If a CREATE_CHILD_SA exchange includes a KEi payload, at least one of
738 the SA offers MUST include the Diffie-Hellman group of the KEi. The
739 Diffie-Hellman group of the KEi MUST be an element of the group the
740 initiator expects the responder to accept (additional Diffie-Hellman
741 groups can be proposed). If the responder selects a proposal using a
742 different Diffie-Hellman group (other than NONE), the responder MUST
743 reject the request and indicate its preferred Diffie-Hellman group in
744 the INVALID_KE_PAYLOAD Notification payload. There are two octets of
745 data associated with this notification: the accepted D-H Group number
746 in big endian order. In the case of such a rejection, the
747 CREATE_CHILD_SA exchange fails, and the initiator will probably retry
748 the exchange with a Diffie-Hellman proposal and KEi in the group that
749 the responder gave in the INVALID_KE_PAYLOAD.
751 The responder sends a NO_ADDITIONAL_SAS notification to indicate that
752 a CREATE_CHILD_SA request is unacceptable because the responder is
753 unwilling to accept any more Child SAs on this IKE SA. Some minimal
754 implementations may only accept a single Child SA setup in the
755 context of an initial IKE exchange and reject any subsequent attempts
758 1.3.1. Creating New Child SAs with the CREATE_CHILD_SA Exchange
760 A Child SA may be created by sending a CREATE_CHILD_SA request. The
761 CREATE_CHILD_SA request for creating a new Child SA is:
764 -------------------------------------------------------------------
765 HDR, SK {SA, Ni, [KEi],
768 The initiator sends SA offer(s) in the SA payload, a nonce in the Ni
769 payload, optionally a Diffie-Hellman value in the KEi payload, and
770 the proposed traffic selectors for the proposed Child SA in the TSi
773 The CREATE_CHILD_SA response for creating a new Child SA is:
775 <-- HDR, SK {SA, Nr, [KEr],
778 The responder replies (using the same Message ID to respond) with the
779 accepted offer in an SA payload, and a Diffie-Hellman value in the
783 Kaufman, et al. Expires January 9, 2010 [Page 14]
785 Internet-Draft IKEv2bis July 2009
788 KEr payload if KEi was included in the request and the selected
789 cryptographic suite includes that group.
791 The traffic selectors for traffic to be sent on that SA are specified
792 in the TS payloads in the response, which may be a subset of what the
793 initiator of the Child SA proposed.
795 The USE_TRANSPORT_MODE notification MAY be included in a request
796 message that also includes an SA payload requesting a Child SA. It
797 requests that the Child SA use transport mode rather than tunnel mode
798 for the SA created. If the request is accepted, the response MUST
799 also include a notification of type USE_TRANSPORT_MODE. If the
800 responder declines the request, the Child SA will be established in
801 tunnel mode. If this is unacceptable to the initiator, the initiator
802 MUST delete the SA. Note: Except when using this option to negotiate
803 transport mode, all Child SAs will use tunnel mode.
805 The ESP_TFC_PADDING_NOT_SUPPORTED notification asserts that the
806 sending endpoint will NOT accept packets that contain Traffic Flow
807 Confidentiality (TFC) padding over the Child SA being negotiated. If
808 neither endpoint accepts TFC padding, this notification is included
809 in both the request and the response. If this notification is
810 included in only one of the messages, TFC padding can still be sent
811 in the other direction.
813 The NON_FIRST_FRAGMENTS_ALSO notification is used for fragmentation
814 control. See [IPSECARCH] for a fuller explanation. Both parties
815 need to agree to sending non-first fragments before either party does
816 so. It is enabled only if NON_FIRST_FRAGMENTS_ALSO notification is
817 included in both the request proposing an SA and the response
818 accepting it. If the responder does not want to send or receive non-
819 first fragments, it only omits NON_FIRST_FRAGMENTS_ALSO notification
820 from its response, but does not reject the whole Child SA creation.
822 Failure of an attempt to create a CHILD SA SHOULD NOT tear down the
823 IKE SA: there is no reason to lose the work done to set up the IKE
824 SA. When an IKE SA is not created, the error message return SHOULD
825 NOT be encrypted because the other party will not be able to
826 authenticate that message. [[ Note: this text may be changed in the
829 1.3.2. Rekeying IKE SAs with the CREATE_CHILD_SA Exchange
831 The CREATE_CHILD_SA request for rekeying an IKE SA is:
834 -------------------------------------------------------------------
835 HDR, SK {SA, Ni, KEi} -->
839 Kaufman, et al. Expires January 9, 2010 [Page 15]
841 Internet-Draft IKEv2bis July 2009
844 The initiator sends SA offer(s) in the SA payload, a nonce in the Ni
845 payload, and a Diffie-Hellman value in the KEi payload. The KEi
846 payload MUST be included. New initiator and responder SPIs are
847 supplied in the SPI fields of the SA payload.
849 The CREATE_CHILD_SA response for rekeying an IKE SA is:
851 <-- HDR, SK {SA, Nr,[KEr]}
853 The responder replies (using the same Message ID to respond) with the
854 accepted offer in an SA payload, and a Diffie-Hellman value in the
855 KEr payload if the selected cryptographic suite includes that group.
857 The new IKE SA has its message counters set to 0, regardless of what
858 they were in the earlier IKE SA. The first IKE requests from both
859 sides on the new IKE SA will have message ID 0. The old IKE SA
860 retains its numbering, so any further requests (for example, to
861 delete the IKE SA) will have consecutive numbering. The new IKE SA
862 also has its window size reset to 1, and the initiator in this rekey
863 exchange is the new "original initiator" of the new IKE SA.
865 1.3.3. Rekeying Child SAs with the CREATE_CHILD_SA Exchange
867 The CREATE_CHILD_SA request for rekeying a Child SA is:
870 -------------------------------------------------------------------
871 HDR, SK {N, SA, Ni, [KEi],
874 The initiator sends SA offer(s) in the SA payload, a nonce in the Ni
875 payload, optionally a Diffie-Hellman value in the KEi payload, and
876 the proposed traffic selectors for the proposed Child SA in the TSi
879 The REKEY_SA notification MUST be included in a CREATE_CHILD_SA
880 exchange if the purpose of the exchange is to replace an existing ESP
881 or AH SA. The SA being rekeyed is identified by the SPI field in the
882 Notify payload; this is the SPI the exchange initiator would expect
883 in inbound ESP or AH packets. There is no data associated with this
884 Notify type. The Protocol ID field of the REKEY_SA notification is
885 set to match the protocol of the SA we are rekeying, for example, 3
886 for ESP and 2 for AH.
888 The CREATE_CHILD_SA response for rekeying a Child SA is:
890 <-- HDR, SK {SA, Nr, [KEr],
895 Kaufman, et al. Expires January 9, 2010 [Page 16]
897 Internet-Draft IKEv2bis July 2009
900 The responder replies (using the same Message ID to respond) with the
901 accepted offer in an SA payload, and a Diffie-Hellman value in the
902 KEr payload if KEi was included in the request and the selected
903 cryptographic suite includes that group.
905 The traffic selectors for traffic to be sent on that SA are specified
906 in the TS payloads in the response, which may be a subset of what the
907 initiator of the Child SA proposed.
909 1.4. The INFORMATIONAL Exchange
911 At various points during the operation of an IKE SA, peers may desire
912 to convey control messages to each other regarding errors or
913 notifications of certain events. To accomplish this, IKE defines an
914 INFORMATIONAL exchange. INFORMATIONAL exchanges MUST ONLY occur
915 after the initial exchanges and are cryptographically protected with
918 Control messages that pertain to an IKE SA MUST be sent under that
919 IKE SA. Control messages that pertain to Child SAs MUST be sent
920 under the protection of the IKE SA which generated them (or its
921 successor if the IKE SA was rekeyed).
923 Messages in an INFORMATIONAL exchange contain zero or more
924 Notification, Delete, and Configuration payloads. The Recipient of
925 an INFORMATIONAL exchange request MUST send some response (else the
926 Sender will assume the message was lost in the network and will
927 retransmit it). That response MAY be a message with no payloads.
928 The request message in an INFORMATIONAL exchange MAY also contain no
929 payloads. This is the expected way an endpoint can ask the other
930 endpoint to verify that it is alive.
932 The INFORMATIONAL exchange is defined as:
935 -------------------------------------------------------------------
938 <-- HDR, SK {[N,] [D,]
941 The processing of an INFORMATIONAL exchange is determined by its
944 1.4.1. Deleting an SA with INFORMATIONAL Exchanges
946 ESP and AH SAs always exist in pairs, with one SA in each direction.
947 When an SA is closed, both members of the pair MUST be closed (that
951 Kaufman, et al. Expires January 9, 2010 [Page 17]
953 Internet-Draft IKEv2bis July 2009
956 is, deleted). Each endpoint MUST close its incoming SAs and allow
957 the other endpoint to close the other SA in each pair. To delete an
958 SA, an INFORMATIONAL exchange with one or more delete payloads is
959 sent listing the SPIs (as they would be expected in the headers of
960 inbound packets) of the SAs to be deleted. The recipient MUST close
961 the designated SAs. Note that one never sends delete payloads for
962 the two sides of an SA in a single message. If there are many SAs to
963 delete at the same time, one includes delete payloads for the inbound
964 half of each SA pair in your Informational exchange.
966 Normally, the reply in the INFORMATIONAL exchange will contain delete
967 payloads for the paired SAs going in the other direction. There is
968 one exception. If by chance both ends of a set of SAs independently
969 decide to close them, each may send a delete payload and the two
970 requests may cross in the network. If a node receives a delete
971 request for SAs for which it has already issued a delete request, it
972 MUST delete the outgoing SAs while processing the request and the
973 incoming SAs while processing the response. In that case, the
974 responses MUST NOT include delete payloads for the deleted SAs, since
975 that would result in duplicate deletion and could in theory delete
978 Half-closed ESP or AH connections are anomalous, and a node with
979 auditing capability should probably audit their existence if they
980 persist. Note that this specification nowhere specifies time
981 periods, so it is up to individual endpoints to decide how long to
982 wait. A node MAY refuse to accept incoming data on half-closed
983 connections but MUST NOT unilaterally close them and reuse the SPIs.
984 If connection state becomes sufficiently messed up, a node MAY close
985 the IKE SA; doing so will implicitly close all SAs negotiated under
986 it. It can then rebuild the SAs it needs on a clean base under a new
987 IKE SA. The response to a request that deletes the IKE SA is an
988 empty Informational response.
990 1.5. Informational Messages outside of an IKE SA
992 If an encrypted IKE request packet arrives on port 500 or 4500 with
993 an unrecognized SPI, it could be because the receiving node has
994 recently crashed and lost state or because of some other system
995 malfunction or attack. If the receiving node has an active IKE SA to
996 the IP address from whence the packet came, it MAY send a
997 notification of the wayward packet over that IKE SA in an
998 INFORMATIONAL exchange. If it does not have such an IKE SA, it MAY
999 send an Informational message without cryptographic protection to the
1000 source IP address. Such a message is not part of an informational
1001 exchange, and the receiving node MUST NOT respond to it. Doing so
1002 could cause a message loop.
1007 Kaufman, et al. Expires January 9, 2010 [Page 18]
1009 Internet-Draft IKEv2bis July 2009
1012 The INVALID_SPI notification MAY be sent in an IKE INFORMATIONAL
1013 exchange when a node receives an ESP or AH packet with an invalid
1014 SPI. The Notification Data contains the SPI of the invalid packet.
1015 This usually indicates a node has rebooted and forgotten an SA. If
1016 this Informational Message is sent outside the context of an IKE SA,
1017 it should only be used by the recipient as a "hint" that something
1018 might be wrong (because it could easily be forged). The recipient of
1019 this notification cannot tell whether the SPI is for AH or ESP, but
1020 this is not important because the SPIs are supposed to be different
1023 There are two cases when such a one-way notification is sent:
1024 INVALID_IKE_SPI and INVALID_SPI. These notifications are sent
1025 outside of an IKE SA. Note that such notifications are explicitly
1026 not Informational exchanges; these are one-way messages that must not
1027 be responded to. (INVALID_MAJOR_VERSION is also a one-way message
1028 which is sent outside of an IKE SA, although it is sent as a response
1029 to the incoming IKE SA creation.)
1031 In case of INVALID_IKE_SPI, the message sent is a response message,
1032 and thus it is sent to the IP address and port from whence it came
1033 with the same IKE SPIs and the Message ID is copied. The Response
1034 bit is set to 1, and the version flags are set in the normal fashion.
1035 For a one-way INVALID_IKE_SPI notification for an unrecognized SPI,
1036 the responder SHOULD copy the Exchange Type from the request.
1038 In case of INVALID_SPI, however, there are no IKE SPI values that
1039 would be meaningful to the recipient of such a notification. Using
1040 zero values or random values are both acceptable. The Initiator flag
1041 is set, the Response bit is set to 0, and the version flags are set
1042 in the normal fashion.
1044 1.6. Requirements Terminology
1046 Definitions of the primitive terms in this document (such as Security
1047 Association or SA) can be found in [IPSECARCH]. It should be noted
1048 that parts of IKEv2 rely on some of the processing rules in
1049 [IPSECARCH], as described in various sections of this document.
1051 Keywords "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT" and
1052 "MAY" that appear in this document are to be interpreted as described
1055 1.7. Differences Between RFC 4306 and This Document
1057 {{ Added this entire section, including this recursive remark. }}
1059 This document contains clarifications and amplifications to IKEv2
1063 Kaufman, et al. Expires January 9, 2010 [Page 19]
1065 Internet-Draft IKEv2bis July 2009
1068 [IKEV2]. The clarifications are mostly based on [Clarif]. The
1069 changes listed in that document were discussed in the IPsec Working
1070 Group and, after the Working Group was disbanded, on the IPsec
1071 mailing list. That document contains detailed explanations of areas
1072 that were unclear in IKEv2, and is thus useful to implementers of
1075 The protocol described in this document retains the same major
1076 version number (2) and minor version number (0) as was used in RFC
1077 4306. That is, the version number is *not* changed from RFC 4306.
1079 This document makes the figures and references a bit more regular
1082 IKEv2 developers have noted that the SHOULD-level requirements are
1083 often unclear in that they don't say when it is OK to not obey the
1084 requirements. They also have noted that there are MUST-level
1085 requirements that are not related to interoperability. This document
1086 has more explanation of some of these requirements. All non-
1087 capitalized uses of the words SHOULD and MUST now mean their normal
1088 English sense, not the interoperability sense of [MUSTSHOULD].
1090 IKEv2 (and IKEv1) developers have noted that there is a great deal of
1091 material in the tables of codes in Section 3.10.1. This leads to
1092 implementers not having all the needed information in the main body
1093 of the document. Much of the material from those tables has been
1094 moved into the associated parts of the main body of the document.
1096 This document removes discussion of nesting AH and ESP. This was a
1097 mistake in RFC 4306 caused by the lag between finishing RFC 4306 and
1098 RFC 4301. Basically, IKEv2 is based on RFC 4301, which does not
1099 include "SA bundles" that were part of RFC 2401. While a single
1100 packet can go through IPsec processing multiple times, each of these
1101 passes uses a separate SA, and the passes are coordinated by the
1102 forwarding tables. In IKEv2, each of these SAs has to be created
1103 using a separate CREATE_CHILD_SA exchange.
1105 This document removes discussion of the INTERNAL_ADDRESS_EXPIRY
1106 configuration attribute because its implementation was very
1107 problematic. Implementations that conform to this document MUST
1108 ignore proposals that have configuration attribute type 5, the old
1109 value for INTERNAL_ADDRESS_EXPIRY.
1111 This document adds the restriction in Section 2.13 that all PRFs used
1112 with IKEv2 MUST take variable-sized keys. This should not affect any
1113 implementations because there were no standardized PRFs that have
1119 Kaufman, et al. Expires January 9, 2010 [Page 20]
1121 Internet-Draft IKEv2bis July 2009
1124 2. IKE Protocol Details and Variations
1126 IKE normally listens and sends on UDP port 500, though IKE messages
1127 may also be received on UDP port 4500 with a slightly different
1128 format (see Section 2.23). Since UDP is a datagram (unreliable)
1129 protocol, IKE includes in its definition recovery from transmission
1130 errors, including packet loss, packet replay, and packet forgery.
1131 IKE is designed to function so long as (1) at least one of a series
1132 of retransmitted packets reaches its destination before timing out;
1133 and (2) the channel is not so full of forged and replayed packets so
1134 as to exhaust the network or CPU capacities of either endpoint. Even
1135 in the absence of those minimum performance requirements, IKE is
1136 designed to fail cleanly (as though the network were broken).
1138 Although IKEv2 messages are intended to be short, they contain
1139 structures with no hard upper bound on size (in particular, X.509
1140 certificates), and IKEv2 itself does not have a mechanism for
1141 fragmenting large messages. IP defines a mechanism for fragmentation
1142 of oversize UDP messages, but implementations vary in the maximum
1143 message size supported. Furthermore, use of IP fragmentation opens
1144 an implementation to denial of service attacks [DOSUDPPROT].
1145 Finally, some NAT and/or firewall implementations may block IP
1148 All IKEv2 implementations MUST be able to send, receive, and process
1149 IKE messages that are up to 1280 octets long, and they SHOULD be able
1150 to send, receive, and process messages that are up to 3000 octets
1151 long. IKEv2 implementations need to be aware of the maximum UDP
1152 message size supported and MAY shorten messages by leaving out some
1153 certificates or cryptographic suite proposals if that will keep
1154 messages below the maximum. Use of the "Hash and URL" formats rather
1155 than including certificates in exchanges where possible can avoid
1156 most problems. Implementations and configuration need to keep in
1157 mind, however, that if the URL lookups are possible only after the
1158 IPsec SA is established, recursion issues could prevent this
1159 technique from working.
1161 The UDP payload of all packets containing IKE messages sent on port
1162 4500 MUST begin with the prefix of four zeros; otherwise, the
1163 receiver won't know how to handle them.
1165 2.1. Use of Retransmission Timers
1167 All messages in IKE exist in pairs: a request and a response. The
1168 setup of an IKE SA normally consists of two request/response pairs.
1169 Once the IKE SA is set up, either end of the security association may
1170 initiate requests at any time, and there can be many requests and
1171 responses "in flight" at any given moment. But each message is
1175 Kaufman, et al. Expires January 9, 2010 [Page 21]
1177 Internet-Draft IKEv2bis July 2009
1180 labeled as either a request or a response, and for each request/
1181 response pair one end of the security association is the initiator
1182 and the other is the responder.
1184 For every pair of IKE messages, the initiator is responsible for
1185 retransmission in the event of a timeout. The responder MUST never
1186 retransmit a response unless it receives a retransmission of the
1187 request. In that event, the responder MUST ignore the retransmitted
1188 request except insofar as it triggers a retransmission of the
1189 response. The initiator MUST remember each request until it receives
1190 the corresponding response. The responder MUST remember each
1191 response until it receives a request whose sequence number is larger
1192 than or equal to the sequence number in the response plus its window
1193 size (see Section 2.3). In order to allow saving memory, responders
1194 are allowed to forget response after a timeout of several minutes.
1195 If the responder receives a retransmitted request for which it has
1196 already forgotten the response, it MUST ignore the request (and not,
1197 for example, attempt constructing a new response).
1199 IKE is a reliable protocol, in the sense that the initiator MUST
1200 retransmit a request until either it receives a corresponding reply
1201 OR it deems the IKE security association to have failed and it
1202 discards all state associated with the IKE SA and any Child SAs
1203 negotiated using that IKE SA. A retransmission from the initiator
1204 MUST be bitwise identical to the original request. That is,
1205 everything starting from the IKE Header (the IKE SA Initiator's SPI
1206 onwards) must be bitwise identical; items before it (such as the IP
1207 and UDP headers, and the zero non-ESP marker) do not have to be
1210 Retransmissions of the IKE_SA_INIT request require some special
1211 handling. When a responder receives an IKE_SA_INIT request, it has
1212 to determine whether the packet is a retransmission belonging to an
1213 existing "half-open" IKE SA (in which case the responder retransmits
1214 the same response), or a new request (in which case the responder
1215 creates a new IKE SA and sends a fresh response), or it belongs to an
1216 existing IKE SA where the IKE_AUTH request has been already received
1217 (in which case the responder ignores it).
1219 It is not sufficient to use the initiator's SPI and/or IP address to
1220 differentiate between these three cases because two different peers
1221 behind a single NAT could choose the same initiator SPI. Instead, a
1222 robust responder will do the IKE SA lookup using the whole packet,
1223 its hash, or the Ni payload.
1231 Kaufman, et al. Expires January 9, 2010 [Page 22]
1233 Internet-Draft IKEv2bis July 2009
1236 2.2. Use of Sequence Numbers for Message ID
1238 Every IKE message contains a Message ID as part of its fixed header.
1239 This Message ID is used to match up requests and responses, and to
1240 identify retransmissions of messages.
1242 The Message ID is a 32-bit quantity, which is zero for the
1243 IKE_SA_INIT messages (including retries of the message due to
1244 responses such as COOKIE and INVALID_KE_PAYLOAD), and incremented for
1245 each subsequent exchange. The Message ID is reset to zero in the new
1246 IKE SA after the IKE SA is rekeyed. Rekeying an IKE SA resets the
1247 sequence numbers. Thus, the first pair of IKE_AUTH messages will
1248 have ID of 1, the second (when EAP is used) will be 2, and so on.
1250 Each endpoint in the IKE Security Association maintains two "current"
1251 Message IDs: the next one to be used for a request it initiates and
1252 the next one it expects to see in a request from the other end.
1253 These counters increment as requests are generated and received.
1254 Responses always contain the same message ID as the corresponding
1255 request. That means that after the initial exchange, each integer n
1256 may appear as the message ID in four distinct messages: the nth
1257 request from the original IKE initiator, the corresponding response,
1258 the nth request from the original IKE responder, and the
1259 corresponding response. If the two ends make very different numbers
1260 of requests, the Message IDs in the two directions can be very
1261 different. There is no ambiguity in the messages, however, because
1262 the (I)nitiator and (R)esponse bits in the message header specify
1263 which of the four messages a particular one is.
1265 Throughout this document, "initiator" refers to the party who
1266 initiated the exchange being described, and "original initiator"
1267 refers to the party who initiated the whole IKE SA. The "original
1268 initiator" always refers to the party who initiated the exchange
1269 which resulted in the current IKE SA. In other words, if the
1270 "original responder" starts rekeying the IKE SA, that party becomes
1271 the "original initiator" of the new IKE SA.
1273 Note that Message IDs are cryptographically protected and provide
1274 protection against message replays. In the unlikely event that
1275 Message IDs grow too large to fit in 32 bits, the IKE SA MUST be
1278 2.3. Window Size for Overlapping Requests
1280 The SET_WINDOW_SIZE notification asserts that the sending endpoint is
1281 capable of keeping state for multiple outstanding exchanges,
1282 permitting the recipient to send multiple requests before getting a
1283 response to the first. The data associated with a SET_WINDOW_SIZE
1287 Kaufman, et al. Expires January 9, 2010 [Page 23]
1289 Internet-Draft IKEv2bis July 2009
1292 notification MUST be 4 octets long and contain the big endian
1293 representation of the number of messages the sender promises to keep.
1294 The window size is always one until the initial exchanges complete.
1296 An IKE endpoint MUST wait for a response to each of its messages
1297 before sending a subsequent message unless it has received a
1298 SET_WINDOW_SIZE Notify message from its peer informing it that the
1299 peer is prepared to maintain state for multiple outstanding messages
1300 in order to allow greater throughput.
1302 After an IKE SA is set up, in order to maximize IKE throughput, an
1303 IKE endpoint MAY issue multiple requests before getting a response to
1304 any of them, up to the limit set by its peer's SET_WINDOW_SIZE.
1305 These requests may pass one another over the network. An IKE
1306 endpoint MUST be prepared to accept and process a request while it
1307 has a request outstanding in order to avoid a deadlock in this
1308 situation. An IKE endpoint may also accept and process multiple
1309 requests while it has a request outstanding.
1311 An IKE endpoint MUST NOT exceed the peer's stated window size for
1312 transmitted IKE requests. In other words, if the responder stated
1313 its window size is N, then when the initiator needs to make a request
1314 X, it MUST wait until it has received responses to all requests up
1315 through request X-N. An IKE endpoint MUST keep a copy of (or be able
1316 to regenerate exactly) each request it has sent until it receives the
1317 corresponding response. An IKE endpoint MUST keep a copy of (or be
1318 able to regenerate exactly) the number of previous responses equal to
1319 its declared window size in case its response was lost and the
1320 initiator requests its retransmission by retransmitting the request.
1322 An IKE endpoint supporting a window size greater than one ought to be
1323 capable of processing incoming requests out of order to maximize
1324 performance in the event of network failures or packet reordering.
1326 The window size is normally a (possibly configurable) property of a
1327 particular implementation, and is not related to congestion control
1328 (unlike the window size in TCP, for example). In particular, it is
1329 not defined what the responder should do when it receives a
1330 SET_WINDOW_SIZE notification containing a smaller value than is
1331 currently in effect. Thus, there is currently no way to reduce the
1332 window size of an existing IKE SA; you can only increase it. When
1333 rekeying an IKE SA, the new IKE SA starts with window size 1 until it
1334 is explicitly increased by sending a new SET_WINDOW_SIZE
1337 The INVALID_MESSAGE_ID notification is sent when an IKE message ID
1338 outside the supported window is received. This Notify MUST NOT be
1339 sent in a response; the invalid request MUST NOT be acknowledged.
1343 Kaufman, et al. Expires January 9, 2010 [Page 24]
1345 Internet-Draft IKEv2bis July 2009
1348 Instead, inform the other side by initiating an INFORMATIONAL
1349 exchange with Notification data containing the four octet invalid
1350 message ID. Sending this notification is optional, and notifications
1351 of this type MUST be rate limited.
1353 2.4. State Synchronization and Connection Timeouts
1355 An IKE endpoint is allowed to forget all of its state associated with
1356 an IKE SA and the collection of corresponding Child SAs at any time.
1357 This is the anticipated behavior in the event of an endpoint crash
1358 and restart. It is important when an endpoint either fails or
1359 reinitializes its state that the other endpoint detect those
1360 conditions and not continue to waste network bandwidth by sending
1361 packets over discarded SAs and having them fall into a black hole.
1363 The INITIAL_CONTACT notification asserts that this IKE SA is the only
1364 IKE SA currently active between the authenticated identities. It MAY
1365 be sent when an IKE SA is established after a crash, and the
1366 recipient MAY use this information to delete any other IKE SAs it has
1367 to the same authenticated identity without waiting for a timeout.
1368 This notification MUST NOT be sent by an entity that may be
1369 replicated (e.g., a roaming user's credentials where the user is
1370 allowed to connect to the corporate firewall from two remote systems
1371 at the same time). The INITIAL_CONTACT notification, if sent, MUST
1372 be in the first IKE_AUTH request or response, not as a separate
1373 exchange afterwards; however, receiving parties MAY ignore it in
1376 Since IKE is designed to operate in spite of Denial of Service (DoS)
1377 attacks from the network, an endpoint MUST NOT conclude that the
1378 other endpoint has failed based on any routing information (e.g.,
1379 ICMP messages) or IKE messages that arrive without cryptographic
1380 protection (e.g., Notify messages complaining about unknown SPIs).
1381 An endpoint MUST conclude that the other endpoint has failed only
1382 when repeated attempts to contact it have gone unanswered for a
1383 timeout period or when a cryptographically protected INITIAL_CONTACT
1384 notification is received on a different IKE SA to the same
1385 authenticated identity. An endpoint should suspect that the other
1386 endpoint has failed based on routing information and initiate a
1387 request to see whether the other endpoint is alive. To check whether
1388 the other side is alive, IKE specifies an empty INFORMATIONAL message
1389 that (like all IKE requests) requires an acknowledgement (note that
1390 within the context of an IKE SA, an "empty" message consists of an
1391 IKE header followed by an Encrypted payload that contains no
1392 payloads). If a cryptographically protected (fresh, i.e. not
1393 retransmitted) message has been received from the other side
1394 recently, unprotected notifications MAY be ignored. Implementations
1395 MUST limit the rate at which they take actions based on unprotected
1399 Kaufman, et al. Expires January 9, 2010 [Page 25]
1401 Internet-Draft IKEv2bis July 2009
1406 Numbers of retries and lengths of timeouts are not covered in this
1407 specification because they do not affect interoperability. It is
1408 suggested that messages be retransmitted at least a dozen times over
1409 a period of at least several minutes before giving up on an SA, but
1410 different environments may require different rules. To be a good
1411 network citizen, retranmission times MUST increase exponentially to
1412 avoid flooding the network and making an existing congestion
1413 situation worse. If there has only been outgoing traffic on all of
1414 the SAs associated with an IKE SA, it is essential to confirm
1415 liveness of the other endpoint to avoid black holes. If no
1416 cryptographically protected messages have been received on an IKE SA
1417 or any of its Child SAs recently, the system needs to perform a
1418 liveness check in order to prevent sending messages to a dead peer.
1419 (This is sometimes called "dead peer detection" or "DPD", although it
1420 is really detecting live peers, not dead ones.) Receipt of a fresh
1421 cryptographically protected message on an IKE SA or any of its Child
1422 SAs ensures liveness of the IKE SA and all of its Child SAs. Note
1423 that this places requirements on the failure modes of an IKE
1424 endpoint. An implementation MUST NOT continue sending on any SA if
1425 some failure prevents it from receiving on all of the associated SAs.
1426 If Child SAs can fail independently from one another without the
1427 associated IKE SA being able to send a delete message, then they MUST
1428 be negotiated by separate IKE SAs.
1430 There is a Denial of Service attack on the initiator of an IKE SA
1431 that can be avoided if the initiator takes the proper care. Since
1432 the first two messages of an SA setup are not cryptographically
1433 protected, an attacker could respond to the initiator's message
1434 before the genuine responder and poison the connection setup attempt.
1435 To prevent this, the initiator MAY be willing to accept multiple
1436 responses to its first message, treat each as potentially legitimate,
1437 respond to it, and then discard all the invalid half-open connections
1438 when it receives a valid cryptographically protected response to any
1439 one of its requests. Once a cryptographically valid response is
1440 received, all subsequent responses should be ignored whether or not
1441 they are cryptographically valid.
1443 Note that with these rules, there is no reason to negotiate and agree
1444 upon an SA lifetime. If IKE presumes the partner is dead, based on
1445 repeated lack of acknowledgement to an IKE message, then the IKE SA
1446 and all Child SAs set up through that IKE SA are deleted.
1448 An IKE endpoint may at any time delete inactive Child SAs to recover
1449 resources used to hold their state. If an IKE endpoint chooses to
1450 delete Child SAs, it MUST send Delete payloads to the other end
1451 notifying it of the deletion. It MAY similarly time out the IKE SA.
1455 Kaufman, et al. Expires January 9, 2010 [Page 26]
1457 Internet-Draft IKEv2bis July 2009
1460 Closing the IKE SA implicitly closes all associated Child SAs. In
1461 this case, an IKE endpoint SHOULD send a Delete payload indicating
1462 that it has closed the IKE SA unless the other endpoint is no longer
1465 2.5. Version Numbers and Forward Compatibility
1467 This document describes version 2.0 of IKE, meaning the major version
1468 number is 2 and the minor version number is 0. This document is a
1469 replacement for [IKEV2]. It is likely that some implementations will
1470 want to support version 1.0 and version 2.0, and in the future, other
1473 The major version number should be incremented only if the packet
1474 formats or required actions have changed so dramatically that an
1475 older version node would not be able to interoperate with a newer
1476 version node if it simply ignored the fields it did not understand
1477 and took the actions specified in the older specification. The minor
1478 version number indicates new capabilities, and MUST be ignored by a
1479 node with a smaller minor version number, but used for informational
1480 purposes by the node with the larger minor version number. For
1481 example, it might indicate the ability to process a newly defined
1482 notification message. The node with the larger minor version number
1483 would simply note that its correspondent would not be able to
1484 understand that message and therefore would not send it.
1486 If an endpoint receives a message with a higher major version number,
1487 it MUST drop the message and SHOULD send an unauthenticated
1488 notification message of type INVALID_MAJOR_VERSION containing the
1489 highest (closest) version number it supports. If an endpoint
1490 supports major version n, and major version m, it MUST support all
1491 versions between n and m. If it receives a message with a major
1492 version that it supports, it MUST respond with that version number.
1493 In order to prevent two nodes from being tricked into corresponding
1494 with a lower major version number than the maximum that they both
1495 support, IKE has a flag that indicates that the node is capable of
1496 speaking a higher major version number.
1498 Thus, the major version number in the IKE header indicates the
1499 version number of the message, not the highest version number that
1500 the transmitter supports. If the initiator is capable of speaking
1501 versions n, n+1, and n+2, and the responder is capable of speaking
1502 versions n and n+1, then they will negotiate speaking n+1, where the
1503 initiator will set a flag indicating its ability to speak a higher
1504 version. If they mistakenly (perhaps through an active attacker
1505 sending error messages) negotiate to version n, then both will notice
1506 that the other side can support a higher version number, and they
1507 MUST break the connection and reconnect using version n+1.
1511 Kaufman, et al. Expires January 9, 2010 [Page 27]
1513 Internet-Draft IKEv2bis July 2009
1516 Note that IKEv1 does not follow these rules, because there is no way
1517 in v1 of noting that you are capable of speaking a higher version
1518 number. So an active attacker can trick two v2-capable nodes into
1519 speaking v1. When a v2-capable node negotiates down to v1, it should
1520 note that fact in its logs.
1522 Also for forward compatibility, all fields marked RESERVED MUST be
1523 set to zero by an implementation running version 2.0, and their
1524 content MUST be ignored by an implementation running version 2.0 ("Be
1525 conservative in what you send and liberal in what you receive"). In
1526 this way, future versions of the protocol can use those fields in a
1527 way that is guaranteed to be ignored by implementations that do not
1528 understand them. Similarly, payload types that are not defined are
1529 reserved for future use; implementations of a version where they are
1530 undefined MUST skip over those payloads and ignore their contents.
1532 IKEv2 adds a "critical" flag to each payload header for further
1533 flexibility for forward compatibility. If the critical flag is set
1534 and the payload type is unrecognized, the message MUST be rejected
1535 and the response to the IKE request containing that payload MUST
1536 include a Notify payload UNSUPPORTED_CRITICAL_PAYLOAD, indicating an
1537 unsupported critical payload was included. In that Notify payload,
1538 the notification data contains the one-octet payload type. If the
1539 critical flag is not set and the payload type is unsupported, that
1540 payload MUST be ignored. Payloads sent in IKE response messages MUST
1541 NOT have the critical flag set. Note that the critical flag applies
1542 only to the payload type, not the contents. If the payload type is
1543 recognized, but the payload contains something which is not (such as
1544 an unknown transform inside an SA payload, or an unknown Notify
1545 Message Type inside a Notify payload), the critical flag is ignored.
1547 Although new payload types may be added in the future and may appear
1548 interleaved with the fields defined in this specification,
1549 implementations SHOULD send the payloads defined in this
1550 specification in the order shown in the figures in Section 2;
1551 implementations MUST NOT reject as invalid a message with those
1552 payloads in any other order.
1554 2.6. IKE SA SPIs and Cookies
1556 The term "cookies" originates with Karn and Simpson [PHOTURIS] in
1557 Photuris, an early proposal for key management with IPsec, and it has
1558 persisted. The Internet Security Association and Key Management
1559 Protocol (ISAKMP) [ISAKMP] fixed message header includes two eight-
1560 octet fields titled "cookies", and that syntax is used by both IKEv1
1561 and IKEv2, although in IKEv2 they are referred to as the "IKE SPI"
1562 and there is a new separate field in a Notify payload holding the
1563 cookie. The initial two eight-octet fields in the header are used as
1567 Kaufman, et al. Expires January 9, 2010 [Page 28]
1569 Internet-Draft IKEv2bis July 2009
1572 a connection identifier at the beginning of IKE packets. Each
1573 endpoint chooses one of the two SPIs and MUST choose them so as to be
1574 unique identifiers of an IKE SA. An SPI value of zero is special and
1575 indicates that the remote SPI value is not yet known by the sender.
1577 Incoming IKE packets are mapped to an IKE SA only using the packet's
1578 SPI, not using (for example) the source IP address of the packet.
1580 Unlike ESP and AH where only the recipient's SPI appears in the
1581 header of a message, in IKE the sender's SPI is also sent in every
1582 message. Since the SPI chosen by the original initiator of the IKE
1583 SA is always sent first, an endpoint with multiple IKE SAs open that
1584 wants to find the appropriate IKE SA using the SPI it assigned must
1585 look at the I(nitiator) Flag bit in the header to determine whether
1586 it assigned the first or the second eight octets.
1588 In the first message of an initial IKE exchange, the initiator will
1589 not know the responder's SPI value and will therefore set that field
1592 An expected attack against IKE is state and CPU exhaustion, where the
1593 target is flooded with session initiation requests from forged IP
1594 addresses. This attack can be made less effective if an
1595 implementation of a responder uses minimal CPU and commits no state
1596 to an SA until it knows the initiator can receive packets at the
1597 address from which it claims to be sending them.
1599 When a responder detects a large number of half-open IKE SAs, it
1600 SHOULD reply to IKE_SA_INIT requests with a response containing the
1601 COOKIE notification. The data associated with this notification MUST
1602 be between 1 and 64 octets in length (inclusive), and its generation
1603 is described later in this section. If the IKE_SA_INIT response
1604 includes the COOKIE notification, the initiator MUST then retry the
1605 IKE_SA_INIT request, and include the COOKIE notification containing
1606 the received data as the first payload, and all other payloads
1607 unchanged. The initial exchange will then be as follows:
1623 Kaufman, et al. Expires January 9, 2010 [Page 29]
1625 Internet-Draft IKEv2bis July 2009
1629 -------------------------------------------------------------------
1630 HDR(A,0), SAi1, KEi, Ni -->
1631 <-- HDR(A,0), N(COOKIE)
1632 HDR(A,0), N(COOKIE), SAi1,
1634 <-- HDR(A,B), SAr1, KEr,
1636 HDR(A,B), SK {IDi, [CERT,]
1637 [CERTREQ,] [IDr,] AUTH,
1639 <-- HDR(A,B), SK {IDr, [CERT,]
1640 AUTH, SAr2, TSi, TSr}
1642 The first two messages do not affect any initiator or responder state
1643 except for communicating the cookie. In particular, the message
1644 sequence numbers in the first four messages will all be zero and the
1645 message sequence numbers in the last two messages will be one. 'A'
1646 is the SPI assigned by the initiator, while 'B' is the SPI assigned
1649 An IKE implementation can implement its responder cookie generation
1650 in such a way as to not require any saved state to recognize its
1651 valid cookie when the second IKE_SA_INIT message arrives. The exact
1652 algorithms and syntax they use to generate cookies do not affect
1653 interoperability and hence are not specified here. The following is
1654 an example of how an endpoint could use cookies to implement limited
1657 A good way to do this is to set the responder cookie to be:
1659 Cookie = <VersionIDofSecret> | Hash(Ni | IPi | SPIi | <secret>)
1661 where <secret> is a randomly generated secret known only to the
1662 responder and periodically changed and | indicates concatenation.
1663 <VersionIDofSecret> should be changed whenever <secret> is
1664 regenerated. The cookie can be recomputed when the IKE_SA_INIT
1665 arrives the second time and compared to the cookie in the received
1666 message. If it matches, the responder knows that the cookie was
1667 generated since the last change to <secret> and that IPi must be the
1668 same as the source address it saw the first time. Incorporating SPIi
1669 into the calculation ensures that if multiple IKE SAs are being set
1670 up in parallel they will all get different cookies (assuming the
1671 initiator chooses unique SPIi's). Incorporating Ni in the hash
1672 ensures that an attacker who sees only message 2 can't successfully
1673 forge a message 3. Also, incorporating Ni in the hash prevents an
1674 attacker from fetching one one cookie from the other end, and then
1675 initiating many IKE_SA_INIT exchanges all with different initiator
1679 Kaufman, et al. Expires January 9, 2010 [Page 30]
1681 Internet-Draft IKEv2bis July 2009
1684 SPIs (and perhaps port numbers) so that the responder thinks that
1685 there are lots of machines behind one NAT box who are all trying to
1688 If a new value for <secret> is chosen while there are connections in
1689 the process of being initialized, an IKE_SA_INIT might be returned
1690 with other than the current <VersionIDofSecret>. The responder in
1691 that case MAY reject the message by sending another response with a
1692 new cookie or it MAY keep the old value of <secret> around for a
1693 short time and accept cookies computed from either one. The
1694 responder should not accept cookies indefinitely after <secret> is
1695 changed, since that would defeat part of the denial of service
1696 protection. The responder should change the value of <secret>
1697 frequently, especially if under attack.
1699 In addition to cookies, there are several cases where the IKE_SA_INIT
1700 exchange does not result in the creation of an IKE SA (such as
1701 INVALID_KE_PAYLOAD or NO_PROPOSAL_CHOSEN). In such a case, sending a
1702 zero value for the Responder's SPI is correct. If the responder
1703 sends a non-zero responder SPI, the initiator should not reject the
1704 response for only that reason.
1706 When one party receives an IKE_SA_INIT request containing a cookie
1707 whose contents do not match the value expected, that party MUST
1708 ignore the cookie and process the message as if no cookie had been
1709 included; usually this means sending a response containing a new
1710 cookie. The initiator should limit the number of cookie exchanges it
1711 tries before giving up. An attacker can forge multiple cookie
1712 responses to the initiator's IKE_SA_INIT message, and each of those
1713 forged cookie reply will trigger two packets: one packet from the
1714 initiator to the responder (which will reject those cookies), and one
1715 reply from responder to initiator that includes the correct cookie.
1717 2.6.1. Interaction of COOKIE and INVALID_KE_PAYLOAD
1719 There are two common reasons why the initiator may have to retry the
1720 IKE_SA_INIT exchange: the responder requests a cookie or wants a
1721 different Diffie-Hellman group than was included in the KEi payload.
1722 If the initiator receives a cookie from the responder, the initiator
1723 needs to decide whether or not to include the cookie in only the next
1724 retry of the IKE_SA_INIT request, or in all subsequent retries as
1727 If the initiator includes the cookie only in the next retry, one
1728 additional roundtrip may be needed in some cases. An additional
1729 roundtrip is needed also if the initiator includes the cookie in all
1730 retries, but the responder does not support this. For instance, if
1731 the responder includes the SAi1 and KEi payloads in cookie
1735 Kaufman, et al. Expires January 9, 2010 [Page 31]
1737 Internet-Draft IKEv2bis July 2009
1740 calculation, it will reject the request by sending a new cookie.
1742 If both peers support including the cookie in all retries, a slightly
1743 shorter exchange can happen.
1746 -----------------------------------------------------------
1747 HDR(A,0), SAi1, KEi, Ni -->
1748 <-- HDR(A,0), N(COOKIE)
1749 HDR(A,0), N(COOKIE), SAi1, KEi, Ni -->
1750 <-- HDR(A,0), N(INVALID_KE_PAYLOAD)
1751 HDR(A,0), N(COOKIE), SAi1, KEi', Ni -->
1752 <-- HDR(A,B), SAr1, KEr, Nr
1754 Implementations SHOULD support this shorter exchange, but MUST NOT
1755 fail if other implementations do not support this shorter exchange.
1757 2.7. Cryptographic Algorithm Negotiation
1759 The payload type known as "SA" indicates a proposal for a set of
1760 choices of IPsec protocols (IKE, ESP, or AH) for the SA as well as
1761 cryptographic algorithms associated with each protocol.
1763 An SA payload consists of one or more proposals. Each proposal
1764 includes one protocol. Each protocol contains one or more transforms
1765 -- each specifying a cryptographic algorithm. Each transform
1766 contains zero or more attributes (attributes are needed only if the
1767 transform identifier does not completely specify the cryptographic
1770 This hierarchical structure was designed to efficiently encode
1771 proposals for cryptographic suites when the number of supported
1772 suites is large because multiple values are acceptable for multiple
1773 transforms. The responder MUST choose a single suite, which may be
1774 any subset of the SA proposal following the rules below:
1776 Each proposal contains one protocol. If a proposal is accepted, the
1777 SA response MUST contain the same protocol. The responder MUST
1778 accept a single proposal or reject them all and return an error. The
1779 error is given in a notification of type NO_PROPOSAL_CHOSEN.
1781 Each IPsec protocol proposal contains one or more transforms. Each
1782 transform contains a transform type. The accepted cryptographic
1783 suite MUST contain exactly one transform of each type included in the
1784 proposal. For example: if an ESP proposal includes transforms
1785 ENCR_3DES, ENCR_AES w/keysize 128, ENCR_AES w/keysize 256,
1786 AUTH_HMAC_MD5, and AUTH_HMAC_SHA, the accepted suite MUST contain one
1787 of the ENCR_ transforms and one of the AUTH_ transforms. Thus, six
1791 Kaufman, et al. Expires January 9, 2010 [Page 32]
1793 Internet-Draft IKEv2bis July 2009
1796 combinations are acceptable.
1798 If an initiator proposes both normal ciphers with integrity
1799 protection as well as combined-mode ciphers, then two proposals are
1800 needed. One of the proposals includes the normal ciphers with the
1801 integrity algoritms for them, and the other proposal includes all the
1802 combined mode ciphers without the integrity algorithms (because
1803 combined mode ciphers are not allowed to have any integrity algorithm
1806 Since the initiator sends its Diffie-Hellman value in the
1807 IKE_SA_INIT, it must guess the Diffie-Hellman group that the
1808 responder will select from its list of supported groups. If the
1809 initiator guesses wrong, the responder will respond with a Notify
1810 payload of type INVALID_KE_PAYLOAD indicating the selected group. In
1811 this case, the initiator MUST retry the IKE_SA_INIT with the
1812 corrected Diffie-Hellman group. The initiator MUST again propose its
1813 full set of acceptable cryptographic suites because the rejection
1814 message was unauthenticated and otherwise an active attacker could
1815 trick the endpoints into negotiating a weaker suite than a stronger
1816 one that they both prefer.
1818 When the IKE_SA_INIT exchange does not result in the creation of an
1819 IKE SA due to INVALID_KE_PAYLOAD, NO_PROPOSAL_CHOSEN, or COOKIE (see
1820 Section 2.6), the responder's SPI will be zero. However, if the
1821 responder sends a non-zero responder SPI, the initiator should not
1822 reject the response for only that reason.
1826 IKE, ESP, and AH security associations use secret keys that should be
1827 used only for a limited amount of time and to protect a limited
1828 amount of data. This limits the lifetime of the entire security
1829 association. When the lifetime of a security association expires,
1830 the security association MUST NOT be used. If there is demand, new
1831 security associations MAY be established. Reestablishment of
1832 security associations to take the place of ones that expire is
1833 referred to as "rekeying".
1835 To allow for minimal IPsec implementations, the ability to rekey SAs
1836 without restarting the entire IKE SA is optional. An implementation
1837 MAY refuse all CREATE_CHILD_SA requests within an IKE SA. If an SA
1838 has expired or is about to expire and rekeying attempts using the
1839 mechanisms described here fail, an implementation MUST close the IKE
1840 SA and any associated Child SAs and then MAY start new ones.
1841 Implementations may wish to support in-place rekeying of SAs, since
1842 doing so offers better performance and is likely to reduce the number
1843 of packets lost during the transition.
1847 Kaufman, et al. Expires January 9, 2010 [Page 33]
1849 Internet-Draft IKEv2bis July 2009
1852 To rekey a Child SA within an existing IKE SA, create a new,
1853 equivalent SA (see Section 2.17 below), and when the new one is
1854 established, delete the old one. To rekey an IKE SA, establish a new
1855 equivalent IKE SA (see Section 2.18 below) with the peer to whom the
1856 old IKE SA is shared using a CREATE_CHILD_SA within the existing IKE
1857 SA. An IKE SA so created inherits all of the original IKE SA's Child
1858 SAs, and the new IKE SA is used for all control messages needed to
1859 maintain those Child SAs. The old IKE SA is then deleted, and the
1860 Delete payload to delete itself MUST be the last request sent over
1861 the old IKE SA. Note that, when rekeying, the new Child SA MAY have
1862 different traffic selectors and algorithms than the old one.
1864 SAs should be rekeyed proactively, i.e., the new SA should be
1865 established before the old one expires and becomes unusable. Enough
1866 time should elapse between the time the new SA is established and the
1867 old one becomes unusable so that traffic can be switched over to the
1870 A difference between IKEv1 and IKEv2 is that in IKEv1 SA lifetimes
1871 were negotiated. In IKEv2, each end of the SA is responsible for
1872 enforcing its own lifetime policy on the SA and rekeying the SA when
1873 necessary. If the two ends have different lifetime policies, the end
1874 with the shorter lifetime will end up always being the one to request
1875 the rekeying. If an SA has been inactive for a long time and if an
1876 endpoint would not initiate the SA in the absence of traffic, the
1877 endpoint MAY choose to close the SA instead of rekeying it when its
1878 lifetime expires. It should do so if there has been no traffic since
1879 the last time the SA was rekeyed.
1881 Note that IKEv2 deliberately allows parallel SAs with the same
1882 traffic selectors between common endpoints. One of the purposes of
1883 this is to support traffic quality of service (QoS) differences among
1884 the SAs (see [DIFFSERVFIELD], [DIFFSERVARCH], and section 4.1 of
1885 [DIFFTUNNEL]). Hence unlike IKEv1, the combination of the endpoints
1886 and the traffic selectors may not uniquely identify an SA between
1887 those endpoints, so the IKEv1 rekeying heuristic of deleting SAs on
1888 the basis of duplicate traffic selectors SHOULD NOT be used.
1890 The node that initiated the surviving rekeyed SA should delete the
1891 replaced SA after the new one is established.
1893 There are timing windows -- particularly in the presence of lost
1894 packets -- where endpoints may not agree on the state of an SA. The
1895 responder to a CREATE_CHILD_SA MUST be prepared to accept messages on
1896 an SA before sending its response to the creation request, so there
1897 is no ambiguity for the initiator. The initiator MAY begin sending
1898 on an SA as soon as it processes the response. The initiator,
1899 however, cannot receive on a newly created SA until it receives and
1903 Kaufman, et al. Expires January 9, 2010 [Page 34]
1905 Internet-Draft IKEv2bis July 2009
1908 processes the response to its CREATE_CHILD_SA request. How, then, is
1909 the responder to know when it is OK to send on the newly created SA?
1911 From a technical correctness and interoperability perspective, the
1912 responder MAY begin sending on an SA as soon as it sends its response
1913 to the CREATE_CHILD_SA request. In some situations, however, this
1914 could result in packets unnecessarily being dropped, so an
1915 implementation MAY defer such sending.
1917 The responder can be assured that the initiator is prepared to
1918 receive messages on an SA if either (1) it has received a
1919 cryptographically valid message on the new SA, or (2) the new SA
1920 rekeys an existing SA and it receives an IKE request to close the
1921 replaced SA. When rekeying an SA, the responder continues to send
1922 traffic on the old SA until one of those events occurs. When
1923 establishing a new SA, the responder MAY defer sending messages on a
1924 new SA until either it receives one or a timeout has occurred. If an
1925 initiator receives a message on an SA for which it has not received a
1926 response to its CREATE_CHILD_SA request, it interprets that as a
1927 likely packet loss and retransmits the CREATE_CHILD_SA request. An
1928 initiator MAY send a dummy message on a newly created SA if it has no
1929 messages queued in order to assure the responder that the initiator
1930 is ready to receive messages.
1932 2.8.1. Simultaneous Child SA rekeying
1934 If the two ends have the same lifetime policies, it is possible that
1935 both will initiate a rekeying at the same time (which will result in
1936 redundant SAs). To reduce the probability of this happening, the
1937 timing of rekeying requests SHOULD be jittered (delayed by a random
1938 amount of time after the need for rekeying is noticed).
1940 This form of rekeying may temporarily result in multiple similar SAs
1941 between the same pairs of nodes. When there are two SAs eligible to
1942 receive packets, a node MUST accept incoming packets through either
1943 SA. If redundant SAs are created though such a collision, the SA
1944 created with the lowest of the four nonces used in the two exchanges
1945 SHOULD be closed by the endpoint that created it. "Lowest" means an
1946 octet-by-octet, lexicographical comparison (instead of, for instance,
1947 comparing the nonces as large integers). In other words, start by
1948 comparing the first octet; if they're equal, move to the next octet,
1949 and so on. If you reach the end of one nonce, that nonce is the
1952 The following is an explanation on the impact this has on
1953 implementations. Assume that hosts A and B have an existing IPsec SA
1954 pair with SPIs (SPIa1,SPIb1), and both start rekeying it at the same
1959 Kaufman, et al. Expires January 9, 2010 [Page 35]
1961 Internet-Draft IKEv2bis July 2009
1965 -------------------------------------------------------------------
1966 send req1: N(REKEY_SA,SPIa1),
1967 SA(..,SPIa2,..),Ni1,.. -->
1968 <-- send req2: N(REKEY_SA,SPIb1),
1972 At this point, A knows there is a simultaneous rekeying going on.
1973 However, it cannot yet know which of the exchanges will have the
1974 lowest nonce, so it will just note the situation and respond as
1977 send resp2: SA(..,SPIa3,..),
1981 Now B also knows that simultaneous rekeying is going on. It responds
1984 <-- send resp1: SA(..,SPIb3,..),
1989 At this point, there are three Child SA pairs between A and B (the
1990 old one and two new ones). A and B can now compare the nonces.
1991 Suppose that the lowest nonce was Nr1 in message resp2; in this case,
1992 B (the sender of req2) deletes the redundant new SA, and A (the node
1993 that initiated the surviving rekeyed SA), deletes the old one.
1995 send req3: D(SPIa1) -->
1996 <-- send req4: D(SPIb2)
1998 <-- send resp3: D(SPIb1)
2000 send resp4: D(SPIa3) -->
2002 The rekeying is now finished.
2004 However, there is a second possible sequence of events that can
2005 happen if some packets are lost in the network, resulting in
2006 retransmissions. The rekeying begins as usual, but A's first packet
2015 Kaufman, et al. Expires January 9, 2010 [Page 36]
2017 Internet-Draft IKEv2bis July 2009
2021 -------------------------------------------------------------------
2022 send req1: N(REKEY_SA,SPIa1),
2025 <-- send req2: N(REKEY_SA,SPIb1),
2028 send resp2: SA(..,SPIa3,..),
2031 <-- send req3: D(SPIb1)
2033 send resp3: D(SPIa1) -->
2036 From B's point of view, the rekeying is now completed, and since it
2037 has not yet received A's req1, it does not even know that there was
2038 simultaneous rekeying. However, A will continue retransmitting the
2039 message, and eventually it will reach B.
2044 To B, it looks like A is trying to rekey an SA that no longer exists;
2045 thus, B responds to the request with something non-fatal such as
2048 <-- send resp1: N(NO_PROPOSAL_CHOSEN)
2051 When A receives this error, it already knows there was simultaneous
2052 rekeying, so it can ignore the error message.
2054 2.8.2. Simultaneous IKE SA Rekeying
2056 Probably the most complex case occurs when both peers try to rekey
2057 the IKE_SA at the same time. Basically, the text in Section 2.8
2058 applies to this case as well; however, it is important to ensure that
2059 the CHILD_SAs are inherited by the right IKE_SA.
2061 The case where both endpoints notice the simultaneous rekeying works
2062 the same way as with CHILD_SAs. After the CREATE_CHILD_SA exchanges,
2063 three IKE_SAs exist between A and B; the one containing the lowest
2064 nonce inherits the CHILD_SAs.
2066 However, there is a twist to the other case where one rekeying
2071 Kaufman, et al. Expires January 9, 2010 [Page 37]
2073 Internet-Draft IKEv2bis July 2009
2077 -------------------------------------------------------------------
2079 SA(..,SPIa1,..),Ni1,.. -->
2080 <-- send req2: SA(..,SPIb1,..),Ni2,..
2082 <-- send resp1: SA(..,SPIb2,..),Nr2,..
2087 At this point, host B sees a request to close the IKE_SA. There's
2088 not much more to do than to reply as usual. However, at this point
2089 host B should stop retransmitting req2, since once host A receives
2090 resp3, it will delete all the state associated with the old IKE_SA
2091 and will not be able to reply to it.
2095 2.8.3. Rekeying the IKE SA Versus Reauthentication
2097 Rekeying the IKE SA and reauthentication are different concepts in
2098 IKEv2. Rekeying the IKE SA establishes new keys for the IKE SA and
2099 resets the Message ID counters, but it does not authenticate the
2100 parties again (no AUTH or EAP payloads are involved).
2102 Although rekeying the IKE SA may be important in some environments,
2103 reauthentication (the verification that the parties still have access
2104 to the long-term credentials) is often more important.
2106 IKEv2 does not have any special support for reauthentication.
2107 Reauthentication is done by creating a new IKE SA from scratch (using
2108 IKE_SA_INIT/IKE_AUTH exchanges, without any REKEY_SA notify
2109 payloads), creating new Child SAs within the new IKE SA (without
2110 REKEY_SA notify payloads), and finally deleting the old IKE SA (which
2111 deletes the old Child SAs as well).
2113 This means that reauthentication also establishes new keys for the
2114 IKE SA and Child SAs. Therefore, while rekeying can be performed
2115 more often than reauthentication, the situation where "authentication
2116 lifetime" is shorter than "key lifetime" does not make sense.
2118 While creation of a new IKE SA can be initiated by either party
2119 (initiator or responder in the original IKE SA), the use of EAP
2120 authentication and/or configuration payloads means in practice that
2121 reauthentication has to be initiated by the same party as the
2122 original IKE SA. IKEv2 does not currently allow the responder to
2123 request reauthentication in this case; however, there are extensions
2127 Kaufman, et al. Expires January 9, 2010 [Page 38]
2129 Internet-Draft IKEv2bis July 2009
2132 that add this functionality such as [REAUTH].
2134 2.9. Traffic Selector Negotiation
2136 When an RFC4301-compliant IPsec subsystem receives an IP packet that
2137 matches a "protect" selector in its Security Policy Database (SPD),
2138 the subsystem protects that packet with IPsec. When no SA exists
2139 yet, it is the task of IKE to create it. Maintenance of a system's
2140 SPD is outside the scope of IKE (see [PFKEY] for an example
2141 programming interface, although it only applies to IKEv1), though
2142 some implementations might update their SPD in connection with the
2143 running of IKE (for an example scenario, see Section 1.1.3).
2145 Traffic Selector (TS) payloads allow endpoints to communicate some of
2146 the information from their SPD to their peers. TS payloads specify
2147 the selection criteria for packets that will be forwarded over the
2148 newly set up SA. This can serve as a consistency check in some
2149 scenarios to assure that the SPDs are consistent. In others, it
2150 guides the dynamic update of the SPD.
2152 Two TS payloads appear in each of the messages in the exchange that
2153 creates a Child SA pair. Each TS payload contains one or more
2154 Traffic Selectors. Each Traffic Selector consists of an address
2155 range (IPv4 or IPv6), a port range, and an IP protocol ID.
2157 The first of the two TS payloads is known as TSi (Traffic Selector-
2158 initiator). The second is known as TSr (Traffic Selector-responder).
2159 TSi specifies the source address of traffic forwarded from (or the
2160 destination address of traffic forwarded to) the initiator of the
2161 Child SA pair. TSr specifies the destination address of the traffic
2162 forwarded to (or the source address of the traffic forwarded from)
2163 the responder of the Child SA pair. For example, if the original
2164 initiator requests the creation of a Child SA pair, and wishes to
2165 tunnel all traffic from subnet 192.0.1.* on the initiator's side to
2166 subnet 192.0.2.* on the responder's side, the initiator would include
2167 a single traffic selector in each TS payload. TSi would specify the
2168 address range (192.0.1.0 - 192.0.1.255) and TSr would specify the
2169 address range (192.0.2.0 - 192.0.2.255). Assuming that proposal was
2170 acceptable to the responder, it would send identical TS payloads
2171 back. (Note: The IP address range 192.0.2.* has been reserved for
2172 use in examples in RFCs and similar documents. This document needed
2173 two such ranges, and so also used 192.0.1.*. This should not be
2174 confused with any actual address.)
2176 IKEv2 allows the responder to choose a subset of the traffic proposed
2177 by the initiator. This could happen when the configurations of the
2178 two endpoints are being updated but only one end has received the new
2179 information. Since the two endpoints may be configured by different
2183 Kaufman, et al. Expires January 9, 2010 [Page 39]
2185 Internet-Draft IKEv2bis July 2009
2188 people, the incompatibility may persist for an extended period even
2189 in the absence of errors. It also allows for intentionally different
2190 configurations, as when one end is configured to tunnel all addresses
2191 and depends on the other end to have the up-to-date list.
2193 When the responder chooses a subset of the traffic proposed by the
2194 initiator, it narrows the traffic selectors to some subset of the
2195 initiator's proposal (provided the set does not become the null set).
2196 If the type of traffic selector proposed is unknown, the responder
2197 ignores that traffic selector, so that the unknown type is not be
2198 returned in the narrowed set.
2200 To enable the responder to choose the appropriate range in this case,
2201 if the initiator has requested the SA due to a data packet, the
2202 initiator SHOULD include as the first traffic selector in each of TSi
2203 and TSr a very specific traffic selector including the addresses in
2204 the packet triggering the request. In the example, the initiator
2205 would include in TSi two traffic selectors: the first containing the
2206 address range (192.0.1.43 - 192.0.1.43) and the source port and IP
2207 protocol from the packet and the second containing (192.0.1.0 -
2208 192.0.1.255) with all ports and IP protocols. The initiator would
2209 similarly include two traffic selectors in TSr. If the initiator
2210 creates the Child SA pair not in response to an arriving packet, but
2211 rather, say, upon startup, then there may be no specific addresses
2212 the initiator prefers for the initial tunnel over any other. In that
2213 case, the first values in TSi and TSr can be ranges rather than
2216 The responder performs the narrowing as follows:
2218 o If the responder's policy does not allow it to accept any part of
2219 the proposed traffic selectors, it responds with TS_UNACCEPTABLE.
2221 o If the responder's policy allows the entire set of traffic covered
2222 by TSi and TSr, no narrowing is necessary, and the responder can
2223 return the same TSi and TSr values.
2225 o If the responder's policy allows it to accept the first selector
2226 of TSi and TSr, then the responder MUST narrow the traffic
2227 selectors to a subset that includes the initiator's first choices.
2228 In this example above, the responder might respond with TSi being
2229 (192.0.1.43 - 192.0.1.43) with all ports and IP protocols.
2231 o If the responder's policy does not allow it to accept the first
2232 selector of TSi and TSr, the responder narrows to an acceptable
2233 subset of TSi and TSr.
2235 When narrowing is done, there may be several subsets that are
2239 Kaufman, et al. Expires January 9, 2010 [Page 40]
2241 Internet-Draft IKEv2bis July 2009
2244 acceptable but their union is not. In this case, the responder
2245 arbitrarily chooses one of them, and MAY include an
2246 ADDITIONAL_TS_POSSIBLE notification in the response. The
2247 ADDITIONAL_TS_POSSIBLE notification asserts that the responder
2248 narrowed the proposed traffic selectors but that other traffic
2249 selectors would also have been acceptable, though only in a separate
2250 SA. There is no data associated with this Notify type. This case
2251 will occur only when the initiator and responder are configured
2252 differently from one another. If the initiator and responder agree
2253 on the granularity of tunnels, the initiator will never request a
2254 tunnel wider than the responder will accept. Such misconfigurations
2255 should be recorded in error logs.
2257 It is possible for the responder's policy to contain multiple smaller
2258 ranges, all encompassed by the initiator's traffic selector, and with
2259 the responder's policy being that each of those ranges should be sent
2260 over a different SA. Continuing the example above, the responder
2261 might have a policy of being willing to tunnel those addresses to and
2262 from the initiator, but might require that each address pair be on a
2263 separately negotiated Child SA. If the initiator generated its
2264 request in response to an incoming packet from 192.0.1.43 to
2265 192.0.2.123, there would be no way for the responder to determine
2266 which pair of addresses should be included in this tunnel, and it
2267 would have to make a guess or reject the request with a status of
2268 SINGLE_PAIR_REQUIRED.
2270 The SINGLE_PAIR_REQUIRED error indicates that a CREATE_CHILD_SA
2271 request is unacceptable because its sender is only willing to accept
2272 traffic selectors specifying a single pair of addresses. The
2273 requestor is expected to respond by requesting an SA for only the
2274 specific traffic it is trying to forward.
2276 Few implementations will have policies that require separate SAs for
2277 each address pair. Because of this, if only some parts of the TSi
2278 and TSr proposed by the initiator are acceptable to the responder,
2279 responders SHOULD narrow the selectors to an acceptable subset rather
2280 than use SINGLE_PAIR_REQUIRED.
2282 2.9.1. Traffic Selectors Violating Own Policy
2284 When creating a new SA, the initiator needs to avoid proposing
2285 traffic selectors that violate its own policy. If this rule is not
2286 followed, valid traffic may be dropped. If you use decorrelated
2287 policies from [IPSECARCH], this kind of policy violations cannot
2290 This is best illustrated by an example. Suppose that host A has a
2291 policy whose effect is that traffic to 192.0.1.66 is sent via host B
2295 Kaufman, et al. Expires January 9, 2010 [Page 41]
2297 Internet-Draft IKEv2bis July 2009
2300 encrypted using AES, and traffic to all other hosts in 192.0.1.0/24
2301 is also sent via B, but must use 3DES. Suppose also that host B
2302 accepts any combination of AES and 3DES.
2304 If host A now proposes an SA that uses 3DES, and includes TSr
2305 containing (192.0.1.0-192.0.1.255), this will be accepted by host B.
2306 Now, host B can also use this SA to send traffic from 192.0.1.66, but
2307 those packets will be dropped by A since it requires the use of AES
2308 for those traffic. Even if host A creates a new SA only for
2309 192.0.1.66 that uses AES, host B may freely continue to use the first
2310 SA for the traffic. In this situation, when proposing the SA, host A
2311 should have followed its own policy, and included a TSr containing
2312 ((192.0.1.0-192.0.1.65),(192.0.1.67-192.0.1.255)) instead.
2314 In general, if (1) the initiator makes a proposal "for traffic X
2315 (TSi/TSr), do SA", and (2) for some subset X' of X, the initiator
2316 does not actually accept traffic X' with SA, and (3) the initiator
2317 would be willing to accept traffic X' with some SA' (!=SA), valid
2318 traffic can be unnecessarily dropped since the responder can apply
2319 either SA or SA' to traffic X'.
2323 The IKE_SA_INIT messages each contain a nonce. These nonces are used
2324 as inputs to cryptographic functions. The CREATE_CHILD_SA request
2325 and the CREATE_CHILD_SA response also contain nonces. These nonces
2326 are used to add freshness to the key derivation technique used to
2327 obtain keys for Child SA, and to ensure creation of strong pseudo-
2328 random bits from the Diffie-Hellman key. Nonces used in IKEv2 MUST
2329 be randomly chosen, MUST be at least 128 bits in size, and MUST be at
2330 least half the key size of the negotiated prf. ("prf" refers to
2331 "pseudo-random function", one of the cryptographic algorithms
2332 negotiated in the IKE exchange.) However, the initiator chooses the
2333 nonce before the outcome of the negotiation is known. Because of
2334 that, the nonce has to be long enough for all the PRFs being
2335 proposed. If the same random number source is used for both keys and
2336 nonces, care must be taken to ensure that the latter use does not
2337 compromise the former.
2339 2.11. Address and Port Agility
2341 IKE runs over UDP ports 500 and 4500, and implicitly sets up ESP and
2342 AH associations for the same IP addresses it runs over. The IP
2343 addresses and ports in the outer header are, however, not themselves
2344 cryptographically protected, and IKE is designed to work even through
2345 Network Address Translation (NAT) boxes. An implementation MUST
2346 accept incoming requests even if the source port is not 500 or 4500,
2347 and MUST respond to the address and port from which the request was
2351 Kaufman, et al. Expires January 9, 2010 [Page 42]
2353 Internet-Draft IKEv2bis July 2009
2356 received. It MUST specify the address and port at which the request
2357 was received as the source address and port in the response. IKE
2358 functions identically over IPv4 or IPv6.
2360 2.12. Reuse of Diffie-Hellman Exponentials
2362 IKE generates keying material using an ephemeral Diffie-Hellman
2363 exchange in order to gain the property of "perfect forward secrecy".
2364 This means that once a connection is closed and its corresponding
2365 keys are forgotten, even someone who has recorded all of the data
2366 from the connection and gets access to all of the long-term keys of
2367 the two endpoints cannot reconstruct the keys used to protect the
2368 conversation without doing a brute force search of the session key
2371 Achieving perfect forward secrecy requires that when a connection is
2372 closed, each endpoint MUST forget not only the keys used by the
2373 connection but also any information that could be used to recompute
2376 Since the computing of Diffie-Hellman exponentials is computationally
2377 expensive, an endpoint may find it advantageous to reuse those
2378 exponentials for multiple connection setups. There are several
2379 reasonable strategies for doing this. An endpoint could choose a new
2380 exponential only periodically though this could result in less-than-
2381 perfect forward secrecy if some connection lasts for less than the
2382 lifetime of the exponential. Or it could keep track of which
2383 exponential was used for each connection and delete the information
2384 associated with the exponential only when some corresponding
2385 connection was closed. This would allow the exponential to be reused
2386 without losing perfect forward secrecy at the cost of maintaining
2389 Decisions as to whether and when to reuse Diffie-Hellman exponentials
2390 is a private decision in the sense that it will not affect
2391 interoperability. An implementation that reuses exponentials MAY
2392 choose to remember the exponential used by the other endpoint on past
2393 exchanges and if one is reused to avoid the second half of the
2394 calculation. See [REUSE] for a security analysis of this practice
2395 and for additional security considerations when reusing ephemeral DH
2398 2.13. Generating Keying Material
2400 In the context of the IKE SA, four cryptographic algorithms are
2401 negotiated: an encryption algorithm, an integrity protection
2402 algorithm, a Diffie-Hellman group, and a pseudo-random function
2403 (prf). The pseudo-random function is used for the construction of
2407 Kaufman, et al. Expires January 9, 2010 [Page 43]
2409 Internet-Draft IKEv2bis July 2009
2412 keying material for all of the cryptographic algorithms used in both
2413 the IKE SA and the Child SAs.
2415 We assume that each encryption algorithm and integrity protection
2416 algorithm uses a fixed-size key and that any randomly chosen value of
2417 that fixed size can serve as an appropriate key. For algorithms that
2418 accept a variable length key, a fixed key size MUST be specified as
2419 part of the cryptographic transform negotiated (see Section 3.3.5 for
2420 the defintion of the Key Length transform attribute). For algorithms
2421 for which not all values are valid keys (such as DES or 3DES with key
2422 parity), the algorithm by which keys are derived from arbitrary
2423 values MUST be specified by the cryptographic transform. For
2424 integrity protection functions based on Hashed Message Authentication
2425 Code (HMAC), the fixed key size is the size of the output of the
2426 underlying hash function.
2428 It is assumed that pseudo-random functions (PRFs) accept keys of any
2429 length, but have a preferred key size. The preferred key size is
2430 used as the length of SK_d, SK_pi, and SK_pr (see Section 2.14). For
2431 PRFs based on the HMAC construction, the preferred key size is equal
2432 to the length of the output of the underlying hash function. Other
2433 types of PRFs MUST specify their preferred key size.
2435 Keying material will always be derived as the output of the
2436 negotiated prf algorithm. Since the amount of keying material needed
2437 may be greater than the size of the output of the prf algorithm, we
2438 will use the prf iteratively. We will use the terminology prf+ to
2439 describe the function that outputs a pseudo-random stream based on
2440 the inputs to a prf as follows: (where | indicates concatenation)
2442 prf+ (K,S) = T1 | T2 | T3 | T4 | ...
2445 T1 = prf (K, S | 0x01)
2446 T2 = prf (K, T1 | S | 0x02)
2447 T3 = prf (K, T2 | S | 0x03)
2448 T4 = prf (K, T3 | S | 0x04)
2450 continuing as needed to compute all required keys. The keys are
2451 taken from the output string without regard to boundaries (e.g., if
2452 the required keys are a 256-bit Advanced Encryption Standard (AES)
2453 key and a 160-bit HMAC key, and the prf function generates 160 bits,
2454 the AES key will come from T1 and the beginning of T2, while the HMAC
2455 key will come from the rest of T2 and the beginning of T3).
2457 The constant concatenated to the end of each string feeding the prf
2458 is a single octet. prf+ in this document is not defined beyond 255
2459 times the size of the prf output.
2463 Kaufman, et al. Expires January 9, 2010 [Page 44]
2465 Internet-Draft IKEv2bis July 2009
2468 2.14. Generating Keying Material for the IKE SA
2470 The shared keys are computed as follows. A quantity called SKEYSEED
2471 is calculated from the nonces exchanged during the IKE_SA_INIT
2472 exchange and the Diffie-Hellman shared secret established during that
2473 exchange. SKEYSEED is used to calculate seven other secrets: SK_d
2474 used for deriving new keys for the Child SAs established with this
2475 IKE SA; SK_ai and SK_ar used as a key to the integrity protection
2476 algorithm for authenticating the component messages of subsequent
2477 exchanges; SK_ei and SK_er used for encrypting (and of course
2478 decrypting) all subsequent exchanges; and SK_pi and SK_pr, which are
2479 used when generating an AUTH payload. The lengths of SK_d, SK_pi,
2480 and SK_pr are the preferred key length of the agreed-to PRF.
2482 SKEYSEED and its derivatives are computed as follows:
2484 SKEYSEED = prf(Ni | Nr, g^ir)
2486 {SK_d | SK_ai | SK_ar | SK_ei | SK_er | SK_pi | SK_pr }
2487 = prf+ (SKEYSEED, Ni | Nr | SPIi | SPIr )
2489 (indicating that the quantities SK_d, SK_ai, SK_ar, SK_ei, SK_er,
2490 SK_pi, and SK_pr are taken in order from the generated bits of the
2491 prf+). g^ir is the shared secret from the ephemeral Diffie-Hellman
2492 exchange. g^ir is represented as a string of octets in big endian
2493 order padded with zeros if necessary to make it the length of the
2494 modulus. Ni and Nr are the nonces, stripped of any headers. For
2495 historical backwards-compatibility reasons, there are two PRFs that
2496 are treated specially in this calculation. If the negotiated PRF is
2497 AES-XCBC-PRF-128 [RFC4434] or AES-CMAC-PRF-128 [RFC4615], only the
2498 first 64 bits of Ni and the first 64 bits of Nr are used in the
2501 The two directions of traffic flow use different keys. The keys used
2502 to protect messages from the original initiator are SK_ai and SK_ei.
2503 The keys used to protect messages in the other direction are SK_ar
2506 2.15. Authentication of the IKE SA
2508 When not using extensible authentication (see Section 2.16), the
2509 peers are authenticated by having each sign (or MAC using a shared
2510 secret as the key) a block of data. For the responder, the octets to
2511 be signed start with the first octet of the first SPI in the header
2512 of the second message (IKE_SA_INIT response) and end with the last
2513 octet of the last payload in the second message. Appended to this
2514 (for purposes of computing the signature) are the initiator's nonce
2515 Ni (just the value, not the payload containing it), and the value
2519 Kaufman, et al. Expires January 9, 2010 [Page 45]
2521 Internet-Draft IKEv2bis July 2009
2524 prf(SK_pr,IDr') where IDr' is the responder's ID payload excluding
2525 the fixed header. Note that neither the nonce Ni nor the value
2526 prf(SK_pr,IDr') are transmitted. Similarly, the initiator signs the
2527 first message (IKE_SA_INIT request), starting with the first octet of
2528 the first SPI in the header and ending with the last octet of the
2529 last payload. Appended to this (for purposes of computing the
2530 signature) are the responder's nonce Nr, and the value
2531 prf(SK_pi,IDi'). In the above calculation, IDi' and IDr' are the
2532 entire ID payloads excluding the fixed header. It is critical to the
2533 security of the exchange that each side sign the other side's nonce.
2535 The initiator's signed octets can be described as:
2537 InitiatorSignedOctets = RealMessage1 | NonceRData | MACedIDForI
2538 GenIKEHDR = [ four octets 0 if using port 4500 ] | RealIKEHDR
2539 RealIKEHDR = SPIi | SPIr | . . . | Length
2540 RealMessage1 = RealIKEHDR | RestOfMessage1
2541 NonceRPayload = PayloadHeader | NonceRData
2542 InitiatorIDPayload = PayloadHeader | RestOfIDPayload
2543 RestOfInitIDPayload = IDType | RESERVED | InitIDData
2544 MACedIDForI = prf(SK_pi, RestOfInitIDPayload)
2546 The responder's signed octets can be described as:
2548 ResponderSignedOctets = RealMessage2 | NonceIData | MACedIDForR
2549 GenIKEHDR = [ four octets 0 if using port 4500 ] | RealIKEHDR
2550 RealIKEHDR = SPIi | SPIr | . . . | Length
2551 RealMessage2 = RealIKEHDR | RestOfMessage2
2552 NonceIPayload = PayloadHeader | NonceIData
2553 ResponderIDPayload = PayloadHeader | RestOfIDPayload
2554 RestOfRespIDPayload = IDType | RESERVED | RespIDData
2555 MACedIDForR = prf(SK_pr, RestOfRespIDPayload)
2557 Note that all of the payloads are included under the signature,
2558 including any payload types not defined in this document. If the
2559 first message of the exchange is sent multiple times (such as with a
2560 responder cookie and/or a different Diffie-Hellman group), it is the
2561 latest version of the message that is signed.
2563 Optionally, messages 3 and 4 MAY include a certificate, or
2564 certificate chain providing evidence that the key used to compute a
2565 digital signature belongs to the name in the ID payload. The
2566 signature or MAC will be computed using algorithms dictated by the
2567 type of key used by the signer, and specified by the Auth Method
2568 field in the Authentication payload. There is no requirement that
2569 the initiator and responder sign with the same cryptographic
2570 algorithms. The choice of cryptographic algorithms depends on the
2571 type of key each has. In particular, the initiator may be using a
2575 Kaufman, et al. Expires January 9, 2010 [Page 46]
2577 Internet-Draft IKEv2bis July 2009
2580 shared key while the responder may have a public signature key and
2581 certificate. It will commonly be the case (but it is not required)
2582 that if a shared secret is used for authentication that the same key
2583 is used in both directions.
2585 Note that it is a common but typically insecure practice to have a
2586 shared key derived solely from a user-chosen password without
2587 incorporating another source of randomness. This is typically
2588 insecure because user-chosen passwords are unlikely to have
2589 sufficient unpredictability to resist dictionary attacks and these
2590 attacks are not prevented in this authentication method.
2591 (Applications using password-based authentication for bootstrapping
2592 and IKE SA should use the authentication method in Section 2.16,
2593 which is designed to prevent off-line dictionary attacks.) The pre-
2594 shared key needs to contain as much unpredictability as the strongest
2595 key being negotiated. In the case of a pre-shared key, the AUTH
2596 value is computed as:
2599 AUTH = prf( prf( Shared Secret,"Key Pad for IKEv2"),
2600 <InitiatorSignedOctets>)
2602 AUTH = prf( prf( Shared Secret,"Key Pad for IKEv2"),
2603 <ResponderSignedOctets>)
2605 where the string "Key Pad for IKEv2" is 17 ASCII characters without
2606 null termination. The shared secret can be variable length. The pad
2607 string is added so that if the shared secret is derived from a
2608 password, the IKE implementation need not store the password in
2609 cleartext, but rather can store the value prf(Shared Secret,"Key Pad
2610 for IKEv2"), which could not be used as a password equivalent for
2611 protocols other than IKEv2. As noted above, deriving the shared
2612 secret from a password is not secure. This construction is used
2613 because it is anticipated that people will do it anyway. The
2614 management interface by which the Shared Secret is provided MUST
2615 accept ASCII strings of at least 64 octets and MUST NOT add a null
2616 terminator before using them as shared secrets. It MUST also accept
2617 a hex encoding of the Shared Secret. The management interface MAY
2618 accept other encodings if the algorithm for translating the encoding
2619 to a binary string is specified.
2621 2.16. Extensible Authentication Protocol Methods
2623 In addition to authentication using public key signatures and shared
2624 secrets, IKE supports authentication using methods defined in RFC
2625 3748 [EAP]. Typically, these methods are asymmetric (designed for a
2626 user authenticating to a server), and they may not be mutual. For
2627 this reason, these protocols are typically used to authenticate the
2631 Kaufman, et al. Expires January 9, 2010 [Page 47]
2633 Internet-Draft IKEv2bis July 2009
2636 initiator to the responder and MUST be used in conjunction with a
2637 public key signature based authentication of the responder to the
2638 initiator. These methods are often associated with mechanisms
2639 referred to as "Legacy Authentication" mechanisms.
2641 While this memo references [EAP] with the intent that new methods can
2642 be added in the future without updating this specification, some
2643 simpler variations are documented here and in Section 3.16. [EAP]
2644 defines an authentication protocol requiring a variable number of
2645 messages. Extensible Authentication is implemented in IKE as
2646 additional IKE_AUTH exchanges that MUST be completed in order to
2647 initialize the IKE SA.
2649 An initiator indicates a desire to use extensible authentication by
2650 leaving out the AUTH payload from message 3. By including an IDi
2651 payload but not an AUTH payload, the initiator has declared an
2652 identity but has not proven it. If the responder is willing to use
2653 an extensible authentication method, it will place an Extensible
2654 Authentication Protocol (EAP) payload in message 4 and defer sending
2655 SAr2, TSi, and TSr until initiator authentication is complete in a
2656 subsequent IKE_AUTH exchange. In the case of a minimal extensible
2657 authentication, the initial SA establishment will appear as follows:
2660 -------------------------------------------------------------------
2661 HDR, SAi1, KEi, Ni -->
2662 <-- HDR, SAr1, KEr, Nr, [CERTREQ]
2663 HDR, SK {IDi, [CERTREQ,]
2666 <-- HDR, SK {IDr, [CERT,] AUTH,
2669 <-- HDR, SK {EAP (success)}
2671 <-- HDR, SK {AUTH, SAr2, TSi, TSr }
2673 As described in Section 2.2, when EAP is used, each pair of IKE SA
2674 initial setup messages will have their message numbers incremented;
2675 the first pair of AUTH messages will have an ID of 1, the second will
2678 For EAP methods that create a shared key as a side effect of
2679 authentication, that shared key MUST be used by both the initiator
2680 and responder to generate AUTH payloads in messages 7 and 8 using the
2681 syntax for shared secrets specified in Section 2.15. The shared key
2682 from EAP is the field from the EAP specification named MSK. This
2683 shared key generated during an IKE exchange MUST NOT be used for any
2687 Kaufman, et al. Expires January 9, 2010 [Page 48]
2689 Internet-Draft IKEv2bis July 2009
2694 EAP methods that do not establish a shared key SHOULD NOT be used, as
2695 they are subject to a number of man-in-the-middle attacks [EAPMITM]
2696 if these EAP methods are used in other protocols that do not use a
2697 server-authenticated tunnel. Please see the Security Considerations
2698 section for more details. If EAP methods that do not generate a
2699 shared key are used, the AUTH payloads in messages 7 and 8 MUST be
2700 generated using SK_pi and SK_pr, respectively.
2702 The initiator of an IKE SA using EAP needs to be capable of extending
2703 the initial protocol exchange to at least ten IKE_AUTH exchanges in
2704 the event the responder sends notification messages and/or retries
2705 the authentication prompt. Once the protocol exchange defined by the
2706 chosen EAP authentication method has successfully terminated, the
2707 responder MUST send an EAP payload containing the Success message.
2708 Similarly, if the authentication method has failed, the responder
2709 MUST send an EAP payload containing the Failure message. The
2710 responder MAY at any time terminate the IKE exchange by sending an
2711 EAP payload containing the Failure message.
2713 Following such an extended exchange, the EAP AUTH payloads MUST be
2714 included in the two messages following the one containing the EAP
2717 When the initiator authentication uses EAP, it is possible that the
2718 contents of the IDi payload is used only for AAA routing purposes and
2719 selecting which EAP method to use. This value may be different from
2720 the identity authenticated by the EAP method. It is important that
2721 policy lookups and access control decisions use the actual
2722 authenticated identity. Often the EAP server is implemented in a
2723 separate AAA server that communicates with the IKEv2 responder. In
2724 this case, the authenticated identity has to be sent from the AAA
2725 server to the IKEv2 responder.
2727 2.17. Generating Keying Material for Child SAs
2729 A single Child SA is created by the IKE_AUTH exchange, and additional
2730 Child SAs can optionally be created in CREATE_CHILD_SA exchanges.
2731 Keying material for them is generated as follows:
2733 KEYMAT = prf+(SK_d, Ni | Nr)
2735 Where Ni and Nr are the nonces from the IKE_SA_INIT exchange if this
2736 request is the first Child SA created or the fresh Ni and Nr from the
2737 CREATE_CHILD_SA exchange if this is a subsequent creation.
2739 For CREATE_CHILD_SA exchanges including an optional Diffie-Hellman
2743 Kaufman, et al. Expires January 9, 2010 [Page 49]
2745 Internet-Draft IKEv2bis July 2009
2748 exchange, the keying material is defined as:
2750 KEYMAT = prf+(SK_d, g^ir (new) | Ni | Nr )
2752 where g^ir (new) is the shared secret from the ephemeral Diffie-
2753 Hellman exchange of this CREATE_CHILD_SA exchange (represented as an
2754 octet string in big endian order padded with zeros in the high-order
2755 bits if necessary to make it the length of the modulus).
2757 For ESP and AH, a single Child SA negotiation results in two security
2758 associations (one in each direction). Keying material MUST be taken
2759 from the expanded KEYMAT in the following order:
2761 o The encryption key (if any) for the SA carrying data from the
2762 initiator to the responder.
2764 o The authentication key (if any) for the SA carrying data from the
2765 initiator to the responder.
2767 o The encryption key (if any) for the SA carrying data from the
2768 responder to the initiator.
2770 o The authentication key (if any) for the SA carrying data from the
2771 responder to the initiator.
2773 Each cryptographic algorithm takes a fixed number of bits of keying
2774 material specified as part of the algorithm, or negotiated in SA
2775 payloads (see Section 2.13 for description of key lengths, and
2776 Section 3.3.5 for the definition of the Key Length transform
2779 2.18. Rekeying IKE SAs Using a CREATE_CHILD_SA Exchange
2781 The CREATE_CHILD_SA exchange can be used to rekey an existing IKE SA
2782 (see Section 2.8). New initiator and responder SPIs are supplied in
2783 the SPI fields in the Proposal structures inside the Security
2784 Association (SA) payloads (not the SPI fields in the IKE header).
2785 The TS payloads are omitted when rekeying an IKE SA. SKEYSEED for
2786 the new IKE SA is computed using SK_d from the existing IKE SA as
2789 SKEYSEED = prf(SK_d (old), g^ir (new) | Ni | Nr)
2791 where g^ir (new) is the shared secret from the ephemeral Diffie-
2792 Hellman exchange of this CREATE_CHILD_SA exchange (represented as an
2793 octet string in big endian order padded with zeros if necessary to
2794 make it the length of the modulus) and Ni and Nr are the two nonces
2795 stripped of any headers.
2799 Kaufman, et al. Expires January 9, 2010 [Page 50]
2801 Internet-Draft IKEv2bis July 2009
2804 The old and new IKE SA may have selected a different PRF. Because
2805 the rekeying exchange belongs to the old IKE SA, it is the old IKE
2806 SA's PRF that is used.
2808 The main reason for rekeying the IKE SA is to ensure that the
2809 compromise of old keying material does not provide information about
2810 the current keys, or vice versa. Therefore, implementations MUST
2811 perform a new Diffie-Hellman exchange when rekeying the IKE SA. In
2812 other words, an initiator MUST NOT propose the value "NONE" for the
2813 D-H transform, and a responder MUST NOT accept such a proposal. This
2814 means that a succesful exchange rekeying the IKE SA always includes
2815 the KEi/KEr payloads.
2817 The new IKE SA MUST reset its message counters to 0.
2819 SK_d, SK_ai, SK_ar, SK_ei, and SK_er are computed from SKEYSEED as
2820 specified in Section 2.14, using SPIi, SPIr, Ni, and Nr from the new
2823 2.19. Requesting an Internal Address on a Remote Network
2825 Most commonly occurring in the endpoint-to-security-gateway scenario,
2826 an endpoint may need an IP address in the network protected by the
2827 security gateway and may need to have that address dynamically
2828 assigned. A request for such a temporary address can be included in
2829 any request to create a Child SA (including the implicit request in
2830 message 3) by including a CP payload. Note, however, it is usual to
2831 only assign one IP address during the IKE_AUTH exchange. That
2832 address persists at least until the deletion of the IKE SA.
2834 This function provides address allocation to an IPsec Remote Access
2835 Client (IRAC) trying to tunnel into a network protected by an IPsec
2836 Remote Access Server (IRAS). Since the IKE_AUTH exchange creates an
2837 IKE SA and a Child SA, the IRAC MUST request the IRAS-controlled
2838 address (and optionally other information concerning the protected
2839 network) in the IKE_AUTH exchange. The IRAS may procure an address
2840 for the IRAC from any number of sources such as a DHCP/BOOTP server
2841 or its own address pool.
2844 -------------------------------------------------------------------
2845 HDR, SK {IDi, [CERT,]
2846 [CERTREQ,] [IDr,] AUTH,
2847 CP(CFG_REQUEST), SAi2,
2849 <-- HDR, SK {IDr, [CERT,] AUTH,
2850 CP(CFG_REPLY), SAr2,
2855 Kaufman, et al. Expires January 9, 2010 [Page 51]
2857 Internet-Draft IKEv2bis July 2009
2860 In all cases, the CP payload MUST be inserted before the SA payload.
2861 In variations of the protocol where there are multiple IKE_AUTH
2862 exchanges, the CP payloads MUST be inserted in the messages
2863 containing the SA payloads.
2865 CP(CFG_REQUEST) MUST contain at least an INTERNAL_ADDRESS attribute
2866 (either IPv4 or IPv6) but MAY contain any number of additional
2867 attributes the initiator wants returned in the response.
2869 For example, message from initiator to responder:
2873 TSi = (0, 0-65535,0.0.0.0-255.255.255.255)
2874 TSr = (0, 0-65535,0.0.0.0-255.255.255.255)
2876 NOTE: Traffic Selectors contain (protocol, port range, address
2879 Message from responder to initiator:
2882 INTERNAL_ADDRESS(192.0.2.202)
2883 INTERNAL_NETMASK(255.255.255.0)
2884 INTERNAL_SUBNET(192.0.2.0/255.255.255.0)
2885 TSi = (0, 0-65535,192.0.2.202-192.0.2.202)
2886 TSr = (0, 0-65535,192.0.2.0-192.0.2.255)
2888 All returned values will be implementation dependent. As can be seen
2889 in the above example, the IRAS MAY also send other attributes that
2890 were not included in CP(CFG_REQUEST) and MAY ignore the non-
2891 mandatory attributes that it does not support.
2893 The FAILED_CP_REQUIRED notification is sent by responder in the case
2894 where CP(CFG_REQUEST) was expected but not received, and so is a
2895 conflict with locally configured policy. There is no associated
2898 The responder MUST NOT send a CFG_REPLY without having first received
2899 a CP(CFG_REQUEST) from the initiator, because we do not want the IRAS
2900 to perform an unnecessary configuration lookup if the IRAC cannot
2901 process the REPLY. In the case where the IRAS's configuration
2902 requires that CP be used for a given identity IDi, but IRAC has
2903 failed to send a CP(CFG_REQUEST), IRAS MUST fail the request, and
2904 terminate the IKE exchange with a FAILED_CP_REQUIRED error. The
2905 FAILED_CP_REQUIRED is not fatal to the IKE SA; it simply causes the
2906 Child SA creation fail. The initiator can fix this by later starting
2907 a new configuration payload request.
2911 Kaufman, et al. Expires January 9, 2010 [Page 52]
2913 Internet-Draft IKEv2bis July 2009
2916 2.19.1. Configuration Payloads
2918 Editor's note: some of this sub-section is redundant and will go away
2919 in the next version of the document.
2921 In support of the scenario described in Section 1.1.3, an initiator
2922 may request that the responder assign an IP address and tell the
2923 initiator what it is. That request is done using configuration
2924 payloads, not traffic selectors. An address in a TSi payload in a
2925 response does not mean that the responder has assigned that address
2926 to the initiator: it only means that if packets matching these
2927 traffic selectors are sent by the initiator, IPsec processing can be
2928 performed as agreed for this SA.
2930 Configuration payloads are of type CFG_REQUEST/CFG_REPLY or CFG_SET/
2931 CFG_ACK (see CFG Type in the payload description below). CFG_REQUEST
2932 and CFG_SET payloads may optionally be added to any IKE request. The
2933 IKE response MUST include either a corresponding CFG_REPLY or CFG_ACK
2934 or a Notify payload with an error type indicating why the request
2935 could not be honored. An exception is that a minimal implementation
2936 MAY ignore all CFG_REQUEST and CFG_SET payloads, so a response
2937 message without a corresponding CFG_REPLY or CFG_ACK MUST be accepted
2938 as an indication that the request was not supported.
2940 "CFG_REQUEST/CFG_REPLY" allows an IKE endpoint to request information
2941 from its peer. If an attribute in the CFG_REQUEST Configuration
2942 Payload is not zero-length, it is taken as a suggestion for that
2943 attribute. The CFG_REPLY Configuration Payload MAY return that
2944 value, or a new one. It MAY also add new attributes and not include
2945 some requested ones. Requestors MUST ignore returned attributes that
2946 they do not recognize.
2948 Some attributes MAY be multi-valued, in which case multiple attribute
2949 values of the same type are sent and/or returned. Generally, all
2950 values of an attribute are returned when the attribute is requested.
2951 For some attributes (in this version of the specification only
2952 internal addresses), multiple requests indicates a request that
2953 multiple values be assigned. For these attributes, the number of
2954 values returned SHOULD NOT exceed the number requested.
2956 If the data type requested in a CFG_REQUEST is not recognized or not
2957 supported, the responder MUST NOT return an error type but rather
2958 MUST either send a CFG_REPLY that MAY be empty or a reply not
2959 containing a CFG_REPLY payload at all. Error returns are reserved
2960 for cases where the request is recognized but cannot be performed as
2961 requested or the request is badly formatted.
2963 "CFG_SET/CFG_ACK" allows an IKE endpoint to push configuration data
2967 Kaufman, et al. Expires January 9, 2010 [Page 53]
2969 Internet-Draft IKEv2bis July 2009
2972 to its peer. In this case, the CFG_SET Configuration Payload
2973 contains attributes the initiator wants its peer to alter. The
2974 responder MUST return a Configuration Payload if it accepted any of
2975 the configuration data and it MUST contain the attributes that the
2976 responder accepted with zero-length data. Those attributes that it
2977 did not accept MUST NOT be in the CFG_ACK Configuration Payload. If
2978 no attributes were accepted, the responder MUST return either an
2979 empty CFG_ACK payload or a response message without a CFG_ACK
2980 payload. There are currently no defined uses for the CFG_SET/CFG_ACK
2981 exchange, though they may be used in connection with extensions based
2982 on Vendor IDs. An minimal implementation of this specification MAY
2983 ignore CFG_SET payloads.
2985 Extensions via the CP payload should not be used for general purpose
2986 management. Its main intent is to provide a bootstrap mechanism to
2987 exchange information within IPsec from IRAS to IRAC. While it MAY be
2988 useful to use such a method to exchange information between some
2989 Security Gateways (SGW) or small networks, existing management
2990 protocols such as DHCP [DHCP], RADIUS [RADIUS], SNMP, or LDAP [LDAP]
2991 should be preferred for enterprise management as well as subsequent
2992 information exchanges.
2994 2.20. Requesting the Peer's Version
2996 An IKE peer wishing to inquire about the other peer's IKE software
2997 version information MAY use the method below. This is an example of
2998 a configuration request within an INFORMATIONAL exchange, after the
2999 IKE SA and first Child SA have been created.
3001 An IKE implementation MAY decline to give out version information
3002 prior to authentication or even after authentication to prevent
3003 trolling in case some implementation is known to have some security
3004 weakness. In that case, it MUST either return an empty string or no
3005 CP payload if CP is not supported.
3008 -------------------------------------------------------------------
3009 HDR, SK{CP(CFG_REQUEST)} -->
3010 <-- HDR, SK{CP(CFG_REPLY)}
3013 APPLICATION_VERSION("")
3015 CP(CFG_REPLY) APPLICATION_VERSION("foobar v1.3beta, (c) Foo Bar
3023 Kaufman, et al. Expires January 9, 2010 [Page 54]
3025 Internet-Draft IKEv2bis July 2009
3028 2.21. Error Handling
3030 There are many kinds of errors that can occur during IKE processing.
3031 If a request is received that is badly formatted or unacceptable for
3032 reasons of policy (e.g., no matching cryptographic algorithms), the
3033 response MUST contain a Notify payload indicating the error. If an
3034 error occurs outside the context of an IKE request (e.g., the node is
3035 getting ESP messages on a nonexistent SPI), the node SHOULD initiate
3036 an INFORMATIONAL exchange with a Notify payload describing the
3039 Errors that occur before a cryptographically protected IKE SA is
3040 established must be handled very carefully. There is a trade-off
3041 between wanting to be helpful in diagnosing a problem and responding
3042 to it and wanting to avoid being a dupe in a denial of service attack
3043 based on forged messages.
3045 If a node receives a message on UDP port 500 or 4500 outside the
3046 context of an IKE SA known to it (and not a request to start one), it
3047 may be the result of a recent crash of the node. If the message is
3048 marked as a response, the node MAY audit the suspicious event but
3049 MUST NOT respond. If the message is marked as a request, the node
3050 MAY audit the suspicious event and MAY send a response. If a
3051 response is sent, the response MUST be sent to the IP address and
3052 port from whence it came with the same IKE SPIs and the Message ID
3053 copied. The response MUST NOT be cryptographically protected and
3054 MUST contain a Notify payload indicating INVALID_IKE_SPI. The
3055 INVALID_IKE_SPI notification indicates an IKE message was received
3056 with an unrecognized destination SPI; this usually indicates&nbs
projects / strongswan.git / blob