Internet-Draft | Multiple Key Exchanges in IKEv2 | September 2021 |
Tjhai, et al. | Expires 3 April 2022 | [Page] |
This document describes how to extend the Internet Key Exchange Protocol Version 2 (IKEv2) to allow multiple key exchanges to take place while computing a shared secret during a Security Association (SA) setup. The primary application of this feature in IKEv2 is the ability to perform one or more post-quantum key exchanges in conjunction with the classical (Elliptic Curve) Diffie-Hellman key exchange, so that the resulting shared key is resistant against quantum computer attacks. Another possible application is the ability to combine several key exchanges in situations when no single key exchange algorithm is trusted by both initiator and responder.¶
This document updates RFC7296 by renaming a transform type 4 from "Diffie-Hellman Group (D-H)" to "Key Exchange Method (KE)" and renaming a field in the Key Exchange Payload from "Diffie-Hellman Group Num" to "Key Exchange Method". It also renames an IANA registry for this transform type from "Transform Type 4 - Diffie-Hellman Group Transform IDs" to "Transform Type 4 - Key Exchange Method Transform IDs". These changes generalize key exchange algorithms that can be used in IKEv2.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 3 April 2022.¶
Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
Internet Key Exchange Protocol (IKEv2) as specified in [RFC7296] uses the Diffie-Hellman (DH) or Elliptic Curve Diffie-Hellman (ECDH) algorithm to establish a shared secret between an initiator and a responder. The security of the DH and ECDH algorithms relies on the difficulty to solve a discrete logarithm problem in multiplicative and elliptic curve groups respectively when the order of the group parameter is large enough. While solving such a problem remains difficult with current computing power, it is believed that general purpose quantum computers will be able to solve this problem, implying that the security of IKEv2 is compromised. There are, however, a number of cryptosystems that are conjectured to be resistant against quantum computer attack. This family of cryptosystems is known as post-quantum cryptography (PQC). It is sometimes also referred to as quantum-safe cryptography (QSC) or quantum-resistant cryptography (QRC).¶
This document describes a method to perform multiple successive key exchanges in IKEv2. It allows integration of QSC in IKEv2, while maintaining backwards compatibility, to derive a set of IKE keys that is resistant to quantum computer attacks. This extension allows the negotiation of one or more QSC algorithm to exchange data, in addition to the existing DH or ECDH key exchange data. We believe that the feature of using more than one post-quantum algorithms is important as many of these algorithms are relatively new and there may be a need to hedge the security risk with multiple key exchange data from several distinct QSC algorithms.¶
The secrets established from each key exchange are combined in a way such that should the post-quantum secrets not be present, the derived shared secret is equivalent to that of the standard IKEv2; on the other hand, a post-quantum shared secret is obtained if both classical and post-quantum key exchange data are present. This extension also applies to key exchanges in IKE Security Associations (SAs) for Encapsulating Security Payload (ESP) [RFC4303] or Authentication Header (AH) [RFC4302], i.e. Child SAs, in order to provide a stronger guarantee of forward security.¶
Some post-quantum key exchange payloads may have sizes larger than the standard maximum transmission unit (MTU) size, and therefore there could be issues with fragmentation at the IP layer. IKE does allow transmission over TCP where fragmentation is not an issue [RFC8229]; however, we believe that a UDP-based solution will be required too. IKE does have a mechanism to handle fragmentation within UDP [RFC7383], however that is only applicable to messages exchanged after the IKE_SA_INIT exchange. To use this mechanism, this specification relies on the IKE_INTERMEDIATE exchange as outlined in [I-D.ietf-ipsecme-ikev2-intermediate]. With this mechanism, we do an initial key exchange, using a smaller, possibly non-quantum resistant primitive, such as ECDH. Then, before we do the IKE_AUTH exchange, we perform one or more IKE_INTERMEDIATE exchanges, each of which contains an additional key exchange. As the IKE_INTERMEDIATE exchange is encrypted, the IKE fragmentation protocol [RFC7383] can be used. The IKE SK_* values are updated after each exchange, and so the final IKE SA keys depend on all the key exchanges, hence they are secure if any of the key exchanges are secure.¶
Note that readers should consider the approach defined in this document as providing a long term solution in upgrading the IKEv2 protocol to support post-quantum algorithms. A short term solution to make IKEv2 key exchange quantum secure is to use post-quantum pre-shared keys as discussed in [RFC8784].¶
Note also, that the proposed approach of performing multiple successive key exchanges in such a way that resulting session keys depend on all of them is not limited to achieving quantum resistance only. It can also be used when all the performed key exchanges are classical (EC)DH ones, where for some reasons (e.g. policy requirements) it is essential to perform multiple of them.¶
This draft does not attempt to address key exchanges with KE payloads longer than 64k; the current IKE payload format does not allow that as a possibility. At the current time, it appears likely that there are a number of key exchanges available that would not require such a requirement. However, if such a requirement is needed, [I-D.tjhai-ikev2-beyond-64k-limit] discusses approaches that should be taken to exchange huge payloads.¶
RFC EDITOR PLEASE DELETE THIS SECTION.¶
Changes in this draft in each version iterations.¶
draft-ietf-ipsecme-ikev2-multiple-ke-04¶
draft-ietf-ipsecme-ikev2-multiple-ke-03¶
draft-ietf-ipsecme-ikev2-multiple-ke-02¶
draft-ietf-ipsecme-ikev2-multiple-ke-01¶
draft-ietf-ipsecme-ikev2-multiple-ke-00¶
draft-tjhai-ipsecme-hybrid-qske-ikev2-04¶
draft-tjhai-ipsecme-hybrid-qske-ikev2-03¶
draft-tjhai-ipsecme-hybrid-qske-ikev2-02¶
draft-tjhai-ipsecme-hybrid-qske-ikev2-01¶
draft-tjhai-ipsecme-hybrid-qske-ikev2-00¶
The remainder of this document is organized as follows. Section 2 summarizes design criteria. Section 3 describes how multiple key exchanges are performed between two IKE peers and how keying materials are derived for both SAs and Child SAs. A summary of alternative approaches that have been considered, but later discarded, are described in Appendix B. Section 4 discusses IANA considerations for the namespaces introduced in this document, and lastly Section 5 discusses security considerations.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
The design of the proposed extension is driven by the following criteria:¶
Most post-quantum key agreement algorithms are relatively new, and thus are not fully trusted. There are also many proposed algorithms, with different trade-offs and relying on different hard problems. The concern is that some of these hard problems may turn out to be easier to solve than anticipated and thus the key agreement algorithm may not be as secure as expected. A hybrid solution, when multiple key exchanges are performed and the calculated shared key depends on all of them, allows us to deal with this uncertainty by combining a classical key exchange with a post-quantum one, as well as leaving open the possibility of multiple post-quantum key exchanges.¶
In order to be able to use IKE fragmentation [RFC7383] for those key exchanges that may have long public keys, the proposed framework utilizes the IKE_INTERMEDIATE exchange defined in [I-D.ietf-ipsecme-ikev2-intermediate]. The initial IKE_INIT messages do not have any inherent fragmentation support within IKE; however that can include a relatively short KE payload. The additional key exchanges are performed using IKE_INTERMEDIATE messages; because these messages are encrypted, the standard IKE fragmentation mechanism is available.¶
In order to minimize communication overhead, only the key shares that are agreed to be used are actually exchanged. To negotiate additional key exchanges seven new Transform Types are defined. These transforms share allowed Transform IDs with Transform Type 4.¶
We assume that new Transform Type 4 identifiers will be assigned later to the various post-quantum key exchanges. We specifically do not make a distinction between classical (DH and ECDH) and post-quantum key exchanges, nor post-quantum algorithms which are true key exchanges versus post-quantum algorithms that act as key transport mechanisms; all are treated equivalently by the protocol. To be more specific, this document renames Transform Type 4 from "Diffie-Hellman Group (D-H)" to "Key Exchange Method (KE)" and renames a field in the Key Exchange Payload from "Diffie-Hellman Group Num" to "Key Exchange Method". The corresponding IANA registry is also renamed from "Diffie-Hellman Group Transform IDs" to "Key Exchange Method Transform IDs".¶
The fact, that newly defined transforms share the same registry for possible Transform IDs with Transform Type 4, allows additional key exchanges to be of any type - either post-quantum or classical (EC)DH one. This approach allows any combination of defined key exchange methods to take place. This also allows performing a single post-quantum key exchange in the IKE_SA_INIT without additional key exchanges, provided that IP fragmentation is not an issue and that hybrid key exchange is not needed.¶
The SA payload in the IKE_SA_INIT message includes one or more newly defined transforms which represent the extra key exchange policy required by the initiator. The responder follows the usual IKEv2 negotiation rules: it selects a single transform of each type, and returns all of them in the IKE_SA_INIT response message.¶
Then, provided that additional key exchanges are negotiated, the initiator and the responder perform one or more IKE_INTERMEDIATE exchanges. Then the IKE_AUTH exchange authenticates peers and completes IKE SA establishment.¶
Initiator Responder --------------------------------------------------------------------- <-- IKE_SA_INIT (additional key exchanges negotiation) --> <-- {IKE_INTERMEDIATE (additional key exchange)} --> ... <-- {IKE_INTERMEDIATE (additional key exchange)} --> <-- {IKE_AUTH} -->¶
Note, that this document assumes, that each key exchange method requires one round trip and consumes exactly one IKE_INTERMEDIATE exchange. This assumption is valid for all classic key exchange methods defined so far and for all post-quantum methods currently known. For hypothetical future key exchange methods requiring multiple round trips to complete, a separate document should define how such methods are splitted into several IKE_INTERMEDIATE exchanges.¶
In the simplest case, the initiator is happy with a single key exchange (and has no interest in supporting multiple), and it is not concerned with possible fragmentation of the IKE_SA_INIT messages (either because the key exchange it selects is small enough not to fragment, or the initiator is confident that fragmentation will be handled either by IP fragmentation, or transport via TCP).¶
In this case, the initiator performs the IKE_SA_INIT as usual, inserting a preferred key exchange (which is possibly a post-quantum algorithm) as the listed Transform Type 4, and including the initiator KE payload. If the responder accepts the policy, it responds with an IKE_SA_INIT response, and IKE continues as usual.¶
If the initiator desires to negotiate multiple key exchanges, then the initiator uses the protocol listed below.¶
Multiple key exchanges are negotiated using the standard IKEv2 mechanism, via SA payload. For this purpose seven new transform types, namely Additional Key Exchange 1 (<TBA by IANA>), Additional Key Exchange 2 (<TBA by IANA>), Additional Key Exchange 3 (<TBA by IANA>), Additional Key Exchange 4 (<TBA by IANA>), Additional Key Exchange 5 (<TBA by IANA>), Additional Key Exchange 6 (<TBA by IANA>) and Additional Key Exchange 7 (<TBA by IANA>) are defined. They are collectively called Additional Key Exchange transforms in this document and have slightly different semantics than existing IKEv2 transform types. They are interpreted as an indication of additional key exchanges methods that peers agreed to perform in a series of IKE_INTERMEDIATE exchanges following the IKE_SA_INIT exchange. The allowed transform IDs for these transform types are the same as IDs for the Transform Type 4, so they all share a single IANA registry for transform IDs.¶
Key exchange method negotiated via Transform Type 4 always takes place in the IKE_SA_INIT exchange, as defined in [RFC7296]. Additional key exchanges negotiated via newly defined transforms MUST take place in a series of IKE_INTERMEDIATE exchanges following the IKE_SA_INIT exchange, performed in an order of the values of their transform types, so that key exchange negotiated using Transform Type n always precedes that of Transform Type n + 1. Each additional key exchange method MUST be fully completed before the next one is started.¶
Note that with this semantics, Additional Key Exchanges transforms are not associated with any particular type of key exchange and do not have any specific per transform type transform IDs IANA registry. Instead they all share a single registry for transform IDs - "Key Exchange Method Transform IDs", as well as Transform Type 4. All new key exchange algorithms (both classical or post-quantum) should be added to this registry. This approach gives peers flexibility in defining the ways they want to combine different key exchange methods.¶
When forming a proposal the initiator adds transforms for the IKE_SA_INIT exchange using Transform Type 4. In most cases they will contain classical key exchange methods (DH or ECDH), however it is not a requirement. Additional key exchange methods are proposed using Additional Key Exchanges transform types. All these transform types are optional, the initiator is free to select any of them for proposing additional key exchange methods. Consequently, if none of Additional Key Exchange transforms are included in the proposal, then this proposal indicates performing standard IKEv2, as defined in [RFC7296]. If the initiator includes any Additional Key Exchanges transform in the proposal, the responder MUST select one of the algorithms proposed using this type. A transform ID NONE MAY be added to those transform types which contain key exchange methods that the initiator believes are optional according to its local policy.¶
The responder performs negotiation using standard IKEv2 procedure described in Section 3.3 of [RFC7296]. However, for the Additional Key Exchange types the responder's choice MUST NOT contain equal algorithms, except for transform ID of NONE. An algorithm is represented as a transform, in some cases the transform could include a set of associated attributes that define details of the algorithm. In this case two ransforms can be the same, but the attributes must be different. Additionally, the order of the attributes does not affect the equality of the algorithm, so two transforms (ID=alg1,ATTR1=attr1,ATTR2=attr2) and (ID=alg1,ATTR2=attr2,ATTR1=attr1) define the same algorithm.¶
If the responder selected NONE for some Additional Key Exchange types (provided they were proposed by the initiator), then the corresponding IKE_INTERMEDIATE exchanges should not take place. The IKE_INTERMEDIATE exchanges MUST only be performed for Additional Key Exchange types containing non-NONE responders choices. It means that if the initiator includes NONE in all Additional Key Exchange transforms and the responder selects this value for all of them, then no IKE_INTERMEDIATE exchanges will take place between the peers. perform additional key exchanges will take place (note that they still may take place for other purposes).¶
Below is an example of the SA payload in the initiator's IKE_SA_INIT request message. Here we use an abbreviation AKE1, AKE 2 etc. to denote Additional Key Exchange 1, Additional Key Exchange 2 etc. transforms, that this document defines, and an abbreviation KE for the Key Exchange transform, that this document renames from the Diffie-Hellman Group transform. We also use not yet defined Transform IDs PQ_KEM_1, PQ_KEM_2 and PQ_KEM_3 to denote some of popular post-quantum key exchange methods.¶
SA Payload | +--- Proposal #1 ( Proto ID = IKE(1), SPI size = 8, | 9 transforms, SPI = 0x35a1d6f22564f89d ) | +-- Transform ENCR ( ID = ENCR_AES_GCM_16 ) | +-- Attribute ( Key Length = 256 ) | +-- Transform KE ( ID = 4096-bit MODP Group ) | +-- Transform PRF ( ID = PRF_HMAC_SHA2_256 ) | +-- Transform AKE2 ( ID = PQ_KEM_1 ) | +-- Transform AKE2 ( ID = PQ_KEM_2 ) | +-- Transform AKE3 ( ID = PQ_KEM_1 ) | +-- Transform AKE3 ( ID = PQ_KEM_2 ) | +-- Transform AKE5 ( ID = PQ_KEM_3 ) | +-- Transform AKE5 ( ID = NONE )¶
In this example the initiator proposes to perform initial key exchange using 4096-bit MODP group following by two mandatory additional key exchanges using PQ_KEM_1 and PQ_KEM_2 methods in any order, following by additional key exchange using PQ_KEM_3 method that may be omitted.¶
The responder might return the following SA payload, indicating that it agrees to perform two additional key exchanges PQ_KEM_2 followed by PQ_KEM_1 and doesn't want to perform PQ_KEM_3 additionally.¶
SA Payload | +--- Proposal #1 ( Proto ID = IKE(1), SPI size = 8, | 6 transforms, SPI = 0x8df52b331a196e7b ) | +-- Transform ENCR ( ID = ENCR_AES_GCM_16 ) | +-- Attribute ( Key Length = 256 ) | +-- Transform KE ( ID = 4096-bit MODP Group ) | +-- Transform PRF ( ID = PRF_HMAC_SHA2_256 ) | +-- Transform AKE2 ( ID = PQ_KEM_2 ) | +-- Transform AKE3 ( ID = PQ_KEM_1 ) | +-- Transform AKE5 ( ID = NONE )¶
If the initiator includes any Additional Key Exchanges transform types into SA payload in the IKE_SA_INIT exchange request message, it MUST also negotiate using IKE_INTERMEDIATE exchange as described in [I-D.ietf-ipsecme-ikev2-intermediate], by including INTERMEDIATE_EXCHANGE_SUPPORTED notification in the same message. If the responder agrees to use additional key exchanges while establishing initial IKE SA, it MUST also return this notification in the IKE_SA_INIT response message, thus confirming that IKE_INTERMEDIATE exchange is supported and will be used for transferring additional key exchange data. If the IKE_INTERMEDIATE exchange is not negotiated, then the peers MUST treat any Additional Key Exchange transforms in the IKE_SA_INIT exchange messages as unknown transform types and skip the proposals they appear in. If no other proposals are present in the SA payload, the peers will proceed as when no proposal is chosen (i.e. the responder will send NO_PROPOSAL_CHOSEN notification).¶
Initiator Responder --------------------------------------------------------------------- HDR, SAi1(.. AKE*...), KEi1, Ni, N(INTERMEDIATE_EXCHANGE_SUPPORTED) ---> HDR, SAr1(.. AKE*...), KEr1, Nr, [CERTREQ], <--- N(INTERMEDIATE_EXCHANGE_SUPPORTED)¶
For each additional key exchange agreed to in the IKE_SA_INIT exchange, the initiator and the responder perform IKE_INTERMEDIATE exchange, as described in [I-D.ietf-ipsecme-ikev2-intermediate].¶
Initiator Responder --------------------------------------------------------------------- HDR, SK {KEi(n)} --> <-- HDR, SK {KEr(n)}¶
The initiator sends key exchange data in the KEi(n) payload. This packet is protected with the current SK_ei/SK_ai keys.¶
On receiving this, the responder sends back key exchange payload KEr(n); again, this packet is protected with the current SK_er/SK_ar keys.¶
The former "Diffie-Hellman Group Num" (now called "Key Exchange Method") field in the KEi(n) and KEr(n) payloads MUST match the n-th negotiated additional key exchange.¶
Once this exchange is done, both sides compute an updated keying material:¶
SKEYSEED(n) = prf(SK_d(n-1), SK(n) | Ni | Nr)¶
where SK(n) is the resulting shared secret of this key exchange, Ni and Nr are nonces from the IKE_SA_INIT exchange and SK_d(n-1) is the last generated SK_d, (derived from the previous IKE_INTERMEDIATE exchange, or the IKE_SA_INIT if there have not already been any IKE_INTERMEDIATE exchanges). Then, SK_d, SK_ai, SK_ar, SK_ei, SK_er, SK_pi, SK_pr are updated as:¶
{SK_d(n) | SK_ai(n) | SK_ar(n) | SK_ei(n) | SK_er(n) | SK_pi(n) | SK_pr(n)} = prf+ (SKEYSEED(n), Ni | Nr | SPIi | SPIr)¶
Both the initiator and the responder use these updated key values in the next exchange (IKE_INTERMEDIATE or IKE_AUTH).¶
After all IKE_INTERMEDIATE exchanges have completed, the initiator and the responder perform an IKE_AUTH exchange. This exchange is the standard IKE exchange, except that the initiator and responder signed octets are modified as described in [I-D.ietf-ipsecme-ikev2-intermediate].¶
The CREATE_CHILD_SA exchange is used in IKEv2 for the purposes of creating additional Child SAs, rekeying them and rekeying IKE SA itself. When creating or rekeying Child SAs, the peers may optionally perform a Diffie-Hellman key exchange to add a fresh entropy into the session keys. In case of IKE SA rekey, the key exchange is mandatory. Peers supporting this specification may want to use multiple key exchanges in these situations.¶
Using multiple key exchanges with CREATE_CHILD_SA exchange is negotiated similarly as in initial exchange, see Section 3.2.1. If the initiator includes any Additional Key Exchanges transform in the SA payload (along with Transform Type 4) and the responder agrees to perform additional key exchanges, then the additional key exchanges are performed in a series of new IKE_FOLLOWUP_KE exchanges that follows the CREATE_CHILD_SA exchange. The IKE_FOLLOWUP_KE exchange is introduced as a dedicated exchange for transferring data of additional key exchanges following the key exchange performed in the CREATE_CHILD_SA. Its Exchange Type is <TBA by IANA>.¶
Key exchange negotiated via Transform Type 4 always takes place in the CREATE_CHILD_SA exchange, as per IKEv2 specification. Additional key exchanges are performed in an order of the values of their transform types, so that key exchange negotiated using Transform Type n always precedes key exchange negotiated using Transform Type n + 1. Each additional key exchange method MUST be fully completed before the next one is started. Note, that this document assumes, that each key exchange method consumes exactly one IKE_FOLLOWUP_KE exchange. For the methods requiring multiple round trips, a separate document should define how such methods are splitted into several IKE_FOLLOWUP_KE exchanges.¶
Since after IKE SA is created the window size may be greater than one and multiple concurrent exchanges may be in progress, it is essential to link the IKE_FOLLOWUP_KE exchanges together and with the corresponding CREATE_CHILD_SA exchange. A new status type notification ADDITIONAL_KEY_EXCHANGE is used for this purpose. Its Notify Message Type is <TBA by IANA>, Protocol ID and SPI Size are both set to 0. The data associated with this notification is a blob meaningful only to the responder, so that the responder can correctly link successive exchanges. For the initiator the content of this notification is an opaque blob.¶
The responder MUST include this notification in a CREATE_CHILD_SA or IKE_FOLLOWUP_KE response message in case the next IKE_FOLLOWUP_KE exchange is expected, filling it with some data that would allow linking current exchange to the next one. The initiator MUST send back this notification intact in the request message of the next IKE_FOLLOWUP_KE exchange.¶
Below is an example of CREATE_CHILD_SA exchange followed by three additional key exchanges.¶
Initiator Responder --------------------------------------------------------------------- HDR(CREATE_CHILD_SA), SK {SA, Ni, KEi} --> <-- HDR(CREATE_CHILD_SA), SK {SA, Nr, KEr, N(ADDITIONAL_KEY_EXCHANGE)(link1)} HDR(IKE_FOLLOWUP_KE), SK {KEi(1), N(ADDITIONAL_KEY_EXCHANGE)(link1)} --> <-- HDR(IKE_FOLLOWUP_KE), SK {KEr(1), N(ADDITIONAL_KEY_EXCHANGE)(link2)} HDR(IKE_FOLLOWUP_KE), SK {KEi(2), N(ADDITIONAL_KEY_EXCHANGE)(link2)} --> <-- HDR(IKE_FOLLOWUP_KE), SK {KEr(2), N(ADDITIONAL_KEY_EXCHANGE)(link3)} HDR(IKE_FOLLOWUP_KE), SK {KEi(3), N(ADDITIONAL_KEY_EXCHANGE)(link3)} --> <-- HDR(IKE_FOLLOWUP_KE), SK {KEr(3)}¶
The former "Diffie-Hellman Group Num" (now called "Key Exchange Method") field in the KEi(n) and KEr(n) payloads MUST match the n-th negotiated additional key exchange.¶
It is possible that due to some unexpected events (e.g. reboot) the initiator may lose its state and forget that it is in the process of performing additional key exchanges and thus never start the remaining IKE_FOLLOWUP_KE exchanges. The responder MUST handle this situation gracefully and delete the associated state if it does not receive the next expected IKE_FOLLOWUP_KE request after some reasonable period of time.¶
If responder receives IKE_FOLLOWUP_KE request containing ADDITIONAL_KEY_EXCHANGE notification and the content of this notify does not correspond to any active key exchange state the responder has, it MUST send back a new error type notification STATE_NOT_FOUND. This is a non-fatal error notification, its Notify Message Type is <TBA by IANA>, Protocol ID and SPI Size are both set to 0 and the data is empty. If the initiator receives this notification in response to IKE_FOLLOWUP_KE exchange performing additional key exchange, it MUST cancel this exchange and MUST treat the whole series of exchanges started from the CREATE_CHILD_SA exchange as failed. In most cases, the receipt of this notification is caused by premature deletion of the corresponding state on the responder (the time period between IKE_FOLLOWUP_KE exchanges appeared too long from the responder's point of view, e.g. due to a temporary network failure). After receiving this notification the initiator MAY start a new CREATE_CHILD_SA exchange (eventually followed by the IKE_FOLLOWUP_KE exchanges) to retry the failed attempt. If the initiator continues to receive STATE_NOT_FOUND notifications after several retries, it MUST treat this situation as a fatal error and delete IKE SA by sending a DELETE payload.¶
When rekeying IKE SA or Child SA, it is possible that the peers start doing this at the same time, which is called simultaneous rekeying. Sections 2.8.1 and 2.8.2 of [RFC7296] describe how IKEv2 handles this situation. In a nutshell IKEv2 follows the rule that if in case of simultaneous rekeying two identical new IKE SAs (or two pairs of Child SAs) are created, then one of them should be deleted. Which one is to be deleted is determined by comparing the values of four nonces, that were used in the colliding CREATE_CHILD_SA exchanges - the IKE SA (or pair of Child SAs) that was created by the exchange in which the smallest nonce was used should be deleted by the initiator of this exchange.¶
With multiple key exchanges the SAs are not yet created when the CREATE_CHILD_SA is completed, they would be created only after the series of IKE_FOLLOWUP_KE exchanges is finished. For this reason if additional key exchanges were negotiated in the CREATE_CHILD_SA initiated by the losing side, there is nothing to delete and this side just stops the rekeying process - this side MUST NOT initiate IKE_FOLLOWUP_KE exchange with next key exchange.¶
In most cases, rekey collisions are resolved in the CREATE_CHILD_SA exchange. However, a situation may occur when due to packet loss, one of the peers receives the CREATE_CHILD_SA message requesting rekey of SA that is already being rekeyed by this peer (i.e. the CREATE_CHILD_SA exchange initiated by this peer has been already completed and the series of IKE_FOLLOWUP_KE exchanges is in progress). In this case, TEMPORARY_FAILURE notification MUST be sent in response to such a request.¶
If multiple key exchanges were negotiated in the CREATE_CHILD_SA exchange, then the resulting keys are computed as follows. In case of IKE SA rekey:¶
SKEYSEED = prf(SK_d, SK(0) | Ni | Nr | SK(1) | ... SK(n))¶
In case of Child SA creation or rekey:¶
KEYMAT = prf+ (SK_d, SK(0) | Ni | Nr | SK(1) | ... SK(n))¶
In both cases SK_d is from existing IKE SA; SK(0), Ni, Nr are the shared key and nonces from the CREATE_CHILD_SA respectively; SK(1)...SK(n) are the shared keys from additional key exchanges.¶
It is also possible to establish a fully quantum-resistant IKE SAs from additional key exchanges without using IKE_INTERMEDIATE exchanges. In this case, the IKE SA created from IKE_SA_INIT exchange can be immediately rekeyed with CREATE_CHILD_SA using additional key exchanges and IKE_FOLLOWUP_KE message to carry the key exchange payload. If only classical key exchange method is used in the IKE_SA_INIT message, the very first Child SA created in IKE_AUTH will not be quantum resistant. Consequently, if the peers' local policy requires that all Child SAs should be fully-protected, then the peers can avoid creating the very first Child SA by adopting [RFC6023]. In this case, the peers exchange CHILDLESS_IKEV2_SUPPORTED notification in the IKE_SA_INIT exchange and a fully-protected Child SA can be created with CREATE_CHILD_SA using additional key exchanges.¶
Note that if the initial IKE SA is used to transfer sensitive information, then this information will not be protected using the additional (e.g. quantum safe) key exchanges, so this scenario may be inappropriate. One such example is in G-IKEv2 protocol [I-D.ietf-ipsecme-g-ikev2] where cryptographic materials are exchanged in IKE_SA_INIT messages between group member and the group controller.¶
This document adds new exchange type into the "IKEv2 Exchange Types" registry:¶
<TBA> IKE_FOLLOWUP_KE¶
This document renames Transform Type 4 defined in "Transform Type Values" registry from "Diffie-Hellman Group (D-H)" to "Key Exchange Method (KE)".¶
This document renames IKEv2 registry "Transform Type 4 - Diffie-Hellman Group Transform IDs" to "Transform Type 4 - Key Exchange Method Transform IDs".¶
This document adds the following Transform Types to the "Transform Type Values" registry:¶
Type Description Used In ----------------------------------------------------------------- <TBA> Additional Key Exchange 1 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 2 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 3 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 4 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 5 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 6 (optional in IKE, AH, ESP) <TBA> Additional Key Exchange 7 (optional in IKE, AH, ESP)¶
This document defines a new Notify Message Type in the "Notify Message Types - Status Types" registry:¶
<TBA> ADDITIONAL_KEY_EXCHANGE¶
and a new Notify Message Type in the "Notify Message Types - Error Types" registry:¶
<TBA> STATE_NOT_FOUND¶
The key length of the Encryption Algorithm (Transform Type 1), the Pseudorandom Function (Transform Type 2) and the Integrity Algorithm (Transform Type 3), all have to be of sufficient length to prevent attacks using Grover's algorithm [GROVER]. In order to use the extension proposed in this document, the key lengths of these transforms MUST be at least 256 bits long in order to provide sufficient resistance to quantum attacks. Accordingly the post-quantum security level achieved is at least 128 bits.¶
SKEYSEED is calculated from shared SK(x) using an algorithm defined in Transform Type 2. While a quantum attacker may learn the value of SK(x), if this value is obtained by means of a classical key exchange, other SK(x) values generated by means of a quantum-resistant algorithm ensure that the final SKEYSEED is not compromised. This assumes that the algorithm defined in the Transform Type 2 is post-quantum.¶
The main focus of this document is to prevent a passive attacker performing a "harvest and decrypt" attack. In other words, an attacker that records messages exchanged today and proceeds to decrypt them once he owns a quantum computer. This attack is prevented due to the hybrid nature of the key exchange. Other attacks involving an active attacker using a quantum-computer are not completely solved by this document. This is for two reasons.¶
The first reason is because the authentication step remains classical. In particular, the authenticity of the SAs established under IKEv2 is protected using a pre-shared key, RSA, DSA, or ECDSA algorithms. Whilst the pre-shared key option, provided the key is long enough, is post-quantum, the other algorithms are not. Moreover, in implementations where scalability is a requirement, the pre-shared key method may not be suitable. Quantum-safe authenticity may be provided by using a quantum-safe digital signature and several quantum-safe digital signature methods are being explored by IETF. For example, if the implementation is able to reliably track state, the hash based method, XMSS has the status of an RFC, see [RFC8391]. Currently, quantum-safe authentication methods are not specified in this document, but are planned to be incorporated in due course.¶
It should be noted that the purpose of post-quantum algorithms is to provide resistance to attacks mounted in the future. The current threat is that encrypted sessions are subject to eavesdropping and archived with decryption by quantum computers taking place at some point in the future. Until quantum computers become available there is no point in attacking the authenticity of a connection because there are no possibilities for exploitation. These only occur at the time of the connection, for example by mounting a man-in-the-middle (MitM) attack. Consequently there is not such a pressing need for quantum-safe authenticity.¶
Performing multiple key exchanges while establishing IKEv2 SA increases the responder's susceptibility to DoS attacks, because of an increased amount of resources needed to spend before the initiator is authenticated. This is especially true for post-quantum key exchange methods, where many of them are more memory and/or CPU intensive than the classical counterparts.¶
Responders may consider recommendations from [RFC8019] to deal with increased DoS attack susceptibility. It is also possible that the responder only agrees to create initial IKE SA without performing additional key exchanges, provided the initiator includes such an option in its proposals. Then peers immediately rekey initial IKE SA with the CREATE_CHILD_SA exchange and additional key exchanges performed via the IKE_FOLLOWUP_KE exchanges. In this case at the point when resource-intensive operations are required, peers have already authenticated each other. However, in the context of hybrid post-quantum key exchange this scenario would leave initial IKE SA (and initial Child SA if it is created) unprotected against quantum computers. Nevertheless the rekeyed IKE SA (and Child SAs that will be created over it) will have full protection. This is similar to the scenario described in [RFC8784]. Depending on peers' policy, this scenario may or may not be appropriate.¶
The authors would like to thank Frederic Detienne and Olivier Pelerin for their comments and suggestions, including the idea to negotiate the post-quantum algorithms using the existing KE payload. The authors are also grateful to Tobias Heider and Tobias Guggemos for valuable comments. Thanks to Paul Wouters for reviewing the document.¶
This appendix shows some examples of multiple key exchanges. These examples are purely for information purposes and they describe some message flow scenarios that may occur in establishing an IKE or CHILD SA. Note that some payloads that are not relevant to multiple key exchanges may be omitted for brevity.¶
The initiator proposes two sets of optional additional key exchanges, but the responder does not support any of them. The responder chooses NONE for each set and consequently, IKE_INTERMEDIATE exchange does not takes place and the exchange proceeds to IKE_AUTH phase. The resulting keying materials are the same as those derived with [RFC7296].¶
Initiator Responder ------------------------------------------------------------------------ HDR(IKE_SA_INIT), SAi1(.. AKE*...), ---> KEi1, Ni, N(IKEV2_FRAG_SUPPORTED), N(INTERMEDIATE_EXCHANGE_SUPPORTED) Proposal #1 Transform ECR (ID = ENCR_AES_GCM_16, 256-bit key) Transform PRF (ID = PRF_HMAC_SHA2_512) Transform KE (ID = Curve25519) Transform AKE1 (ID = PQ_KEM_1) Transform AKE1 (ID = PQ_KEM_2) Transform AKE1 (ID = NONE) Transform AKE2 (ID = PQ_KEM_3) Transform AKE2 (ID = PQ_KEM_4) Transform AKE2 (ID = NONE) <--- HDR(IKE_SA_INIT), SAr1(.. AKE*...), KEr1, Nr, N(IKEV2_FRAG_SUPPORTED), N(INTERMEDIATE_EXCHANGE_SUPPORTED) Proposal #1 Transform ECR (ID = ENCR_AES_GCM_16, 256-bit key) Transform PRF (ID = PRF_HMAC_SHA2_512) Transform KE (ID = Curve25519) Transform AKE1 (ID = NONE) Transform AKE2 (ID = NONE) HDR(IKE_AUTH), SK{ IDi, AUTH, SAi2, TSi, TSr } ---> <--- HDR(IKE_AUTH), SK{ IDr, AUTH, SAr2, TSi, TSr }¶
The exchanges below show that the initiator does not propose the use of additional key exchanges to establish an IKE SA, but they are required in order to establish a Child SA. In order to establish a fully quantum-resistant IPsec SA, both peers include CHILDLESS_IKEV2_SUPPORTED notification in their exchange so that the first Child SA is not created in IKE_AUTH, but instead the IKE SA is immediately rekeyed using CREATED_CHILD_SA. Any Child SA will have to be created via subsequent CREATED_CHILD_SA exchange.¶
Initiator Responder ------------------------------------------------------------------------ HDR(IKE_SA_INIT), SAi1, ---> KEi1, Ni, N(IKEV2_FRAG_SUPPORTED), N(CHILDLESS_IKEV2_SUPPORTED) <--- HDR(IKE_SA_INIT), SAr1, KEr1, Nr, N(IKEV2_FRAG_SUPPORTED), N(CHILDLESS_IKEV2_SUPPORTED) HDR(IKE_AUTH), SK{ IDi, AUTH } ---> <--- HDR(IKE_AUTH), SK{ IDr, AUTH } HDR(CREATE_CHILD_SA), SK{ SAi(.. AKE*...), Ni, KEi } ---> Proposal #1 Transform ECR (ID = ENCR_AES_GCM_16, 256-bit key) Transform PRF (ID = PRF_HMAC_SHA2_512) Transform KE (ID = Curve25519) Transform AKE1 (ID = PQ_KEM_1) Transform AKE1 (ID = PQ_KEM_2) Transform AKE2 (ID = PQ_KEM_5) Transform AKE2 (ID = PQ_KEM_6) Transform AKE2 (ID = NONE) <--- HDR(CREATE_CHILD_SA), SK{ SAr(.. AKE*...), Nr, KEr, N(ADDITIONAL_KEY_EXCHANGE)(link1) } Proposal #1 Transform ECR (ID = ENCR_AES_GCM_16, 256-bit key) Transform PRF (ID = PRF_HMAC_SHA2_512) Transform KE (ID = Curve25519) Transform AKE1 (ID = PQ_KEM_2) Transform AKE2 (ID = PQ_KEM_5) HDR(IKE_FOLLOWUP_KE), SK{ KEi(1), ---> N(ADDITIONAL_KEY_EXCHANGE)(link1) } <--- HDR(IKE_FOLLOWUP_KE), SK{ KEr(1), N(ADDITIONAL_KEY_EXCHANGE)(link2) } HDR(IKE_FOLLOWUP_KE), SK{ KEi(2), ---> N(ADDITIONAL_KEY_EXCHANGE)(link2) } <--- HDR(IKE_FOLLOWUP_KE), SK{ KEr(2) }¶
The initiator proposes the combination of PQ_KEM_1, PQ_KEM_2, PQ_KEM_3, and PQ_KEM_4 as the additional key exchanges. The initiator indicates, using the key exchange method NONE, that either PQ_KEM_1 or PQ_KEM_2 must be used to establish a security association. The responder, although supports the optional PQ_KEM_3 and PQ_KEM_4 method, does not support either PQ_KEM_1 or PQ_KEM_2 mandatory method and therefore responds with NO_PROPOSAL_CHOSEN notification.¶
Initiator Responder ------------------------------------------------------------------------ HDR(IKE_SA_INIT), SAi1(.. AKE*...), ---> KEi1, Ni, N(IKEV2_FRAG_SUPPORTED), N(INTERMEDIATE_EXCHANGE_SUPPORTED) Proposal #1 Transform ECR (ID = ENCR_AES_GCM_16, 256-bit key) Transform PRF (ID = PRF_HMAC_SHA2_512) Transform KE (ID = Curve25519) Transform AKE1 (ID = PQ_KEM_1) Transform AKE1 (ID = PQ_KEM_2) Transform AKE2 (ID = PQ_KEM_3) Transform AKE2 (ID = PQ_KEM_4) Transform AKE2 (ID = NONE) <--- HDR(IKE_SA_INIT), N(NO_PROPOSAL_CHOSEN)¶
This section gives an overview on a number of alternative approaches that we have considered, but later discarded. These approaches are:¶
Sending the classical and post-quantum key exchanges as a single transform¶
We considered combining the various key exchanges into a single large KE payload; this effort is documented in a previous version of this draft (draft-tjhai-ipsecme-hybrid-qske-ikev2-01). This does allow us to cleanly apply hybrid key exchanges during the child SA; however it does add considerable complexity, and requires an independent fragmentation solution.¶
Sending post-quantum proposals and policies in KE payload only¶
With the objective of not introducing unnecessary notify payloads, we considered communicating the hybrid post-quantum proposal in the KE payload during the first pass of the protocol exchange. Unfortunately, this design is susceptible to the following downgrade attack. Consider the scenario where there is an MitM attacker sitting between an initiator and a responder. The initiator proposes, through SAi payload, to use a hybrid post-quantum group and as a backup a Diffie-Hellman group, and through KEi payload, the initiator proposes a list of hybrid post-quantum proposals and policies. The MitM attacker intercepts this traffic and replies with N(INVALID_KE_PAYLOAD) suggesting to downgrade to the backup Diffie-Hellman group instead. The initiator then resends the same SAi payload and the KEi payload containing the public value of the backup Diffie-Hellman group. Note that the attacker may forward the second IKE_SA_INIT message only to the responder, and therefore at this point in time, the responder will not have the information that the initiator prefers the hybrid group. Of course, it is possible for the responder to have a policy to reject an IKE_SA_INIT message that (a) offers a hybrid group but not offering the corresponding public value in the KEi payload; and (b) the responder has not specifically acknowledged that it does not supported the requested hybrid group. However, the checking of this policy introduces unnecessary protocol complexity. Therefore, in order to fully prevent any downgrade attacks, using KE payload alone is not sufficient and that the initiator MUST always indicate its preferred post-quantum proposals and policies in a notify payload in the subsequent IKE_SA_INIT messages following a N(INVALID_KE_PAYLOAD) response.¶
New payload types to negotiate hybrid proposal and to carry post- quantum public values¶
Semantically, it makes sense to use a new payload type, which mimics the SA payload, to carry a hybrid proposal. Likewise, another new payload type that mimics the KE payload, could be used to transport hybrid public value. Although, in theory a new payload type could be made backwards compatible by not setting its critical flag as per Section 2.5 of RFC7296, we believe that it may not be that simple in practice. Since the original release of IKEv2 in RFC4306, no new payload type has ever been proposed and therefore, this creates a potential risk of having a backward compatibility issue from non-conforming RFC IKEv2 implementations. Since we could not see any other compelling advantages apart from a semantic one, we use the existing transform type and notify payloads instead. In fact, as described above, we use the KE payload in the first IKE_SA_INIT request round and the notify payload to carry the post-quantum proposals and policies. We use one or more of the existing KE payloads to carry the hybrid public values.¶
Hybrid public value payload¶
One way to transport the negotiated hybrid public payload, which contains one classical Diffie-Hellman public value and one or more post-quantum public values, is to bundle these into a single KE payload. Alternatively, these could also be transported in a single new hybrid public value payload, but following the same reasoning as above, this may not be a good idea from a backward compatibility perspective. Using a single KE payload would require an encoding or formatting to be defined so that both peers are able to compose and extract the individual public values. However, we believe that it is cleaner to send the hybrid public values in multiple KE payloads--one for each group or algorithm. Furthermore, at this point in the protocol exchange, both peers should have indicated support of handling multiple KE payloads.¶
Fragmentation¶
Handling of large IKE_SA_INIT messages has been one of the most challenging tasks. A number of approaches have been considered and the two prominent ones that we have discarded are outlined as follows.¶
The first approach was to treat the entire IKE_SA_INIT message as a stream of bytes, which we then split it into a number of fragments, each of which is wrapped onto a payload that would fit into the size of the network MTU. The payload that wraps each fragment is a new payload type and it was envisaged that this new payload type will not cause a backward compatibility issue because at this stage of the protocol, both peers should have indicated support of fragmentation in the first pass of the IKE_SA_INIT exchange. The negotiation of fragmentation is performed using a notify payload, which also defines supporting parameters such as the size of fragment in octets and the fragment identifier. The new payload that wraps each fragment of the messages in this exchange is assigned the same fragment identifier. Furthermore, it also has other parameters such as a fragment index and total number of fragments. We decided to discard this approach due to its blanket approach to fragmentation. In cases where only a few payloads need to be fragmented, we felt that this approach is overly complicated.¶
Another idea that was discarded was fragmenting an individual payload without introducing a new payload type. The idea was to use the 9-th bit (the bit after the critical flag in the RESERVED field) in the generic payload header as a flag to mark that this payload is fragmented. As an example, if a KE payload is to be fragmented, it may look as follows.¶
1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Payload |C|F| RESERVED | Payload Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Diffie-Hellman Group Number | Fragment Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Fragment Index | Total Fragments | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Total KE Payload Data Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | ~ Fragmented KE Payload ~ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+¶
Group sub-identifier¶
As discussed before, each group identifier is used to distinguish a post-quantum algorithm. Further classification could be made on a particular post-quantum algorithm by assigning additional value alongside the group identifier. This sub- identifier value may be used to assign different security parameter sets to a given post-quantum algorithm. However, this level of details does not fit the principles of the document where it should deal with generic hybrid key exchange protocol, not a specific ciphersuite. Furthermore, there are enough Diffie- Hellman group identifiers should this be required in the future.¶