IKEv2 Resumption: Internet Key Exchange Protocol Version 2 (IKEv2) Session Resumption
Letting a VPN client pick up a dropped session where it left off, without starting the whole handshake again.
Sheffer, Y. and H. Tschofenig, "Internet Key Exchange Protocol Version 2 (IKEv2) Session Resumption", RFC 5723, DOI 10.17487/RFC5723, January 2010, <https://www.rfc-editor.org/info/rfc5723>.
@misc{rfc5723,
series = {Request for Comments},
number = 5723,
howpublished = {RFC 5723},
publisher = {RFC Editor},
doi = {10.17487/RFC5723},
url = {https://www.rfc-editor.org/info/rfc5723},
author = {Y. Sheffer and H. Tschofenig},
title = {{Internet Key Exchange Protocol Version 2 (IKEv2) Session Resumption}},
pagetotal = 26,
year = 2010,
month = jan,
} RFC text
Contents
Internet Engineering Task Force (IETF) Y. Sheffer
Request for Comments: 5723 Check Point
Category: Standards Track H. Tschofenig
ISSN: 2070-1721 Nokia Siemens Networks
January 2010
Internet Key Exchange Protocol Version 2 (IKEv2)
Session Resumption
Abstract
The Internet Key Exchange version 2 (IKEv2) protocol has a certain
computational and communication overhead with respect to the number
of round trips required and the cryptographic operations involved.
In remote access situations, the Extensible Authentication Protocol
(EAP) is used for authentication, which adds several more round trips
and consequently latency.
To re-establish security associations (SAs) upon a failure recovery
condition is time consuming especially when an IPsec peer (such as a
VPN gateway) needs to re-establish a large number of SAs with various
endpoints. A high number of concurrent sessions might cause
additional problems for an IPsec peer during SA re-establishment.
In order to avoid the need to re-run the key exchange protocol from
scratch, it would be useful to provide an efficient way to resume an
IKE/IPsec session. This document proposes an extension to IKEv2 that
allows a client to re-establish an IKE SA with a gateway in a highly
efficient manner, utilizing a previously established IKE SA.
A client can reconnect to a gateway from which it was disconnected.
The proposed approach encodes partial IKE state into an opaque
ticket, which can be stored on the client or in a centralized store,
and is later made available to the IKEv2 responder for re-
authentication. We use the term ticket to refer to the opaque data
that is created by the IKEv2 responder. This document does not
specify the format of the ticket but examples are provided.
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Further information on
Internet Standards is available in Section 2 of RFC 5741.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
http://www.rfc-editor.org/info/rfc5723.
Copyright Notice
Copyright (c) 2010 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
(http://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.
This document may contain material from IETF Documents or IETF
Contributions published or made publicly available before November
10, 2008. The person(s) controlling the copyright in some of this
material may not have granted the IETF Trust the right to allow
modifications of such material outside the IETF Standards Process.
Without obtaining an adequate license from the person(s) controlling
the copyright in such materials, this document may not be modified
outside the IETF Standards Process, and derivative works of it may
not be created outside the IETF Standards Process, except to format
it for publication as an RFC or to translate it into languages other
than English.
Table of Contents
1. Introduction ....................................................4
2. Terminology .....................................................5
3. Usage Scenario ..................................................5
4. Protocol Sequences ..............................................7
4.1. Requesting a Ticket ........................................7
4.2. Receiving a Ticket .........................................8
4.3. Presenting a Ticket ........................................9
4.3.1. Prologue ............................................9
4.3.2. IKE_SESSION_RESUME Exchange ........................10
4.3.3. IKE_AUTH Exchange ..................................11
4.3.4. Epilogue ...........................................12
5. IKE and IPsec State after Resumption ...........................12
5.1. Generating Cryptographic Material for the Resumed IKE SA ..15
6. Ticket Handling ................................................16
6.1. Ticket Content ............................................16
6.2. Ticket Identity and Lifecycle .............................16
7. IKE Notifications ..............................................17
7.1. TICKET_LT_OPAQUE Notify Payload ...........................17
7.2. TICKET_OPAQUE Notify Payload ..............................18
8. IANA Considerations ............................................18
9. Security Considerations ........................................19
9.1. Stolen Tickets ............................................19
9.2. Forged Tickets ............................................19
9.3. Denial-of-Service Attacks .................................20
9.4. Detecting the Need for Resumption .........................20
9.5. Key Management for "Tickets by Value" .....................20
9.6. Ticket Lifetime ...........................................21
9.7. Tickets and Identity ......................................21
9.8. Ticket Revocation .........................................21
9.9. Ticket by Value Format ....................................21
9.10. Identity Privacy, Anonymity, and Unlinkability ...........22
10. Acknowledgements ..............................................22
11. References ....................................................23
11.1. Normative References .....................................23
11.2. Informative References ...................................23
Appendix A. Ticket Format ........................................25
A.1. Example "Ticket by Value" Format ..........................25
A.2. Example "Ticket by Reference" Format ......................25
1. Introduction
The Internet Key Exchange version 2 (IKEv2) protocol has a certain
computational and communication overhead with respect to the number
of round trips required and the cryptographic operations involved.
In particular, the Extensible Authentication Protocol (EAP) is used
for authentication in remote access cases, which increases latency.
To re-establish security associations (SAs) upon a failure recovery
condition is time-consuming, especially when an IPsec peer, such as a
VPN gateway, needs to re-establish a large number of SAs with various
endpoints. A high number of concurrent sessions might cause
additional problems for an IPsec responder. Usability is also
affected when the re-establishment of an IKE SA involves user
interaction for re-authentication.
In many failure cases, it would be useful to provide an efficient way
to resume an interrupted IKE/IPsec session. This document proposes
an extension to IKEv2 that allows a client to re-establish an IKE SA
with a gateway in a highly efficient manner, utilizing a previously
established IKE SA.
The client (IKEv2 initiator) stores the state about the previous IKE
SA locally. The gateway (IKEv2 responder) has two options for
maintaining the IKEv2 state about the previous IKE SA:
o In the "ticket by reference" approach, the gateway stores the
state locally, and gives the client a protected and opaque
reference (e.g., an index to the gateway's table) that the gateway
can later use to find the state. The client includes this opaque
reference when it resumes the session.
o In the "ticket by value" approach, the gateway stores its state in
a ticket (data structure) that is encrypted and integrity-
protected by a key known only to the gateway. The ticket is
passed to the client (who treats the ticket as an opaque string)
and sent back to the gateway when the session is resumed. The
gateway can then decrypt the ticket and recover the state.
Note that the client behaves identically in both cases, and in
general does not know which approach the gateway is using. Since the
ticket (or reference) is only interpreted by the same party that
created it, this document does not specify the exact format for it.
However, Appendix A contains examples for both "ticket by reference"
and "ticket by value" formats.
This approach is similar to the one taken by Transport Layer Security
(TLS) session resumption [RFC5077] with the required adaptations for
IKEv2, e.g., to accommodate the two-phase protocol structure. We
have borrowed heavily from that specification.
The proposed solution should additionally meet the following goals:
o Using only symmetric cryptography to minimize CPU consumption.
o Providing cryptographic agility.
o Having no negative impact on IKEv2 security features.
The following are non-goals of this solution:
o Failover from one gateway to another. This use case may be added
in a future specification.
o Providing load balancing among gateways.
o Specifying how a client detects the need for resumption.
2. Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119].
This document uses terminology defined in [RFC4301] and [RFC4306].
In addition, this document uses the following term:
Ticket: An IKEv2 ticket is a data structure that contains all the
necessary information that allows an IKEv2 responder to re-
establish an IKEv2 security association.
In this document, we use the term "ticket" and thereby refer to an
opaque data structure that may either contain IKEv2 state as
described above or a reference pointing to such state.
3. Usage Scenario
This specification envisions two usage scenarios for efficient IKEv2
and IPsec SA session re-establishment.
The first is similar to the use case specified in Section 1.1.3 of
the IKEv2 specification [RFC4306], where the IPsec tunnel mode is
used to establish a secure channel between a remote access client and
a gateway; the traffic flow may be between the client and entities
beyond the gateway. This scenario is further discussed below.
The second use case focuses on the usage of transport (or tunnel)
mode to secure the communicate between two endpoints (e.g., two
servers). The two endpoints have a client-server relationship with
respect to a protocol that runs using the protections afforded by the
IPsec SA.
(a)
+-+-+-+-+-+ +-+-+-+-+-+
| | IKEv2/IKEv2-EAP | | Protected
| Remote |<------------------------>| | Subnet
| Access | | Access |<--- and/or
| Client |<------------------------>| Gateway | Internet
| | IPsec tunnel | |
+-+-+-+-+-+ +-+-+-+-+-+
(b)
+-+-+-+-+-+ +-+-+-+-+-+
| | IKE_SESSION_RESUME | |
| Remote |<------------------------>| |
| Access | | Access |
| Client |<------------------------>| Gateway |
| | IPsec tunnel | |
+-+-+-+-+-+ +-+-+-+-+-+
Figure 1: Resuming a Session with a Remote Access Gateway
In the first use case above, an end host (an entity with a host
implementation of IPsec [RFC4301]) establishes a tunnel mode IPsec SA
with a gateway in a remote network using IKEv2. The end host in this
scenario is sometimes referred to as a remote access client. At a
later stage, when a client needs to re-establish the IKEv2 session,
it may choose to establish IPsec SAs using a full IKEv2 exchange or
the IKE_SESSION_RESUME exchange (shown in Figure 1).
For either of the above use cases, there are multiple possible
situations where the mechanism specified in this document could be
useful. These include the following (note that this list is not
meant to be exhaustive, and any particular deployment may not care
about all of these):
o If a client temporarily loses network connectivity (and the IKE SA
times out through the liveness test facility, a.k.a. "dead peer
detection"), this mechanism could be used to re-establish the SA
with less overhead (network, CPU, authentication infrastructure)
and without requiring user interaction for authentication.
o If the connectivity problems affect a large number of clients
(e.g., a large remote access VPN gateway), when the connectivity
is restored, all the clients might reconnect almost
simultaneously. This mechanism could be used to reduce the load
spike for cryptographic operations and authentication
infrastructure.
o Losing connectivity can also be predictable and planned; for
example, putting a laptop to "stand-by" mode before traveling.
This mechanism could be used to re-establish the SA when the
laptop is switched back on (again, with less overhead and without
requiring user interaction for authentication). However, such
user-level "resumption" may often be disallowed by policy.
Moreover, this document requires the client to destroy the ticket
when the user explicitly "logs out" (Section 6.2).
4. Protocol Sequences
This section provides protocol details and contains the normative
parts. This document defines two protocol exchanges, namely
requesting a ticket, see Section 4.1, and presenting a ticket, see
Section 4.3.
4.1. Requesting a Ticket
A client MAY request a ticket in the following exchanges:
o In an IKE_AUTH exchange, as shown in the example message exchange
in Figure 2 below.
o In a CREATE_CHILD_SA exchange, when an IKE SA is rekeyed (and only
when this exchange is initiated by the client).
o In an Informational exchange at any time, e.g., if the gateway
previously replied with an N(TICKET_ACK) instead of providing a
ticket, or when the ticket lifetime is about to expire, or
following a gateway-initiated IKE rekey. All such Informational
exchanges MUST be initiated by the client.
o While resuming an IKE session, i.e., in the IKE_AUTH exchange that
follows an IKE_SESSION_RESUME exchange, see Section 4.3.3.
Normally, a client requests a ticket in the third message of an IKEv2
exchange (the first of IKE_AUTH). Figure 2 shows the message
exchange for this typical case.
Initiator Responder
----------- -----------
HDR, SAi1, KEi, Ni -->
<-- HDR, SAr1, KEr, Nr [, CERTREQ]
HDR, SK {IDi, [CERT,] [CERTREQ,] [IDr,]
AUTH, SAi2, TSi, TSr, N(TICKET_REQUEST)} -->
Figure 2: Example Message Exchange for Requesting a Ticket
The notification payloads are described in Section 7. The above is
an example, and IKEv2 allows a number of variants on these messages.
Refer to [RFC4306] and [IKEV2-BIS] for more details on IKEv2.
When an IKEv2 responder receives a request for a ticket using the
N(TICKET_REQUEST) payload, it MUST perform one of the following
operations if it supports the extension defined in this document:
o it creates a ticket and returns it with the N(TICKET_LT_OPAQUE)
payload in a subsequent message towards the IKEv2 initiator. This
is shown in Figure 3.
o it returns an N(TICKET_NACK) payload, if it refuses to grant a
ticket for some reason.
o it returns an N(TICKET_ACK), if it cannot grant a ticket
immediately, e.g., due to packet size limitations. In this case,
the client MAY request a ticket later using an Informational
exchange, at any time during the lifetime of the IKE SA.
Regardless of this choice, there is no change to the behavior of the
responder with respect to the IKE exchange, and the proper IKE
response (e.g., an IKE_AUTH response or an error notification) MUST
be sent.
4.2. Receiving a Ticket
The IKEv2 initiator receives the ticket and may accept it, provided
the IKEv2 exchange was successful. The ticket may be used later with
an IKEv2 responder that supports this extension. Figure 3 shows how
the initiator receives the ticket.
Initiator Responder
----------- -----------
<-- HDR, SK {IDr, [CERT,] AUTH, SAr2, TSi,
TSr, N(TICKET_LT_OPAQUE) }
Figure 3: Receiving a Ticket
When a multi-round-trip IKE_AUTH exchange is used, the
N(TICKET_REQUEST) payload MUST be included in the first IKE_AUTH
request, and N(TICKET_LT_OPAQUE) (or TICKET_NACK/TICKET_ACK) MUST
only be returned in the final IKE_AUTH response.
When the client accepts the ticket, it stores it in its local storage
for later use, along with the IKE SA to which the ticket refers.
Since the ticket itself is opaque to the client, the local storage
MUST also include all items marked as "from the ticket" in the table
of Section 5.
4.3. Presenting a Ticket
When the client wishes to recover from an interrupted session, it
presents the ticket to resume the session. This section describes
the resumption process, consisting of some preparations, an
IKE_SESSION_RESUME exchange, an IKE_AUTH exchange and finalization.
4.3.1. Prologue
It is up to the client's local policy to decide when the
communication with the IKEv2 responder is seen as interrupted and the
session resumption procedure is to be initiated.
A client MAY initiate a regular (non-ticket-based) IKEv2 exchange
even if it is in possession of a valid, unexpired ticket. A client
MUST NOT present a ticket when it knows that the ticket's lifetime
has expired.
Tickets are intended for one-time use, i.e., a client MUST NOT reuse
a ticket. A reused ticket SHOULD be rejected by a gateway. Note
that a ticket is considered as used only when an IKE SA has been
established successfully with it.
4.3.2. IKE_SESSION_RESUME Exchange
This document specifies a new IKEv2 exchange type called
IKE_SESSION_RESUME whose value is 38. This exchange is equivalent to
the IKE_SA_INIT exchange, and MUST be followed by an IKE_AUTH
exchange. The client SHOULD NOT use this exchange type unless it
knows that the gateway supports it (this condition is trivially true
in the context of the current document, since the client always
resumes into the same gateway that generated the ticket).
Initiator Responder
----------- -----------
HDR, [N(COOKIE),] Ni, N(TICKET_OPAQUE) [,N+] -->
Figure 4: IKEv2 Initiator Wishes to Resume an IKE SA
The exchange type in HDR is set to 'IKE_SESSION_RESUME'. The
initiator sets the SPIi (Security Parameter Index, Initiator) value
in the HDR to a new, unique value and the SPIr value is set to 0.
When the IKEv2 responder receives a ticket using the N(TICKET_OPAQUE)
payload, it MUST perform one of the following steps if it supports
the extension defined in this document:
o If it is willing to accept the ticket, it responds as shown in
Figure 5.
o It responds with an unprotected N(TICKET_NACK) notification, if it
rejects the ticket for any reason. In that case, the initiator
should re-initiate a regular IKE exchange. One such case is when
the responder receives a ticket for an IKE SA that has previously
been terminated on the responder itself, which may indicate
inconsistent state between the IKEv2 initiator and the responder.
However, a responder is not required to maintain the state for
terminated sessions.
Initiator Responder
----------- -----------
<-- HDR, Nr [,N+]
Figure 5: IKEv2 Responder Accepts the Ticket
Again, the exchange type in HDR is set to 'IKE_SESSION_RESUME'. The
responder copies the SPIi value from the request, and the SPIr value
is set to a new, unique value.
Where not specified otherwise, the IKE_SESSION_RESUME exchange
behaves exactly like the IKE_SA_INIT exchange. Specifically:
o The client MAY resume the IKE exchange from any IP address and
port, regardless of its original address. The gateway MAY reject
the resumed exchange if its policy depends on the client's address
(although this rarely makes sense).
o The first message MAY be rejected in denial-of-service (DoS)
situations, with the initiator instructed to send a cookie.
o Notifications normally associated with IKE_SA_INIT can be sent.
In particular, NAT detection payloads.
o The client's NAT traversal status SHOULD be determined anew in
IKE_SESSION_RESUME. If NAT is detected, the initiator switches to
UDP encapsulation on port 4500, as per [RFC4306], Section 2.23.
NAT status is explicitly not part of the session resumption state.
o The SPI values and Message ID fields behave similarly to
IKE_SA_INIT.
Although the IKE SA is not fully valid until the completion of the
IKE_AUTH exchange, the peers must create much of the SA state
(Section 5) now. Specifically, the shared key values are required to
protect the IKE_AUTH payloads. Their generation is described in
Section 5.1.
4.3.3. IKE_AUTH Exchange
Following the IKE_SESSION_RESUME exchange, the client MUST initiate
an IKE_AUTH exchange, which is largely as specified in [RFC4306].
This section lists the differences and constraints compared to the
base document.
The value of the AUTH payload is derived in a manner similar to the
usage of IKEv2 pre-shared secret authentication:
AUTH = prf(SK_px, <message octets>)
Each of the initiator and responder uses its own value for SK_px,
namely SK_pi for the initiator and SK_pr for the responder. Both are
taken from the newly generated IKE SA (Section 5.1).
The exact material to be signed is defined in Section 2.15 of
[RFC4306].
The IDi value sent in the IKE_AUTH exchange MUST be identical to the
value included in the ticket. A CERT payload MUST NOT be included in
this exchange, and therefore a new IDr value cannot be negotiated
(since it would not be authenticated). As a result, the IDr value
sent (by the gateway, and optionally by the client) in this exchange
MUST also be identical to the value included in the ticket.
When resuming a session, a client will typically request a new ticket
immediately, so that it is able to resume the session again in the
case of a second failure. The N(TICKET_REQUEST) and
N(TICKET_LT_OPAQUE) notifications will be included in the IKE_AUTH
exchange that follows the IKE_SESSION_RESUME exchange, with similar
behavior to a ticket request during a regular IKE exchange,
Section 4.1. The returned ticket (if any) will correspond to the IKE
SA created per the rules described in Section 5.
4.3.4. Epilogue
Following the IKE_AUTH exchange, a new IKE SA is created by both
parties, see Section 5, and a Child SA is derived, per Section 2.17
of [RFC4306].
When the responder receives a ticket for an IKE SA that is still
active and if the responder accepts it (i.e., following successful
completion of the IKE_AUTH exchange), the old SA SHOULD be silently
deleted without sending a DELETE informational exchange.
Consequently, all the dependent IPsec Child SAs are also deleted.
5. IKE and IPsec State after Resumption
During the resumption process, both peers create IKE and IPsec state
for the resumed IKE SA. Although the SA is only completed following
a successful IKE_AUTH exchange, many of its components are created
earlier, notably the SA's crypto material (Section 5.1).
When a ticket is presented, the gateway needs to obtain the ticket
state. In case a "ticket by reference" was provided by the client,
the gateway needs to resolve the reference in order to obtain this
state. In case the client has already provided a "ticket by value",
the gateway can parse the ticket to obtain the state directly. In
either case, the gateway needs to process the ticket state in order
to restore the state of the old IKE SA, and the client retrieves the
same state from its local store.
The following table describes the IKE and IPsec state of the peers
after session resumption, and how it is related to their state before
the IKE SA was interrupted. When the table mentions that a certain
state item is taken "from the ticket", this should be construed as:
o The client retrieves this item from its local store.
o In the case of "ticket by value", the gateway encodes this
information in the ticket.
o In the case of "ticket by reference", the gateway fetches this
information from the ticket store.
+--------------------------------+----------------------------------+
| State Item | After Resumption |
+--------------------------------+----------------------------------+
| IDi | From the ticket (but must also |
| | be exchanged in IKE_AUTH). See |
| | also Note 1. |
| | |
| IDr | From the ticket (but must also |
| | be exchanged in IKE_AUTH). |
| | |
| Authentication method (PKI, | From the ticket. |
| pre-shared secret, EAP, | |
| PKI-less EAP [EAP-AUTH] etc.) | |
| | |
| Certificates (when applicable) | From the ticket, see Note 2. |
| | |
| Local IP address/port, peer IP | Selected by the client, see Note |
| address/port | 3. |
| | |
| NAT detection status | From new exchange. |
| | |
| SPIs | From new exchange, see Note 4. |
| | |
| Which peer is the "original | Determined by the initiator of |
| initiator"? | IKE_SESSION_RESUME. |
| | |
| IKE SA sequence numbers | Reset to 0 in |
| (Message ID) | IKE_SESSION_RESUME, and |
| | subsequently incremented |
| | normally. |
| | |
| IKE SA algorithms (SAr) | From the ticket. |
| | |
| IKE SA keys (SK_*) | The old SK_d is obtained from |
| | the ticket and all keys are |
| | refreshed, see Section 5.1. |
| | |
| IKE SA window size | Reset to 1. |
| | |
| Child SAs (ESP/AH) | Created in new exchange, see |
| | Note 6. |
| | |
| Internal IP address | Not resumed, but see Note 5. |
| | |
| Other Configuration Payload | Not resumed. |
| information | |
| | |
| Peer Vendor IDs | Not resumed, resent in new |
| | exchange if required. |
| | |
| Peer supports MOBIKE [RFC4555] | From new exchange. |
| | |
| MOBIKE additional addresses | Not resumed, should be resent by |
| | client if necessary. |
| | |
| Time until re-authentication | From new exchange (but ticket |
| [RFC4478] | lifetime is bounded by this |
| | duration). |
| | |
| Peer supports redirects | From new exchange. |
| [RFC5685] | |
+--------------------------------+----------------------------------+
Note 1: The authenticated peer identity used for policy lookups may
not be the same as the IDi payload. This is possible when
using certain EAP methods, see Section 3.5 of [RFC4718]. If
these identities are indeed different, then the
authenticated client identity MUST be included in the
ticket. Note that the client may not have access to this
value.
Note 2: Certificates don't need to be stored if the peer never uses
them for anything after the IKE SA is up; however, if they
are needed, e.g., if exposed to applications via IPsec APIs,
they MUST be stored in the ticket.
Note 3: If the certificate has an iPAddress SubjectAltName, and the
implementation requires it to match the peer's source IP
address, the same check needs to be performed on session
resumption and the required information saved locally or in
the ticket.
Note 4: SPI values of the old SA MAY be stored in the ticket, to
help the gateway locate corresponding old IKE state. These
values MUST NOT be used for the resumed SA.
Note 5: The client can request the address it was using earlier, and
if possible, the gateway SHOULD honor the request.
Note 6: Since information about Child SAs and configuration payloads
is not resumed, IKEv2 features related to Child SA
negotiation (such as IPCOMP_SUPPORTED,
ESP_TFC_PADDING_NOT_SUPPORTED, ROHC-over-IPsec [ROHCoIPsec]
and configuration) aren't usually affected by session
resumption.
IKEv2 features that affect only the IKE_AUTH exchange (including
HTTP_CERT_LOOKUP_SUPPORTED, multiple authentication exchanges
[RFC4739], Elliptic Curve Digital Signature Algorithm (ECDSA)
authentication [RFC4754], and the Online Certificate Status Protocol
(OCSP) [RFC4806]) don't usually need any state in the IKE SA (after
the IKE_AUTH exchanges are done), so resumption doesn't affect them.
New IKEv2 features that are not covered by Note 6 or by the previous
paragraph should specify how they interact with session resumption.
5.1. Generating Cryptographic Material for the Resumed IKE SA
The cryptographic material is refreshed based on the ticket and the
nonce values, Ni, and Nr, from the current exchange. A new SKEYSEED
value is derived as follows:
SKEYSEED = prf(SK_d_old, "Resumption" | Ni | Nr)
where SK_d_old is taken from the ticket. The literal string is
encoded as 10 ASCII characters, with no NULL terminator.
The keys are derived as follows, unchanged from IKEv2:
{SK_d | SK_ai | SK_ar | SK_ei | SK_er | SK_pi | SK_pr} =
prf+(SKEYSEED, Ni | Nr | SPIi | SPIr)
where SPIi, SPIr are the SPI values created in the new IKE exchange.
See [RFC4306] for the notation. "prf" is determined from the SA value
in the ticket.
6. Ticket Handling
6.1. Ticket Content
When passing a "ticket by value" to the client, the ticket content
MUST be integrity protected and encrypted.
A "ticket by reference" does not need to be encrypted, as it does not
contain any sensitive material, such as keying material. However,
access to the storage where that sensitive material is stored MUST be
protected so that only authorized access is allowed. We note that
such a ticket is analogous to the concept of 'stub', as defined in
[SA-SYNC], or the concept of a Session ID from TLS.
Although not strictly required for cryptographic protection, it is
RECOMMENDED to integrity-protect the "ticket by reference". Failing
to do so could result in various security vulnerabilities on the
gateway side, depending on the format of the reference. Potential
vulnerabilities include access by the gateway to unintended URLs
(similar to cross-site scripting) or SQL injection.
When the state is passed by value, the ticket MUST encode all state
information marked "from the ticket" in the table on Section 5. The
same state MUST be stored in the ticket store, in the case of "ticket
by reference".
A "ticket by value" MUST include a protected expiration time, which
is an absolute time value and SHOULD correspond to the value included
in the TICKET_LT_OPAQUE payload.
The "ticket by value" MUST additionally include a key identity field,
so that keys for ticket encryption and authentication can be changed,
and when necessary, algorithms can be replaced.
6.2. Ticket Identity and Lifecycle
Each ticket is associated with a single IKE SA. In particular, when
an IKE SA is deleted by the client or the gateway, the client MUST
delete its stored ticket. Similarly, when credentials associated
with the IKE SA are invalidated (e.g., when a user logs out), the
ticket MUST be deleted. When the IKE SA is rekeyed, the ticket is
invalidated, and the client SHOULD request a new ticket. When a
client does not follow these rules, it might present an invalid
ticket to the gateway. See Section 9.8 for more about this issue.
The lifetime of the ticket sent by the gateway SHOULD be the minimum
of the IKE SA lifetime (per the gateway's local policy) and its re-
authentication time, according to [RFC4478]. Even if neither of
these are enforced by the gateway, a finite lifetime MUST be
specified for the ticket.
The key that is used to protect the ticket MUST have a lifetime that
is significantly longer than the lifetime of an IKE SA.
In normal operation, the client will request a ticket when
establishing the initial IKE SA, and then every time the SA is
rekeyed or re-established because of re-authentication.
7. IKE Notifications
This document defines a number of notifications. The following
Notify Message types have been assigned by IANA.
+-------------------+-------+-----------------+
| Notification Name | Value | Data |
+-------------------+-------+-----------------+
| TICKET_LT_OPAQUE | 16409 | See Section 7.1 |
| | | |
| TICKET_REQUEST | 16410 | None |
| | | |
| TICKET_ACK | 16411 | None |
| | | |
| TICKET_NACK | 16412 | None |
| | | |
| TICKET_OPAQUE | 16413 | See Section 7.2 |
+-------------------+-------+-----------------+
For all these notifications, the Protocol ID and the SPI Size fields
MUST both be sent as 0.
7.1. TICKET_LT_OPAQUE Notify Payload
The data for the TICKET_LT_OPAQUE Notify payload consists of the
Notify message header, a Lifetime field and the ticket itself. The
four octet Lifetime field contains a relative time value, the number
of seconds until the ticket expires (encoded as an unsigned integer,
in network byte order).
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Next Payload |C| Reserved | Payload Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Protocol ID | SPI Size = 0 | Notify Message Type |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Lifetime |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
~ Ticket ~
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 6: TICKET_LT_OPAQUE Notify Payload
7.2. TICKET_OPAQUE Notify Payload
The data for the TICKET_OPAQUE Notify payload consists of the Notify
message header, and the ticket itself. Unlike the TICKET_LT_OPAQUE
payload, no lifetime value is included in the TICKET_OPAQUE Notify
payload.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Next Payload |C| Reserved | Payload Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Protocol ID | SPI Size = 0 | Notify Message Type |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| |
~ Ticket ~
| |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 7: TICKET_OPAQUE Notify Payload
8. IANA Considerations
Section 4.3.2 defines a new IKEv2 exchange type, IKE_SESSION_RESUME,
whose value has been allocated from the "IKEv2 Exchange Types"
registry.
Section 7 defines several new IKEv2 notifications whose Message Type
values have been allocated from the "IKEv2 Notify Message Types -
Status Types" registry.
9. Security Considerations
This section addresses security issues related to the usage of a
ticket.
9.1. Stolen Tickets
A man in the middle may try to eavesdrop on an exchange to obtain a
"ticket by value" and use it to establish a session with the IKEv2
responder. Since all exchanges where the client obtains a ticket are
encrypted, this is only possible by listening in on a client's use of
the ticket to resume a session. However, since the ticket's contents
are encrypted and the attacker does not know the corresponding secret
key, a stolen ticket cannot be used by an attacker to successfully
resume a session. An IKEv2 responder MUST use strong encryption and
integrity protection of the ticket to prevent an attacker from
obtaining the ticket's contents, e.g., by using a brute force attack.
A "ticket by reference" does not need to be encrypted. When an
adversary is able to eavesdrop on a resumption attempt, as described
in the previous paragraph, then the "ticket by reference" may be
obtained. A "ticket by reference" cannot be used by an attacker to
successfully resume a session, for the same reasons as for a "ticket
by value", namely because the attacker would not be able to prove,
during IKE_AUTH, its knowledge of the secret part of the IKE state
embedded in the ticket. Moreover, the adversary MUST NOT be able to
resolve the ticket via the reference, i.e., access control MUST be
enforced to ensure disclosure only to authorized entities.
9.2. Forged Tickets
A malicious user could forge or alter a "ticket by value" in order to
resume a session, to extend its lifetime, to impersonate as another
user, or to gain additional privileges. This attack is not possible
if the content of the "ticket by value" is protected using a strong
integrity protection algorithm.
In the case of a "ticket by reference" an adversary may attempt to
construct a fake "ticket by reference" to point to state information
stored by the IKEv2 responder. This attack will fail because the
adversary is not in possession of the keying material associated with
the IKEv2 SA. As noted in Section 6.1, it is often useful to
integrity-protect the "ticket by reference", too.
9.3. Denial-of-Service Attacks
An adversary could generate and send a large number of "tickets by
value" to a gateway for verification. Such an attack could burden
the gateway's CPU, and/or exhaust its memory with half-open IKE
state. To minimize the possibility of such denial of service, ticket
verification should be lightweight (e.g., using efficient symmetric
key cryptographic algorithms).
When an adversary chooses to send a large number of "tickets by
reference" then this may lead to an amplification attack as the IKEv2
responder is forced to resolve the reference to a ticket in order to
determine that the adversary is not in possession of the keying
material corresponding to the stored state or that the reference is
void. To minimize this attack, the protocol to resolve the reference
should be as lightweight as possible and should not generate a large
number of messages.
Note also that the regular IKEv2 cookie mechanism can be used to
handle state-overflow DoS situations.
9.4. Detecting the Need for Resumption
Detecting when an old IKE SA is no longer usable and needs to be
resumed is out of scope of the current document. However, clients
are warned against implementing a more liberal policy than that used
to detect failed IKE SAs (Section 2.4 of RFC 4306). In particular,
untrusted messages MUST NOT be relied upon to make this decision.
9.5. Key Management for "Tickets by Value"
A full description of the management of the keys used to protect the
"ticket by value" is beyond the scope of this document. A list of
RECOMMENDED practices is given below.
o The keys should be generated securely following the randomness
recommendations in [RFC4086].
o The keys and cryptographic protection algorithms should be at
least 128 bits in strength.
o The keys should not be used for any other purpose than generating
and verifying tickets.
o The keys should be changed regularly.
o The keys should be changed if the ticket format or cryptographic
protection algorithms change.
9.6. Ticket Lifetime
An IKEv2 responder controls the validity period of the state
information by attaching a lifetime to a ticket. The chosen lifetime
is based on the operational and security requirements of the
environment in which this IKEv2 extension is deployed. The responder
provides information about the ticket lifetime to the IKEv2
initiator, allowing it to manage its tickets.
9.7. Tickets and Identity
A ticket is associated with a certain identity, and MUST be managed
securely on the client side. Section 6.2 requires that a ticket be
deleted when the credentials associated with the ticket's identity
are no longer valid, e.g., when a user whose credentials were used to
create the SA logs out.
9.8. Ticket Revocation
A misbehaving client could present a ticket in its possession to the
gateway resulting in session resumption, even though the IKE SA
associated with this ticket had previously been deleted. This is
disallowed by Section 6.2. This issue is unique to "ticket by value"
cases, since a "ticket by reference" will have been deleted from the
ticket store.
To avoid this issue for "ticket by value", an Invalid Ticket List
(ITL) may be maintained by the gateway, see [TOKENS]. This can be a
simple blacklist of revoked tickets. Alternatively, [TOKENS]
suggests to use Bloom Filters [Bloom70] to maintain the list in
constant space. Management of such lists is outside the scope of the
current document. Note that a policy that requires tickets to have
shorter lifetimes (e.g., 1 hour) significantly mitigates this issue.
9.9. Ticket by Value Format
The ticket's format is not defined by this document, since this is
not required for interoperability. However, great care must be taken
when defining a ticket format such that the requirements outlined in
Section 6.1 are met. The "ticket by value" MUST have its integrity
and confidentiality protected with strong cryptographic techniques to
prevent a breach in the security of the system.
9.10. Identity Privacy, Anonymity, and Unlinkability
Since opaque state information is passed around between the IKEv2
initiator and the IKEv2 responder it is important that leakage of
information, such as the identities of an IKEv2 initiator and a
responder, MUST be avoided.
When an IKEv2 initiator presents a ticket as part of the
IKE_SESSION_RESUME exchange, confidentiality is not provided for the
exchange. There is thereby the possibility for an on-path adversary
to observe multiple exchange handshakes where the same state
information is used and therefore to conclude that they belong to the
same communication endpoints.
This document therefore requires that the ticket be presented to the
IKEv2 responder only once; under normal circumstances (e.g., no
active attacker), there should be no multiple use of the same ticket.
We are not aware of additional security issues associated with ticket
reuse: the protocol guarantees freshness of the generated crypto
material even in such cases. As noted in Section 4.3.1, the gateway
SHOULD prevent multiple uses of the same ticket. But this is only an
extra precaution, to ensure that clients do not implement reuse. In
other words, the gateway is not expected to cache old tickets for
extended periods of time.
10. Acknowledgements
We would like to thank Paul Hoffman, Pasi Eronen, Florian Tegeler,
Stephen Kent, Sean Shen, Xiaoming Fu, Stjepan Gros, Dan Harkins, Russ
Housely, Yoav Nir, Peny Yang, Sean Turner, and Tero Kivinen for their
comments. We would like to particularly thank Florian Tegeler and
Stjepan Gros for their implementation efforts and Florian Tegeler for
a formal verification using the Casper tool set.
We would furthermore like to thank the authors of [SA-SYNC] (Yan Xu,
Peny Yang, Yuanchen Ma, Hui Deng, and Ke Xu) for their input on the
stub concept.
We would like to thank Hui Deng, Tero Kivinen, Peny Yang, Ahmad
Muhanna, and Stephen Kent for their feedback regarding the "ticket by
reference" concept.
Vidya Narayanan and Lakshminath Dondeti coauthored several past
versions of this document, and we acknowledge their significant
contribution.
11. References
11.1. Normative References
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC4306] Kaufman, C., "Internet Key Exchange (IKEv2) Protocol",
RFC 4306, December 2005.
11.2. Informative References
[Bloom70] Bloom, B., "Space/time trade-offs in hash coding with
allowable errors", Comm. ACM 13(7):422-6, July 1970.
[EAP-AUTH] Eronen, P., Tschofenig, H., and Y. Sheffer, "An
Extension for EAP-Only Authentication in IKEv2", Work
in Progress, October 2009.
[IKEV2-BIS] Kaufman, C., Hoffman, P., Nir, Y., and P. Eronen,
"Internet Key Exchange Protocol: IKEv2", Work
in Progress, October 2009.
[RFC4086] Eastlake, D., Schiller, J., and S. Crocker, "Randomness
Requirements for Security", BCP 106, RFC 4086,
June 2005.
[RFC4301] Kent, S. and K. Seo, "Security Architecture for the
Internet Protocol", RFC 4301, December 2005.
[RFC4478] Nir, Y., "Repeated Authentication in Internet Key
Exchange (IKEv2) Protocol", RFC 4478, April 2006.
[RFC4555] Eronen, P., "IKEv2 Mobility and Multihoming Protocol
(MOBIKE)", RFC 4555, June 2006.
[RFC4718] Eronen, P. and P. Hoffman, "IKEv2 Clarifications and
Implementation Guidelines", RFC 4718, October 2006.
[RFC4739] Eronen, P. and J. Korhonen, "Multiple Authentication
Exchanges in the Internet Key Exchange (IKEv2)
Protocol", RFC 4739, November 2006.
[RFC4754] Fu, D. and J. Solinas, "IKE and IKEv2 Authentication
Using the Elliptic Curve Digital Signature Algorithm
(ECDSA)", RFC 4754, January 2007.
[RFC4806] Myers, M. and H. Tschofenig, "Online Certificate Status
Protocol (OCSP) Extensions to IKEv2", RFC 4806,
February 2007.
[RFC5077] Salowey, J., Zhou, H., Eronen, P., and H. Tschofenig,
"Transport Layer Security (TLS) Session Resumption
without Server-Side State", RFC 5077, January 2008.
[RFC5685] Devarapalli, V. and K. Weniger, "Redirect Mechanism for
the Internet Key Exchange Protocol Version 2 (IKEv2)",
RFC 5685, November 2009.
[ROHCoIPsec] Ertekin, E., Christou, C., Jasani, R., Kivinen, T., and
C. Bormann, "IKEv2 Extensions to Support Robust Header
Compression over IPsec (ROHCoIPsec)", Work in Progress,
December 2009.
[SA-SYNC] Xu, Y., Yang, P., Ma, Y., Deng, H., and H. Deng, "IKEv2
SA Synchronization for session resumption", Work
in Progress, October 2008.
[TOKENS] Rescorla, E., "How to Implement Secure (Mostly)
Stateless Tokens", Work in Progress, March 2007.
Appendix A. Ticket Format
This document does not specify a particular ticket format nor even
the suggested contents of a ticket: both are entirely up to the
implementer. The formats described in the following sub-sections are
provided as useful examples, and implementers are free to adopt them
as-is or change them in any way necessary.
A.1. Example "Ticket by Value" Format
struct {
[authenticated] struct {
octet format_version; // 1 for this version of the protocol
octet reserved[3]; // sent as 0, ignored by receiver.
octet key_id[8]; // arbitrary byte string
opaque IV[0..255]; // actual length (possibly 0) depends
// on the encryption algorithm
[encrypted] struct {
opaque IDi, IDr; // the full payloads
octet SPIi[8], SPIr[8];
opaque SA; // the full SAr payload
octet SK_d[0..255]; // actual length depends on SA value
enum ... authentication_method;
int32 expiration; // an absolute time value, seconds
// since Jan. 1, 1970
} ikev2_state;
} protected_part;
opaque MAC[0..255]; // the length (possibly 0) depends
// on the integrity algorithm
} ticket;
Note that the key defined by "key_id" determines the encryption and
authentication algorithms used for this ticket. Those algorithms are
unrelated to the transforms defined by the SA payload.
The reader is referred to [TOKENS] that recommends a similar (but not
identical) ticket format, and discusses related security
considerations in depth.
A.2. Example "Ticket by Reference" Format
For implementations that prefer to pass a reference to IKE state in
the ticket, rather than the state itself, we suggest the following
format:
struct {
[authenticated] struct {
octet format_version; // 1 for this version of the protocol
octet reserved[3]; // sent as 0, ignored by receiver.
octet key_id[8]; // arbitrary byte string
struct {
opaque state_ref; // reference to IKE state
int32 expiration; // an absolute time value, seconds
// since Jan. 1, 1970
} ikev2_state_ref;
} protected_part;
opaque MAC[0..255]; // the length depends
// on the integrity algorithm
} ticket;
Authors' Addresses
Yaron Sheffer
Check Point Software Technologies Ltd.
5 Hasolelim St.
Tel Aviv 67897
Israel
EMail: yaronf@checkpoint.com
Hannes Tschofenig
Nokia Siemens Networks
Linnoitustie 6
Espoo 02600
Finland
Phone: +358 (50) 4871445
EMail: Hannes.Tschofenig@gmx.net
URI: http://www.tschofenig.priv.at
ISO Plain Language
On this page
§1Why Resume Instead of Starting Over
The Internet Key Exchange version 2 protocol has a computational and communication overhead in terms of the number of round trips required and the cryptographic operations involved. In remote access situations the Extensible Authentication Protocol is used for authentication, which adds several more round trips and consequently latency.
Re-establishing security associations upon a failure recovery condition is time-consuming, especially when an IPsec peer such as a VPN gateway has to re-establish a large number of SAs with various endpoints. A high number of concurrent sessions might cause additional problems for an IPsec peer during SA re-establishment, and usability is affected when the re-establishment of an IKE SA involves user interaction for re-authentication.
This document proposes an extension to IKEv2 that allows a client to re-establish an IKE SA with a gateway in a highly efficient manner, using a previously established IKE SA. The client, the IKEv2 initiator, stores the state about the previous IKE SA locally. The gateway, the IKEv2 responder, has two options for maintaining the state.
In the ticket by reference approach the gateway stores the state locally and gives the client a protected and opaque reference, for example an index into the gateway's table, which the gateway can later use to find the state. In the ticket by value approach the gateway stores its state in a ticket, a data structure that is encrypted and integrity-protected by a key known only to the gateway. The ticket is passed to the client, who treats it as an opaque string, and sent back to the gateway when the session is resumed, so the gateway can decrypt it and recover the state.
The client behaves identically in both cases and in general does not know which approach the gateway is using. Since the ticket or reference is interpreted only by the party that created it, the document does not specify an exact format for it, but Appendix A gives examples of both. The approach is similar to TLS session resumption, with the adaptations IKEv2 needs for its two-phase structure. The stated goals are to use only symmetric cryptography, to provide cryptographic agility, and to have no negative impact on IKEv2 security features. The non-goals are failover from one gateway to another, load balancing among gateways, and specifying how a client detects the need for resumption.
Setting up a protected connection from scratch takes several round trips and a lot of cryptography, and with a password prompt it can take even longer. This RFC lets a client that was disconnected pick up where it left off, using a small piece of data the gateway gave it earlier.
| Aspect | Ticket by reference | Ticket by value |
|---|---|---|
| Where the state lives | On the gateway, in a ticket store | Inside the ticket, encrypted |
| What the client holds | An opaque reference, such as an index | The opaque ticket itself |
| Encryption | Not needed, but integrity protection is recommended | MUST be encrypted and integrity-protected |
| What the client does | The same in both cases | The same in both cases |
§3Usage Scenarios
The specification envisions two usage scenarios. The first is like the use case in the IKEv2 specification, where IPsec tunnel mode is used to establish a secure channel between a remote access client and a gateway, and the traffic flow may be between the client and entities beyond the gateway. The second focuses on using transport or tunnel mode to secure the communication between two endpoints, such as two servers, where the two endpoints have a client-server relationship with respect to a protocol that runs using the protections afforded by the IPsec SA.
In the first use case an end host with a host implementation of IPsec establishes a tunnel mode IPsec SA with a gateway in a remote network using IKEv2. At a later stage, when the client needs to re-establish the IKEv2 session, it may choose to establish IPsec SAs using a full IKEv2 exchange or the IKE_SESSION_RESUME exchange.
The RFC lists several situations where the mechanism is useful. If a client temporarily loses network connectivity and the IKE SA times out through the liveness test facility, known as dead peer detection, the mechanism re-establishes the SA with less overhead in network, CPU and authentication infrastructure, and without requiring user interaction for authentication. If connectivity problems affect a large number of clients, as at a large remote access VPN gateway, then when connectivity is restored all the clients might reconnect almost simultaneously, and the mechanism reduces the load spike for cryptographic operations and authentication infrastructure. Losing connectivity can also be predictable and planned, for example putting a laptop into stand-by mode before travelling, and the mechanism re-establishes the SA when the laptop is switched back on. The RFC notes that such user-level resumption may often be disallowed by policy, and that this document requires the client to destroy the ticket when the user explicitly logs out.
The main case is a remote access client that lost its connection to a VPN gateway. The other case is two servers that keep a protected channel between them. Resumption helps most when many clients reconnect at once, because the gateway does far less work than if every client authenticated from scratch.
§4.1Requesting and Receiving a Ticket
A client MAY request a ticket in several exchanges: in an IKE_AUTH exchange, which is the typical case; in a CREATE_CHILD_SA exchange when an IKE SA is rekeyed, and only when the client initiates that exchange; in an Informational exchange at any time, for example when the gateway previously replied with an acknowledgement instead of a ticket, when the ticket lifetime is about to expire, or following a gateway-initiated IKE rekey, all of which must be client-initiated; and while resuming an IKE session, in the IKE_AUTH exchange that follows an IKE_SESSION_RESUME exchange.
Normally the client requests a ticket in the third message of an IKEv2 exchange, the first of IKE_AUTH, by adding an N(TICKET_REQUEST) notification.
When a responder receives that request and supports the extension, it MUST do one of three things. It may create a ticket and return it with an N(TICKET_LT_OPAQUE) payload in a subsequent message. It may return an N(TICKET_NACK) if it refuses to grant a ticket for some reason. Or it may return an N(TICKET_ACK) if it cannot grant a ticket immediately, for example because of packet size limitations, in which case the client MAY request a ticket later using an Informational exchange at any time during the lifetime of the IKE SA. Whichever it does, there is no change to the behaviour of the responder with respect to the IKE exchange, and the proper IKE response, such as an IKE_AUTH response or an error notification, MUST still be sent.
The initiator receives the ticket and may accept it, provided the IKEv2 exchange was successful, and it may be used later with a responder that supports the extension. When a multi-round-trip IKE_AUTH exchange is used, the request MUST be included in the first IKE_AUTH request, and the ticket, nack or ack MUST only be returned in the final IKE_AUTH response. When the client accepts the ticket it stores it in its local storage for later use, along with the IKE SA to which the ticket refers. Since the ticket itself is opaque to the client, the local storage MUST also include all the items the RFC marks as coming from the ticket.
The client asks for a ticket during the normal handshake, by adding a small notification to the third message. The gateway then either sends the ticket, refuses, or says it cannot do it now and will send it later. The client keeps the ticket with the details of the session it belongs to.
- IKE_SA_INITHDR, SAi1, KEi, Ni
- IKE_SA_INIT REPLYHDR, SAr1, KEr, Nr [, CERTREQ]
- IKE_AUTH WITH TICKET REQUESTHDR, SK {IDi, AUTH, SAi2, TSi, TSr, N(TICKET_REQUEST)}
- IKE_AUTH WITH TICKETHDR, SK {IDr, AUTH, SAr2, TSi, TSr, N(TICKET_LT_OPAQUE)}
§4.3Presenting the Ticket
It is up to the client's local policy to decide when communication with the responder is seen as interrupted and the resumption procedure is started. A client MAY initiate a regular, non-ticket-based IKEv2 exchange even if it is in possession of a valid, unexpired ticket, and it MUST NOT present a ticket when it knows that the ticket's lifetime has expired. Tickets are intended for one-time use: a client MUST NOT reuse a ticket, and a reused ticket SHOULD be rejected by a gateway. A ticket counts as used only when an IKE SA has been established successfully with it.
The RFC specifies a new IKEv2 exchange type called IKE_SESSION_RESUME, whose value is 38. It is equivalent to the IKE_SA_INIT exchange and MUST be followed by an IKE_AUTH exchange. The initiator sets the SPIi value in the header to a new, unique value and the SPIr value to 0. When the responder receives a ticket it MUST either accept it, responding with the SPIi copied from the request and a new, unique SPIr, or reject it for any reason with an unprotected N(TICKET_NACK), in which case the initiator should re-initiate a regular IKE exchange. One such case is a ticket for an IKE SA that the responder itself has previously terminated, which may indicate inconsistent state between the initiator and the responder.
Where not otherwise specified, the exchange behaves exactly like the IKE_SA_INIT exchange. The client MAY resume from any IP address and port, regardless of its original address, although the gateway MAY reject the exchange if its policy depends on the client's address. The first message MAY be rejected in denial-of-service situations, with the initiator instructed to send a cookie. Notifications normally associated with IKE_SA_INIT can be sent, in particular NAT detection payloads, and the client's NAT traversal status SHOULD be determined anew, because NAT status is explicitly not part of the session resumption state.
Following the IKE_SESSION_RESUME exchange the client MUST initiate an IKE_AUTH exchange. The AUTH payload is derived in a manner similar to IKEv2 pre-shared secret authentication, as AUTH = prf(SK_px, <message octets>), where each side uses its own value for SK_px, namely SK_pi for the initiator and SK_pr for the responder, both taken from the newly generated IKE SA. The IDi value sent in the IKE_AUTH exchange MUST be identical to the value included in the ticket, a CERT payload MUST NOT be included, and the IDr value sent by the gateway, and optionally by the client, MUST also be identical to the value included in the ticket. When the responder accepts the ticket and the IKE_AUTH exchange completes, the old SA SHOULD be silently deleted without sending a DELETE informational exchange, and consequently all the dependent IPsec Child SAs are also deleted.
To resume, the client sends the ticket in a short exchange that works like the first message of a normal handshake, then confirms the session in the usual authentication exchange. The ticket already carries the proof of identity, so no certificate or password is needed again.
- IKE_SESSION_RESUMEHDR, [N(COOKIE),] Ni, N(TICKET_OPAQUE). New SPIi, SPIr set to 0.
- IKE_SESSION_RESUME REPLYHDR, Nr [,N+]. The responder copies SPIi and sets a new SPIr.
- IKE_AUTHAUTH = prf(SK_px, message octets). The IDi and IDr values must match the ticket.
- IKE_AUTH REPLYThe old SA is then silently deleted.
§5.1Generating Cryptographic Material for the Resumed IKE SA
The cryptographic material is refreshed based on the ticket and the nonce values Ni and Nr from the current exchange. A new SKEYSEED value is derived as SKEYSEED = prf(SK_d_old, "Resumption" | Ni | Nr), where SK_d_old is taken from the ticket and the literal string is encoded as 10 ASCII characters with no NULL terminator. The keys are then derived as they are in IKEv2, with {SK_d | SK_ai | SK_ar | SK_ei | SK_er | SK_pi | SK_pr} = prf+(SKEYSEED, Ni | Nr | SPIi | SPIr), where the SPI values are those created in the new exchange and the pseudorandom function is determined from the SA value in the ticket.
The RFC also tabulates the IKE and IPsec state after resumption and how it relates to the state before the IKE SA was interrupted. An item marked as coming from the ticket is retrieved by the client from its local store, encoded into the ticket by the gateway in the ticket-by-value case, or fetched by the gateway from the ticket store in the ticket-by-reference case. So IDi and IDr come from the ticket, as do the authentication method and, where applicable, the certificates. The NAT detection status and the SPIs come from the new exchange, the IKE SA sequence numbers are reset to 0 in IKE_SESSION_RESUME and then incremented normally, and the IKE SA algorithms come from the ticket. The IKE SA window size is reset to 1, Child SAs are created in the new exchange, and the internal IP address, other Configuration Payload information and the peer vendor IDs are not resumed.
The new keys are computed from the old session key and the random numbers in the new exchange, so both ends end up with fresh keys that nobody has seen before. The identity and the authentication method come from the ticket, but anything about the network connection is worked out again.
SKEYSEED = prf(SK_d_old, "Resumption" | Ni | Nr)
{SK_d | SK_ai | SK_ar | SK_ei | SK_er | SK_pi | SK_pr} =
prf+(SKEYSEED, Ni | Nr | SPIi | SPIr)| State item | After resumption |
|---|---|
| IDi and IDr | From the ticket, and exchanged again in IKE_AUTH |
| Authentication method and certificates | From the ticket |
| NAT detection status | From the new exchange |
| IKE SA sequence numbers | Reset to 0 in IKE_SESSION_RESUME |
| IKE SA window size | Reset to 1 |
| Child SAs | Created in the new exchange |
| Internal IP address and other configuration payloads | Not resumed |
§6Ticket Handling and Lifetime
When passing a ticket by value to the client, the ticket content MUST be integrity-protected and encrypted. A ticket by reference does not need to be encrypted, as it holds no sensitive material such as keying material, but access to the storage where that material is kept MUST be protected so that only authorised access is allowed. The RFC compares such a ticket to the stub concept and to a TLS session identifier. Although not strictly required for cryptographic protection, it is RECOMMENDED to integrity-protect a ticket by reference, because failing to do so could result in security vulnerabilities on the gateway side, depending on the format of the reference, such as the gateway accessing unintended URLs or SQL injection.
A ticket by value MUST encode all the state information marked as coming from the ticket, and the same state MUST be stored in the ticket store in the ticket-by-reference case. A ticket by value MUST include a protected expiration time, an absolute time value that SHOULD correspond to the value included in the TICKET_LT_OPAQUE payload, and it MUST additionally include a key identity field, so that the keys for ticket encryption and authentication can be changed and, when necessary, algorithms replaced.
Each ticket is associated with a single IKE SA. When an IKE SA is deleted by the client or the gateway, the client MUST delete its stored ticket, and when credentials associated with the IKE SA are invalidated, for example when a user logs out, the ticket MUST be deleted. When the IKE SA is rekeyed the ticket is invalidated and the client SHOULD request a new one.
The lifetime of the ticket sent by the gateway SHOULD be the minimum of the IKE SA lifetime, per the gateway's local policy, and its re-authentication time. Even if neither of these is enforced by the gateway, a finite lifetime MUST be specified for the ticket. The key used to protect the ticket MUST have a lifetime significantly longer than the lifetime of an IKE SA. In normal operation the client requests a ticket when establishing the initial IKE SA, and then every time the SA is rekeyed or re-established because of re-authentication.
A ticket is only good for one session and one try. The client has to throw it away when the session ends or when the user logs out, and the gateway has to give it an end date. The key that protects the ticket must live longer than any single session.
§7The Notify Payloads
The RFC defines five notifications. The TICKET_LT_OPAQUE payload carries a ticket together with its remaining lifetime: it consists of the Notify message header, a Lifetime field and the ticket itself. The four octet Lifetime field contains a relative time value, the number of seconds until the ticket expires, encoded as an unsigned integer in network byte order. The TICKET_OPAQUE payload consists of the Notify message header and the ticket itself, with no lifetime value, and it is the payload the client sends back when it presents the ticket. TICKET_REQUEST, TICKET_ACK and TICKET_NACK carry no data.
For all of these notifications the Protocol ID and the SPI Size fields MUST both be sent as 0.
The ticket travels in a small notification message. One kind carries the ticket and how long it stays valid, and another carries just the ticket, which is what the client sends back when it wants to resume.
- 16409TICKET_LT_OPAQUE — Carries a ticket and the number of seconds until it expires.
- 16410TICKET_REQUEST — The client asks for a ticket. Carries no data.
- 16411TICKET_ACK — The gateway cannot issue a ticket right now. Carries no data.
- 16412TICKET_NACK — The gateway refuses to issue a ticket. Carries no data.
- 16413TICKET_OPAQUE — Carries the ticket the client presents when it resumes.
§9Security Considerations
A man in the middle may try to eavesdrop on an exchange to obtain a ticket by value and use it to establish a session with the responder. Since all the exchanges where the client obtains a ticket are encrypted, this is only possible by listening in on the client's use of the ticket to resume a session. Because the ticket's contents are encrypted and the attacker does not know the corresponding secret key, a stolen ticket cannot be used to resume a session successfully, and the responder MUST use strong encryption and integrity protection of the ticket. A ticket by reference cannot be used by an attacker to resume a session either, for the same reason, because the attacker would not be able to prove during IKE_AUTH its knowledge of the secret part of the IKE state embedded in the ticket. Access control must also ensure the reference is disclosed only to authorised entities.
A malicious user could forge or alter a ticket by value in order to resume a session, extend its lifetime, impersonate another user or gain additional privileges. This is not possible if the content is protected using a strong integrity protection algorithm. For a ticket by reference, an adversary may attempt to construct a fake reference pointing at state information stored by the responder, but the attack fails because the adversary is not in possession of the keying material associated with the IKE SA.
For denial of service, an adversary could send a large number of tickets by value to burden the gateway's CPU or exhaust its memory with half-open IKE state, so ticket verification should be lightweight and use efficient symmetric key algorithms. A large number of tickets by reference may lead to an amplification attack, because the responder is forced to resolve the reference to discover that the adversary holds no keying material, so the resolution protocol should be as lightweight as possible and should not generate a large number of messages. The regular IKEv2 cookie mechanism can be used for state-overflow denial-of-service situations.
On key management, the RFC recommends that the keys used to protect a ticket by value be generated securely following the randomness recommendations in RFC 4086, that the keys and cryptographic protection algorithms be at least 128 bits in strength, that the keys not be used for any other purpose than generating and verifying tickets, that the keys be changed regularly, and that they be changed if the ticket format or the algorithms change. A misbehaving client could present a ticket for an IKE SA that has already been deleted, so the gateway may maintain an Invalid Ticket List, a simple blacklist of revoked tickets, or use Bloom filters to keep the list in constant space; a policy requiring shorter ticket lifetimes, an hour for example, significantly mitigates the issue. Finally, since the IKE_SESSION_RESUME exchange provides no confidentiality, an on-path adversary could observe several exchanges that use the same state and conclude that they belong to the same endpoints, so this document requires the ticket to be presented to the responder only once.
The ticket acts as proof that the client was here before, so it has to be protected like a password. If it is stolen it cannot be used, because the attacker cannot produce the matching keys, and it must not be used twice, because repeated use would let an observer link the sessions together.