EAP
Internet Draft H. Tschofenig
D. Kroeselberg
Siemens
Y. Ohba
Toshiba
Document: draft-tschofenig-eap-ikev2-02.txt draft-tschofenig-eap-ikev2-03.txt
Expires: April 2002 October 2003 August 2004 February 2004
EAP IKEv2 Method
(EAP-IKEv2)
Status of this Memo
This document is an Internet-Draft and is subject to all provisions
of Section 10 of RFC2026.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
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."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/1id-abstracts.html
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html
Abstract
EAP-IKEv2 is an EAP method which reuses the cryptography and the
payloads of IKEv2, creating a flexible EAP method that supports both
symmetric and asymmetric authentication. Furthermore protection of
legacy authentication mechanisms is supported. This EAP method offers the
security benefits of IKEv2 authentication and key agreement without
the goal of establishing IPsec security associations.
Table of Contents
1. Introduction..................................................2
2. IKEv2 and EAP-IKEv2 Overview..................................3
3. Terminology...................................................4
4. Protocol overview.............................................4
5. Identities used in EAP-IKEv2..................................7
6. Packet Format.................................................9 Format.................................................8
7. Retransmission...............................................10
8. Key derivation...............................................10
9. Error Handling...............................................11
10. Security Considerations.....................................13 Fast Reconnect..............................................12
11. Open Issues.................................................13 Channel Binding.............................................14
11.1 Channel Binding Procedure in Full Authentication........15
11.2 Channel Binding Procedure in Fast Reconnect.............16
11.3 Channel Binding Error Indication........................16
11.4 Notify Payload Types for Channel Binding................16
11.5 Examples................................................18
12. Normative References........................................13 Security Considerations.....................................22
12.1 General Considerations..................................22
12.2 Security Claims.........................................22
13. Open Issues.................................................23
14. Normative References........................................24
15. Informative References......................................14
Acknowledgments.................................................14 References......................................25
Acknowledgments.................................................25
Author's Addresses..............................................15 Addresses..............................................26
Full Copyright Statement........................................15 Statement........................................26
1. Introduction
This document specifies the EAP-IKEv2 authentication method. EAP-
IKEv2 The
main design goal for EAP-IKEv2 is to provide a flexible and
efficient EAP method which makes the IKEv2 protocolĘs protocol's features
available for scenarios using EAP-based authentication.
The main advantage of EAP-IKEv2 is that it does not define a new
cryptographic protocol, but re-uses the IKEv2 authentication
protocol,
exchanges, and thereby provides strong, well-analyzed, cryptographic
properties as well as broad flexibility. This includes the support
of authentication methods and configuration payloads for remote
access scenarios.
EAP-IKEv2 can be used directly to mutually authenticate provides mutual authentication between EAP peers. This may
be based on either symmetric methods using pre-shared keys, or on
asymmetric methods based on public/private key pairs, Certificates
and CRLs. In addition, EAP-IKEv2 It is possible to use different types of authentication
for the different directions, e.g. the server uses certificate-based
authentication whereas the client uses a symmetric method.
IKEv2 supports two-phased authentication schemes by establishing a
server-authenticated secure
tunnel, tunnel and by subsequently protecting an
EAP authentication allowing for legacy client authentication
methods. Hence, EAP-IKEv2
provides a secure EAP EAP-IKEv2, however, does not support this optional
tunneling method. feature of IKEv2 in this version, which allows to increase
the EAP-IKEv2 method performance and decrease implementation
complexity.
A non-goal of EAP-IKEv2 (and basically the major difference to plain
IKEv2) is the establishment of IPsec security associations, as this
would not make much sense in the standard AAA three-party scenario,
consisting of an EAP peer, an authenticator (NAS) and a back-end
authentication server terminating EAP. IPsec SA establishment may be
required locally (i.e., between the EAP peer and some access
server). However, SA establishment within an EAP method would only
provide SAs between the EAP peer and the back-end authentication
server. Other approaches as e.g., those of the IETF PANA group are
considered more appropriate in this case.
2. IKEv2 and EAP-IKEv2 Overview
IKEv2 [Kau03] [Kau04] is a protocol which consists of two exchanges:
(1) an authentication and key exchange protocol which establishes an
IKE-SA.
(2) messages and payloads which focus on the negotiation of
parameters in order to establish IPsec security associations (i.e.,
Child-SAs). These payloads contain algorithm parameters and traffic
selector fields.
In addition to the above-mentioned parts IKEv2 also includes some
payloads and messages which allow configuration parameters to be
exchanged primarily for remote access scenarios.
The EAP-IKEv2 method defined by this document uses the IKEv2
payloads and messages used for the initial IKEv2 exchange which
establishes an IKE-SA.
IKEv2 provides an improvement over IKEv1 [RFC2409] as described in
Appendix A of [Kau03]. [Kau04]. Important for this document are the reduced
number of initial exchanges, support of legacy authentication, decreased latency of the initial
exchange, optional Denial-of-
Service (DoS) protection capability and some other fixes (e.g., hash problem). IKEv2 is a
cryptographically sound protocol that has received a considerable
amount of expert review and that benefits from a long practical
experience with IKE.
The goal of EAP-IKEv2 is to inherit these properties within an
efficient, secure EAP method.
In addition, IKEv2 provides authentication and key exchange
capabilities which allow an entity to use symmetric as well as
asymmetric authentication within a single protocol. Such flexibility
is considered important for an EAP method and is provided by EAP-
IKEv2.
[Per03] provides a good tutorial for IKEv2 design decisions.
EAP-IKEv2 therefore provides
a) well-known IKEv2 symmetric/asymmetric authentication and
b) a new EAP tunneling method.
EAP-IKEv2 provides a secure fragmentation mechanism in which
integrity protection is performed for each fragment of an IKEv2
message.
3. Terminology
This document does not introduce new terms other than those defined
in [RFC2284] or in [Kau03]. [Kau04].
The keywords MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD,
SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, when they appear in this
document, are to be interpreted as described in [RFC2119].
4. Protocol overview
This section provides some overview over EAP-IKEv2 message
exchanges. Note that some mandatory IKEv2 payloads are omitted omitted, or
profiled (such as SAi2 and
SAr2) which are mandatory for IKEv2 but are SAr2), as it is not required in EAP-
IKEv2 since they are used supported to
establish an IPsec SA. (ESP, AH) SAs in EAP-IKEv2.
IKEv2 uses the same protocol message exchanges for both symmetric
and asymmetric authentication. The difference lies only in the
computation of the AUTH payload. See Section 2.15 of [Kau03] [Kau04] for
more information about the details of the AUTH payload computation.
It is even possible to combine symmetric (e.g., from the client to
the server) with asymmetric authentication (e.g., from the server to
the client) in a single protocol exchange. Figure 1 depicts such a
protocol exchange.
Message exchanges are reused from [Kau03], [Kau04], and are adapted. Since
this document does not describe frameworks or particular
architectures the message exchange takes place between two parties -
between the Initiator (I) and the Responder (R). In the context of
EAP the Initiator is often called Authenticating Peer whereas takes the
Responder is referred as Authenticator. role of the EAP server and the responder
matches the EAP peer.
The first message flow shows EAP-IKEv2 without the optional DoS
protection exchanges. EAP-IKEv2 full successful exchange.
The core EAP-IKEv2 exchange (message (4) (3) -
(7)) (6)) consists of four
messages (two round trips)_only. The first two messages constitute
the standard EAP identity exchange and are optional; they are not
mandatory if
required in case the EAP server is known. In the exchange, the EAP
server (B) takes the role of the IKEv2 initiator and the EAP peer
(A) acts as the IKEv2 responder.
1) I A <-- R: B: EAP-Request/Identity
2) I A --> R: B: EAP-Response/Identity(Id)
3) I A <-- R: EAP-Request/EAP-Type=EAP-IKEv2(Start)
4) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(HDR(A,0), B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
5) I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ])
6) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,], AUTH})
7) I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDr, [CERT,] AUTH})
8) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(Finish)
9) I
7) A <-- R: B: EAP-Success
Figure 1: EAP-IKEv2 successful message flow
The subsequent
Figure 2 shows the message flow shows EAP-IKEv2 with DoS protection
enabled. The IKEv2 DoS protection mechanism uses cookies and keeps in case the responder stateless when it receives EAP peer fails to
authenticate the first IKEv2 message,
preventing it from performing heavy cryptographic operations based
on this first incoming message. As a consequence of DoS protection
an additional round trip (message (5) and (6)) is required. EAP server.
1) I A <-- R: B: EAP-Request/Identity
2) I A --> R: B: EAP-Response/Identity(Id)
3) I A <-- R: EAP-Request/EAP-Type=EAP-IKEv2(Start)
4) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(HDR(A,0), B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
5) I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,0), N(COOKIE-REQUIRED), N(COOKIE))
6) I
4) A --> R: B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,0), N(COOKIE), SAi1, KEi, Ni)
7) I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ])
8) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,], AUTH})
9) I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDr, [CERT,] AUTH})
10) I --> R: EAP-Response/EAP-Type=EAP-IKEv2(Finish)
11) I {N(AUTHENTICATION_FAILED)})
7) A <-- R: EAP-Success B: EAP-Failure
Figure 2: EAP-IKEv2 with Cookies
The Secure Legacy Authentication (SLA) EAP message exchange shown in failed server authentication
Figure 3 is taken from Section 2.16 of [Kau03] and adapted. It
provides an example of a successful inner EAP exchange using shows the
EAP-SIM Authentication method [HS03], which is secured by message flow in case the IKE-
SA.
Implementations EAP server fails to
authenticate the EAP peer. The EAP peer MUST ensure that infinite recursions of send an empty EAP-IKEv2
informational message in reply to the EAP and EAP-
IKEv2 exchanges are not allowed. (TBD: some limit necessary)
I server's error indication.
The EAP server answers with an EAP-Failure.
1) A <-- R: B: EAP-Request/Identity
I
2) A --> R: B: EAP-Response/Identity(Id)
I
3) A <-- R: EAP-Request/EAP-Type=EAP-IKEv2(Start)
I --> R: EAP-Response/EAP-Type=EAP-IKEv2(
HDR, B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(
HDR,
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ])
I --> R: EAP-Response/EAP-Type=EAP-IKEv2(
HDR,
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,]})
I <-- R: EAP-Request/EAP-Type=EAP-IKEv2(HDR, [IDr,], AUTH})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDr, [CERT,] AUTH, EAP(EAP-Request /SIM
/Start(AT_VERSION_LIST))})
I --> R: EAP-Response/EAP-Type=EAP-IKEv2(HDR, SK {EAP(EAP-
Response/SIM/Start(AT_NONCE_MT, AT_SELECTED_VERSION)),
[AUTH]})
I AUTH})
7) A <-- R: EAP-Request/EAP-Type=EAP-IKEv2(HDR, B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {EAP(EAP-
Request/SIM/Challenge(AT_RAND, AT_MAC)), [AUTH]})
I {N(AUTHENTICATION_FAILED)})
8) A --> R: B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR,
HDR(A,B), SK {EAP(EAP-Response/SIM/Challenge(AT_MAC) ), [AUTH]})
I {})
9) A <-- R: EAP-Success B: EAP-Failure
Figure 3: EAP-IKEv2 SLA with EAP-SIM Authentication
Please note that the message flow in Figure 3 does not include an
EAP-Request/Identity and the corresponding EAP-Response/Identity
message inside the EAP-IKEv2 tunnel. Although it would be possible
to perform such an exchange IKEv2 suggests using the IDi payload for
this purpose. As a consequence the initiators identity is not
protected against active attacks. failed client authentication
Since the goal of this EAP method is not to establish an IPsec SA
some payloads used in IKEv2 are omitted. In particularly the
following messages and payloads are SHOULD not required: be sent:
- Traffic Selectors Selector (TS)payloads
- IPsec SA negotiation payloads
(e.g., CREATE_CHILD_SA exchange that carry SA proposals for protocol IDs other than
1(IKE), i.e., SA payloads with protocol ID 2 (ESP) or SAx2 payloads)
- ECN Notification
- Port handling 3 (AH)
- NAT traversal ESN (extended sequence number) transforms
Some of these messages and payloads are optional in IKEv2.
In general it does not make sense to directly negotiate IPsec SAs
with
within EAP-IKEv2, as such SAs were unlikely to are not required between the EAP
endpoints and as SAs cannot be transferred to different AAA entities
by standard AAA protocols.
Consequently, mechanisms and payloads that are not supported by EAP-
IKEv2 are:
- ECN Notifications as specified in section 2.24 of [Kau04].
- IKE-specific port handling
- NAT traversal
Since the EAP server acts as the initiator of the initial IKEv2
exchange, a number of optional payloads used between for realizing specific
features in IKEv2 are not supported by EAP-IKEv2, as they are
intended for the client side (e.g. for corporate access scenarios)
in plain IKEv2. These payloads MUST not be sent by an EAP-IKEv2
entity. EAP-IKEv2 entities receiving such payloads MUST respond with
the appropriate error messages as defined in [Kau04]. These payloads
are:
- Configuration (CFG) payloads as specified in 3.15 of [Kau04].
These payloads MUST not be sent by an EAP-IKEv2 implementation. EAP-
IKEv2 entities receiving such payloads MUST ignore configuration
payloads as described for minimal implementations in 3.15 of
[Kau04].
- EAP
endpoints. payloads as specified in section 3.16 of [Kau04]. These
payloads allow to run an inner EAP exchange for secure legacy
authentication through an IKE SA. EAP-IKEv2 implementations acting
as initiator MUST include and AUTH payload in the initial IKE_AUTH
message (message 3 of the initial IKE exchange). EAP-IKEv2
implementations receiving initial IKE_AUTH messages as responders
that indicate the initiator's desire to start extended
authentication MUST be answered with an AUTHENTICATION_FAILED
notification as the response.
IKEv2 also provides optional functionality for additional DoS protection
by adding a roundtrip to the initiator initial exchanges, see section 2.xx of
[Kau04]. As this is intended to request
address information from protect the IKEv2 responder as described but in Section 2.19
EAP-IKEv2 the EAP server takes the role of [Kau03]. Using this functionality the initiator, it is possible for an end host not
recommended to securely request address configuration information from the local
network. use this feature of IKEv2 for server protection.
5. Identities used in EAP-IKEv2
A number of different places allow to convey identity information in
IKEv2, when combined with EAP. This section describes their function
within the different exchanges of EAP-IKEv2. Note that EAP-IKEv2
does not introduce more identities than any other tunneling
approach. non-tunneling EAP
methods. Figure 4 shows which identities are used during the
individual phases of the protocol.
+-------+ +-------------+ +---------+ +--------+
|Client | |Front-End | |Local AAA| |Home AAA| |AAA |
| |Authenticator| |Server | |Authenticator| |Server |
+-------+ +-------------+ +---------+ +--------+
EAP/Identity-Request
<---------------------
(a) EAP/Identity-Response
---------------------------------->
Tunnel-Establishment
(b) (Identities of IKEv2 are used)
Server (Network) Authentication
<----------------------------------
...
---------------------------------->
+---------------------------------+
| Secure Tunnel |
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~+
| Secure Legacy Authentication |
| protected with the IKE-SA |
(c) | (Identities of the tunneled |
| EAP method are used) |
| Client Authentication |
|---------------------------------+---------------->
|<--------------------------------+-----------------
+---------------------------------+
Figure 4: Identities used in EAP-IKEv2
a) The first part of the (outer) EAP message exchange provides
information about the identities of the EAP endpoints. This message
exchange mainly is an identity request/response. This exchange is
optional if the EAP server is known already or can be learned by
other means.
b) The identities used Identities exchanged within EAP-IKEv2 for both the initiator and
the responder. The initiator identity is often associated with a
user identity such as a fully-qualified RFC 822 email address. The
identity of the responder might be a FQDN. The identity is of
importance for authorization.
c)
For secure legacy authentication an EAP message exchange is
protected with the established IKE-SA as shown in Figure 3. This
exchange again adds EAP identities.
This inner EAP message exchange serves the purpose of client
authentication. The two carrying identities used thereby are the EAP identity
(i.e., a NAI) and possibly a separate identity for in EAP-IKEv2, implementations MUST follow
the selected EAP
method.
The large number of identities is required due rules given in [Kau04], section 3.5, i.e., MUST be configurable
to nesting send at least one of
authentication methods ID_IPV4_ADDR, ID_FQDN, ID_RFC822_ADDR, or
ID_KEY_ID, and due MUST be configurable to overloaded function of the
identity for routing (i.e., authentication end point indication).
The number of recursions of EAP and IKEv2 is limited, see Section 4.
Hence with this additional (nested) EAP exchange the end point accept all of
the EAP-IKEv2 exchange might not these types.
Implementations SHOULD be the same as the end point capable of the
inner EAP exchange which is protected by the IKE-SA (which in this
case is not protected by the IKE-SA any more between the EAP-IKEv2
endpoint generating and the endpoint accepting all of the inner EAP exchange, but might be
protected by other means that are not considered in this document).
these types.
6. Packet Format
The IKEv2 payloads, which are defined in [Kau03], [Kau04], are embedded into
the Data field of the standard EAP Request/Response packets. The
Code, Identifier, Length and Type field is described in [RFC2284].
The Type-Data field carries a one byte Flags field following the
IKEv2 payloads. Each IKEv2 payload starts with a header field HDR
(see [Kau03]). [Kau04]).
The packet format is shown in Figure 5.
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Code | Identifier | Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Type | Flags | Message Length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Message Length | Data ... ~
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Integrity Checksum Data |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 5: Packet Format
No additional packet formats other than those defined in [Kau03] [Kau04] are
required for this EAP method.
The Flags field is required to indicate Start used for fragmentation support. The S and Finish messages
which F bits
are required due to the asymmetric nature of IKEv2 and the
Request/Response message handling of EAP. reserved for future use.
Currently five bits of the eight bit flags field are defined. The
remaining bits are set to zero.
0 1 2 3 4 5 6 7
+-+-+-+-+-+-+-+-+
|S F L M 0 I 0 0 0|
+-+-+-+-+-+-+-+-+
S = EAP-IKEv2 start message (reserved)
F = EAP-IKEv2 finish message (reserved)
L = Length included
M = More fragments
I = Integrity Checksum Data included
EAP-IKEv2 messages which have neither the S nor the F flag set
contain regular IKEv2 message payloads inside the Data field.
With regard to fragmentation we follow the suggestions and
descriptions given in Section 2.8 of [PS+03]: The L indicates that a
length field is present and the M flag indicates fragments. The L
flag MUST be set for the first fragment and the M flag MUST be set
on all fragments expect for the last one. Each fragment sent must
subsequently be acknowledged.
The Message Length field is four octets long and present only if the
L bit is set. This field provides the total message length that is
being fragmented (i.e., the length of the Data field.).
The Integrity Checksum Data is the cryptographic checksum of the
entire EAP message starting with the Code field through the Data
field. This field presents only if the I bit is set. The field
immediately follows the Data field without adding any padding octet
before or after itself. The checksum MUST be computed for each
fragment (including the case where the entire IKEv2 message is
carried in a single fragment) by using the same key (i.e., SK_ai or
SK_ar) that is used for computing the checksum for the IKEv2
Encrypted payload in the encapsulated IKEv2 message. The Integrity
Checksum Data field is omitted for other packets. To minimize DoS
attacks on fragmented packets, messages that are not protected
SHOULD NOT be fragmented. Note that IKE_SA_INIT messages are the
only ones that are not encrypted or integrity protected, however,
such messages are not likely to be fragmented since they do not
carry certificates.
The EAP Type for this EAP method is <TBD>.
7. Retransmission
Since EAP authenticators support a timer-based retransmission
mechanism for EAP Requests and EAP peers retransmit the last valid
EAP Response when receiving a duplicate EAP Request message, IKEv2
messages MUST NOT be retransmitted based on timers, when used as EAP
authentication method.
8. Key derivation
The EAP-IKEv2 method described in this document generates session
keys. These On the one hand, these session keys are used to establish an IKE-SA which
provides within the IKE-
SA, for protection of subsequent EAP-IKEv2 payloads. To export a
session key payloads, e.g., AUTH exchanges or
notifications. On the other hand, additional keys are derived to be
exported as part of the EAP keying framework [AS+03] it is
required to derive additional session keys for usage with EAP (i.e., MSK,
EMSK and IV). It is good cryptographic security practice to use
different keys for different "applications". Hence we suggest
reusing of the key derivation function suggested in Section 2.17 of
[Kau03]
[Kau04] to create keying material KEYMAT.
The key derivation function defined is KEYMAT = prf+(SK_d, Ni | Nr),
where Ni and Nr are the Nonces from the IKE_SA_INIT exchange.
Since the required amount of keying material is greater than the
size of the output of the prf algorithm the prf is used iteratively.
Section 2.13 of [Kau04] describes this mechanism in detail.
According to [AS+03] the keying material of MSK, EMSK and IV have to
be at minimum 64, 64 and 64 octets long.
The produced keying material for MSK, EMSK and IV MUST be twice at least
the minimum size (i.e., 128 64 octets). The keying material KEYMAT is
split into the MSK, EMSK and IV part.
Figure 6 describes the keying hierarchy of EAP-IKEv2 graphically.
This figure is adopted from Figure 2 of [AS+03].
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-++ ---+
| | ^
| EAP-IKEv2 Method | |
| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-++ +------------------+ | |
| | EAP-IKEv2 Diffie-Hellmann | | EAP-IKEv2 prot. | | |
| | derived and authenticated key | | session specific | | |
| | SK_d | | state (Nonce i,j)| | |
| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-++ +-------------+----+ | |
| | | | Local |
| | | | to EAP |
| | | | Method |
| | | | |
| | | | |
| | | | |
| | | | |
| +---------------+-------------+ | | |
| | | | | | |
| +-+-+-+-+-++ +-+-+-+-+-++ +-+-+-+-+-++ | |
| | MSK | |EMSK | | IV | | |
| |Derivation| |Derivation| |Derivation| | |
| +-+-+-+-+-++ +-+-+-+-+-++ +-+-+-+-+-++ | |
| | | | | V
+-+-+-+-+-+-+-+-+-+-+-+-+-+--+-+-++-+-+-+-+-------+-+-+----+ ---+
| | | ^
|MSK |EMSK |IV |
| | | |
| | | Exported |
| | | by EAP |
V V V Method |
+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+ |
| AAA Key Derivation | | Known | |
| Naming & Binding | |(Not Secret) | |
+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+ V
Legend:
MSK = Master Session Key (512 Bit)
EMSK = Extended Master Session Key (512 Bit)
SK_d = Session key derived by EAP-IKEv2
IV = Initialization Vector
Figure 6: EAP-IKEv2 Keying Hierarchy
9. Error Handling
As described in the IKEv2 specification, there are many kinds of
errors that can occur during IKE processing (i.e., processing the
Data field of EAP-IKEv2 Request and Response messages) and detailed
processing rules. EAP-IKEv2 follows the error handling rules
specified in the IKEv2 specification for errors on the Data field of
EAP-IKEv2 messages, with the following additional rules:
o
For an IKEv2 error that triggers an initiation of an IKEv2 exchange
(i.e., an INFORMATIONAL exchange), an EAP-IKEv2 message that
contains the IKEv2 request that is generated for the IKEv2 exchange
MUST be sent to the peering entity. In this case, the EAP message
that caused the IKEv2 error MUST be treated as a valid EAP message.
o
For an IKEv2 error for which the IKEv2 message that caused the error
is discarded without triggering an initiation of an IKEv2 exchange,
the EAP message that carries the the erroneous IKEv2 message MUST be
treated as an invalid EAP message and discarded as if it were not
received at EAP layer.
For an error occurred outside the Data field of EAP-IKEv2 messages,
including defragmentation failures, integrity check failures, errors
in Flag and Message Length fields, the EAP message that caused the
error MUST be treated as an invalid EAP message and discarded as if
it were not received at EAP layer.
When the EAP-IKEv2 method runs on a backend EAP server, an
outstanding EAP Request is not retransmitted based on timer and thus
there is a possibility of EAP conversation stall when the EAP server
receives an invalid EAP Response. To avoid this, the EAP server MAY
retransmit the outstanding EAP Request in response to an invalid EAP
Response. Alternatively, the EAP server MAY send a new EAP Request
in response to an invalid EAP Response with assigning a new
Identifier and putting the last transmitted IKEv2 message in the
Data field.
10. Fast Resume
TLS provides the capability Reconnect
EAP-IKEv2 supports fast reconnect, i.e., a successful reconnect
exchange creates a new IKE-SA by using an IKE CHILD_SA exchange. The
purpose of resuming a session. This offers
primarily performance improvement re-authentication exchange is to allow for a efficient
re-keying based on the existing IKE-SA in situations where
(depending on the given security policy) no full authentication is
required in case of an existing EAP-IKEv2 security context.
The fast reconnect exchange uses the IKE-SA rekeying as specified in
section 2.18 of [IKEv2]. However, the exchanges for EAP-IKEv2 do not
use the payloads for rekeying other than IKE SAs:
- The TS (traffic selector) payloads SHOULD not be sent by EAP-IKEv2
implementations.
- The [N] payload (REKEY_SA notification) SHOULD not be sent by EAP-
IKEv2 implementations.
During fast re-authentication, the new IKE_SA is computed as
specified in [IKEv2], section 2.18. The new keying material derived
from this IKE_SA is computed as in an initial EAP-IKEv2
authentication exchange.
Fast re-authentication allows for an optional new Diffie-Hellman
exchange.
The following exchange provides fast reconnect for EAP-IKEv2, where
A is the EAP peer (IKE responder) and key B is the EAP server (IKE
initiator):
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR, SK {SA, Ni, [KEi]})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR, SK {SA, Nr, [KEr]})
5) A <-- B: EAP-Success
Figure 7: Fast Reconnect Message Flow
The first two messages constitute the standard EAP identity exchange protocol run. In order
and are optional; they are not required in case the EAP server is
known.
Figure 8 shows the fast reconnect message flow in case the EAP peer
fails to resume a session two approaches
can be taken:
a) Generic approach
b) Method-specific approach re-authenticate the EAP server.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2
(HDR, SK {SA, Ni, [KEi]})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR, SK {N(AUTHENTICATION_FAILED)})
5) A <-- B: EAP-Failure
Figure 8: EAP-IKEv2 fast reconnect
(server authentication failed)
Figure 9 shows the fast reconnect message flow in case the EAP
server fails to re-authenticate the EAP peer. The idea of approach (a) is EAP peer MUST send
an empty EAP-IKEv2 informational message in reply to
- force each the EAP method
server's error indication. The EAP server answers with an EAP-
Failure.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR, SK {SA, Ni, [KEi]})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR, SK {SA, Nr, [KEr]})
5) A <-- B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(AUTHENTICATION_FAILED)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {})
7) A <-- B: EAP-Failure
Figure 9: EAP-IKEv2 fast reconnect
(client authentication failed)
IKE_SAs do not have lifetimes. Such lifetimes are therefore set by
local policies of the peers. Typically the peer setting the shorter
lifetime will therefore trigger the reconnect procedure.
Note: IKEv2 supports fast rekeying to create be initiated by both peers. As
the EAP server initiating the rekeying better fits the EAP message
flow, the EAP server, from an IKE perspective, is the initiator in
the rekeying exchange.
11. Channel Binding
EAP-IKEv2 provides a channel binding functionality [RFC2284bis] in
order for the EAP SA. peer and EAP server to make sure that the both
entities are given the same network access attributes such as
Calling-Station-Id, Called-Station-Id, and NAS-Port-Type by the NAS.
This SA is kept at achieved by using Notify payloads to exchange attribute data
between the EAP peer and the EAP server server.
A Notify payload that carries a null channel binding attribute is
referred to as a channel binding request. A Notify payload which
contains a non-null channel binding attribute and is sent in
response to a channel binding request is referred to as a channel
binding response. A pair of channel binding request and channel
binding response constitute a channel binding exchange. A distinct
Notify payload type is used for subsequent exchanges.
- built this functionality into EAP itself.
Approach (b) a particular type of channel binding
attribute, which is referred to as a channel binding attribute type.
It is allowed to carry multiple channel binding requests and/or
responses with different channel binding attribute types in a single
IKEv2 message. A set of channel binding exchanges performed in a
single round of EAP-IKEv2 full authentication or fast reconnect is
referred to as a channel binding procedure.
A Notify payload that is already used by existing methods using TLS. Choosing
(b) does not require any changes for reporting an error occurred during
a channel binding exchange is referred to as a channel binding error
indication.
EAP-IKEv2 offers a protected result indication mechanism (see
section 12.2). To receive protected result indication, the EAP itself since each
server MUST initiate a channel binding exchange as specified in
Figure 10, message 5. As a result of this channel binding exchange,
the client will receive a protected result indication, because the
server will initiate an informational exchange as part of the
channel binding procedure(messages 7 and 8) through the new IKE-SA
that results from a successful reconnect procedure.
11.1 Channel Binding Procedure in Full Authentication
In the case of EAP-IKEv2 full authentication procedure, the channel
binding procedure is performed in the following way.
The EAP method
has peer MAY include one or more channel binding request in an
IKE_SA_INIT response. The EAP server MAY include one or more channel
binding request in an IKE_AUTH request. When the EAP server receives
an IKE_SA_INIT response with one or more channel binding request, it
MUST include the corresponding channel binding response(s) an
IKE_AUTH request (in addition to implement its own mechanism.
So far channel binding request(s) if
any). When the EAP peer receives an IKE_AUTH request with one or
more channel binding request, it has not been decided which approach should be suggested
for EAP. MUST include the corresponding
channel binding response(s) in an IKE_AUTH response.
When the EAP server successfully validates all the channel binding
response(s) sent by the EAP server, it initiates an INFORMATIONAL
exchange, where an empty payload is used in both INFORMATIONAL
request and INFORMATIONAL response. This exchange serves as a
protected success indication. After completion of this
INFORMATIONAL exchange, the EAP server sends Success message.
11.2 Channel Binding Procedure in Fast Reconnect
In any the case it seems of EAP-IKEv2 fast reconnect, the channel binding
procedure is performed in the following way.
In the pair of CREATE_CHILD_SA exchange, the EAP peer and/or the EAP
server MAY include one or more channel binding request, one for each
channel binding attribute that a generic approach contains some
difficulties since needs validation. When the EAP methods need to negotiate peer
receives a CREATE_CHILD_SA request with containing one or more
channel binding request, it MUST contain channel binding response(s)
in the necessary
parameters CREATE_CHILD_SA response, as well as its channel binding
request(s) if any. This piggybacking is possible because the
CREATE_CHILD_SA exchange is protected with are required the old IKE_SA. When the
EAP server receives a CREATE_CHILD_SA response, if it has one or
more channel binding response to send to build the EAP SA (lifetime,
algorithms, identifiers, etc.). Furthermore, peer, it initiates
an INFORMATIONAL exchange immediately after completion of the
CREATE_CHILD_SA exchange, where one or more channel binding response
is necessary to
cover carried in the INFORMATIONAL request. If the EAP peer
successfully validates the channel binding response(s), it MUST
respond with an empty INFORMATIONAL response. This exchange serves
as a protected success indication. After completion of this
INFORMATIONAL exchange, the EAP server sends Success message.
11.3 Channel Binding Error Indication
A channel binding error cases which happen if is detected by the wrong AAA EAP peer or EAP server
when (i) a channel binding response is selected
(due to failover not contained in the expected
IKEv2 message or load balancing) (ii) a channel binding response is contained in the
expected IKEv2 message but the channel binding attribute does not
have the expected value. Whenever a channel binding error is
detected, the detecting entity MUST send a channel binding error
indication to its peering entity. In case of (ii), the channel
binding error indication MUST contain the channel binding attribute
that caused the error.
When the EAP server detects a channel binding error, a channel
binding error indication MUST be carried in an INFORMATIONAL
request, and the EAP SA is not found.
For both cases peer MUST respond with an empty INFORMATIONAL
response.
When the EAP peer detects a channel binding error, a channel binding
error indication MUST be carried in an IKEv2 error reporting message
for which the R-flag of the IKEv2 header MUST be set. The EAP server
MUST respond with EAP-Failure message when it is necessary receives such a
channel binding error indication.
11.4 Notify Payload Types for Channel Binding
The following Notify Payload types are defined for the purpose of
channel binding exchange.
CALLING_STATION_ID TBD
The payload data in a channel binding response of this type
contains octet string representation of Calling-Station-Id
value known to establish the EAP server by using an external mechanism.
CALLED_STATION_ID TBD
The payload data in a channel binding response of this type
contains octet string representation of Called-Station-Id
value known to keep some state
information. An additional motivation for establishing state is the
ability EAP peer by using an external mechanism.
NAS_PORT_TYPE TBD
The payload data in a channel binding response of this type
contains 4-octet unsigned integer value of NAS-Port-Type
known to provide passive user identity confidentiality as
exercised the EAP peer by using an external mechanism.
The following Notify Payload types are defined for the purpose of
reporting when there is an error in [AH03]. Subsequent protocol exchanges use a pseudonym
instead channel binding exchange.
INVALID_CALLING_STATION_ID TBD
The payload data (if non-null) contains octet string
representation of Calling-Station-Id value that caused the long-term user identity.
Additionally it
error.
INVALID_CALLED_STATION_ID TBD
The payload data (if non-null) contains octet string
representation of Called-Station-Id value that caused the
error.
INVALID_NAS_PORT_TYPE TBD
The payload data (if non-null) contains 4-octet unsigned
integer value of NAS-Port-Type that caused the error.
Table 1 shows the the entity that is necessary allowed to list some requirements send a channel
binding request for
establishing each channel binding attribute type.
channel binding The entity that is allowed to send
attribute type channel binding request
----------------------+------------------------------------------
CALLING_STATION_ID EAP server
CALLED_STATION_ID EAP peer
NAS_PORT_TYPE EAP server
Table 1: Channel Binding Attribute Types and Requesting Entities
11.5 Examples
In the figures of this section, a Notify payload tagged with '*'
indicates a Notify payload with null data (i.e., a channel binding
request). a Notify payload no tagged with '*' indicates a Notify
payload with non-null data (i.e., a channel binding response).
Figure 10 shows an example of EAP-IKEv2 authentication sequence with
a successful channel binding procedure. The first two messages
constitute the standard EAP SA identity exchange and for running are optional.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ,]
N(CALLED_STATION_ID*))
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,], AUTH,
N(CALLED_STATION_ID),
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDr, [CERT,] AUTH,
N(CALLING_STATION_ID),
N(NAS_PORT_TYPE)})
7) A <-- B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {})
8) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {})
9) A <-- B: EAP-Success
Figure 10: EAP-IKEv2 with successful channel binding
Figure 11 shows an example of EAP-IKEv2 authentication sequence when
the EAP server detects an error in a fast resume. For example,
does channel binding procedure. The
first two messages constitute the standard EAP identity exchange and
are optional. In this case, message 7) and 8) MUST constitute an
INFORMATIONAL exchange.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ,]
N(CALLED_STATION_ID*))
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,], AUTH,
N(CALLED_STATION_ID),
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDr, [CERT,] AUTH,
N(CALLING_STATION_ID),
N(NAS_PORT_TYPE)})
7) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(INVALID_CALLING_STATION_ID)})
8) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {})
9) A <-- B: EAP-Failure
Figure 11: EAP-IKEv2 with channel binding error
(detected by EAP server)
Figure 12 shows an example of EAP-IKEv2 authentication sequence when
the EAP peer detects an error in a channel binding procedure. The
first two messages constitute the standard EAP identity exchange and
are optional.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR(A,0), SAi1, KEi, Ni)
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SAr1, KEr, Nr, [CERTREQ,]
N(CALLED_STATION_ID*))
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {IDi, [CERT,] [CERTREQ,] [IDr,], AUTH,
N(CALLED_STATION_ID),
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(INVALID_CALLED_STATION_ID)})
7) A <-- B: EAP-Failure
Figure 12: EAP-IKEv2 with channel binding error
(detected by EAP peer)
Figure 13 shows an example of EAP-IKEv2 fast resume reconnection sequence
with a successful channel binding procedure. The first two messages
constitute the standard EAP identity exchange need to provide key agreement or key
transport functionality?
Once and are optional.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR, SK {SA, Ni, [KEi,]
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(HDR, SK {SA, Nr, [KEr,]
N(CALLED_STATION_ID*),
N(CALLING_STATION_ID),
N(NAS_PORT_TYPE)})
5) A <-- B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(CALLED_STATION_ID)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(HDR(A,B), SK {})
7) A <-- B: EAP-Success
Figure 13: Fast reconnect with channel binding error
(fast reconnect)
Figure 14 shows an example of EAP-IKEv2 fast reconnect sequence when
the above-raised issues have been addressed EAP server detects an error in a channel binding procedure. The
first two messages constitute the standard EAP working
group identity exchange and
are optional. In this case, message 7) and 8) MUST constitute an
INFORMATIONAL exchange.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR, SK {SA, Ni, [KEi,]
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(HDR, SK {SA, Nr, [KEr,]
N(CALLED_STATION_ID*),
N(CALLING_STATION_ID),
N(NAS_PORT_TYPE)})
5) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(INVALID_CALLING_STATION_ID)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {})
7) A <-- B: EAP-Failure
Figure 14: Fast reconnect with channel binding error
(detected by EAP server)
Figure 15 shows an example of EAP-IKEv2 fast reconnect sequence when
the EAP peer detects an error in a solution will be added to EAP-IKEv2.
11. channel binding procedure. The
first two messages constitute the standard EAP identity exchange and
are optional.
1) A <-- B: EAP-Request/Identity
2) A --> B: EAP-Response/Identity(Id)
3) A <-- B: EAP-Request/EAP-Type=EAP-IKEv2(HDR, SK {SA, Ni, [KEi,]
N(CALLING_STATION_ID*),
N(NAS_PORT_TYPE*)})
4) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(HDR, SK {SA, Nr, [KEr,]
N(CALLED_STATION_ID*),
N(CALLING_STATION_ID),
N(NAS_PORT_TYPE)})
5) A <-- B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(CALLED_STATION_ID)})
6) A --> B: EAP-Response/EAP-Type=EAP-IKEv2(
HDR(A,B), SK {N(INVALID_CALLED_STATION_ID)})
7) A <-- B: EAP-Failure
Figure 15: Fast reconnect with channel binding error
(detected by EAP peer)
12. Security Considerations
12.1 General Considerations
The security of the proposed EAP method is intentionally based on
IKEv2 [Kau03]. [Kau04]. Therefore, the security claims of EAP-IKEv2 are
derived from the security offered by the supported features of
IKEv2.
12.2 Security Claims
Authentication mechanism:
Mutual authentication is supported based on either pre-shared
symmetric keys or public/private key pairs. Besides certificates,
plain public keys can be used. It is possible to use different types
of authentication for the different directions within one
authentication exchange. An example is the server using certificate-
based authentication and the client using pre-shared secrets.
Password-based authentication should only be used in IKEv2 with
extended authentication (EAP tunneling), which is not supported by
this version of EAP-IKEv2. Therefore, dictionary attacks are not
applicable in the context of EAP-IKEv2.
Man-in-the-middle attacks discovered in the context of tunneled
authentication protocols (see [AN03] and [PL+03]) are not applicable
to IKEv2 if legacy EAP-IKEv2 as the extended authentication with EAP [RFC2284] feature of IKEv2 is
used. To counter this threat not
supported. Hence, the cryptographic binding claim is not applicable.
Ciphersuite negotiation is supported as specified in IKEv2 provides a compound
authentication for IKE-
SAs. The negotiation for IPsec (Child) SAs is not supported, as such
SAs are not generated by including EAP-IKEv2.
Protected result indication as described in section 7.16 of
[RFC2284bis] is optionally provided by EAP-IKEv2. In message 5 of
figure 1 (full successful authentication) the EAP provided session key inside server
authenticates to the
AUTH payload.
12. Open Issues
The following issues are still under consideration:
- Reducing client. Message 6 authenticates the number of messages
The client to
the server, and the client by authenticating the server and by
sending message flows given 6 expresses that it is willing to accept access. The
client, however, does not get a protected result indication from the
server in this document finish with case. An attacker could potentially forge an EAP-Success
message. EAP
success/failure message which could result in DoS to the client. In
some cases it might situations, synchronization may be possible to skip these messages.
Furthermore it achieved by lower layer
indications.
Protected result indication is optionally provided as specified in
section 11.
If this mechanism is not used, the recommended behavior for the
client is possible to omit assume the first exchange if correct establishment of a new IKE-SA after
sending message 6, independent of the
identity can receipt of an EAP
success/failure. In case of unsuccessful authentication, the server
would answer with an IKEv2 notification (which, in case of the fast
reconnect exchange, would be learned protected by other means. the old IKE-SA). In case
of a lost message 6, the server would retransmit message 5,
indicating the message loss to the client.
The client implementation can minimize potential DoS risks due to
missing protected result indications by assuming the correct
establishment of a new IKE-SA after not receiving one of the above
messages within a certain time window after sending message 6. In
the fast reconnect case, the client needs to hold both the old and
the new IKE-SA in parallel during this time window.
Session independence is optionally provided if the fast reconnect
exchange includes the KE payloads (new Diffie-Hellman) as described
in section 10, Figure 7.
Security claims:
Ciphersuite negotiation: Yes
Mutual authentication: Yes
Integrity protection: Yes
Replay protection: Yes
Confidentiality: Yes
Key derivation: Yes
Key strength: N/A
Dictionary attack prot.: N/A
Fast reconnect: Yes
Crypt. binding: N/A
Protected result ind.: yes
Session independence: yes
Fragmentation: Yes
Channel binding: Yes
13. Open Issues
The following issues are still under consideration:
- Notifications
IKEv2 provides the concept of notifications to exchange messages at
any time (e.g., dead peer detection). It remains for further study
which of these messages are required for this EAP method.
- Roles of initiator and responder
Figure 4 shows supported identities
Can the initiator starting NAI be carried by the EAP-IKEv2 exchange.
However, RFC822 ID type of IKEv2? Are there
other formats to be supported? Additional profiling may be required
in section 5.
- protected result indication
This method provides protected result indications as an optional
feature by running a channel binding exchange. It is also the possibility ffs whether
this feature should be mandated for all message flows (which would
require to add an additional informational exchange to have the message
flows without channel binding.
- tunneled method
To reduce the method's complexity, EAP server tunneling through EAP-IKEv2
that is in principal possible with IKEv2 is not supported. If
tunneling support is, however, required (e.g. for sequencing), it is
possible to
start develop an EAP-IKEv2-tunneled method from the exchange which saves roundtrips. present
one. The major change would be to reverse the roles of IKEv2
initiator and responder, as the initiator is EAP-authenticated in
the tunneled case.
It remains for further
study is not considered a good approach by the authors to analyze have both the resulting security properties.
13.
tunneled and the non-tunneled method in a single specification, as
this would result in a rather complex method description. The
tunneled-method EAP-IKEv2 specification, if required, will therefore
come with a separate document.
14. Normative References
[RFC2284] L. Blunk and J. Vollbrecht: "PPP Extensible Authentication
Protocol (EAP)", RFC 2284, March 1998.
[Kau03]
[Kau04] C. Kaufman: "Internet Key Exchange (IKEv2) Protocol",
internet draft, Internet Engineering Task Force, October 2003. January 2004. Work
in progress.
[RFC2119] S. Bradner: "Key words for use in RFCs to Indicate
Requirement Levels", RFC 2119, Internet Engineering Task Force,
March 1997.
14.
15. Informative References
[AN03] N. Asokan, V. Niemi, and K. Nyberg: "Man-in-the-middle in
tunnelled authentication", In the Proceedings of the 11th
International Workshop on Security Protocols, Cambridge, UK, April
2003. To be published in the Springer-Verlag LNCS series.
[PL+03] J. Puthenkulam, V. Lortz, A. Palekar, D. Simon, and B.
Aboba, "The compound authentication binding problem," internet
draft, Internet Engineering Task Force, 2003. Work in progress.
[RFC2409] Harkins, D., Carrel, D., "The Internet Key Exchange
(IKE)", RFC 2409, November 1998.
[Per03] R. Perlman: "Understanding IKEv2: Tutorial, and rationale
for decisions", internet draft, Internet Engineering Task Force,
2003. Work in progress.
[AS+03] B. Aboba, D. Simon and J. Arkko: " EAP Key Management
Framework", internet draft, Internet Engineering Task Force,
October, 2003. Work in progress.
[HS03] H. Haverinen, J. Salowey: "EAP SIM Authentication", internet
draft, Internet Engineering Task Force, 2003. Work in progress.
[PS+03] A. Palekar, D. Simon, G. Zorn and S. Josefsson: "Protected
EAP Protocol (PEAP)", internet draft, Internet Engineering Task
Force, March 2003. Work in progress.
[AH03] J. Arkko and H. Haverinen: "EAP AKA Authentication", internet
draft, Internet Engineering Task Force, June 2003. Work in
progress.
Acknowledgments
We would like to thank Bernard Aboba, Jari Arkko, Guenther Horn,
Paoulo Pagliusi and John Vollbrecht for their comments to this
draft.
Additionally we would like to thank members of the PANA design team
(namely D. Forsberg and A. Yegin) for their comments and input to
the initial version of the draft.
Finally we would like to thank the members of the EAP keying design
team for their discussion in the area of the EAP Key Management
Framework.
Author's Addresses
Hannes Tschofenig
Siemens AG
Otto-Hahn-Ring 6
81739 Munich
Germany
EMail: Hannes.Tschofenig@siemens.com
Dirk Kroeselberg
Siemens AG
Otto-Hahn-Ring 6
81739 Munich
Germany
EMail: Dirk.Kroeselberg@siemens.com
Yoshihiro Ohba
Toshiba America Research, Information Systems, Inc.
P.O. Box 136
Convent Station, NJ, 07961-0136
9740 Irvine Blvd.
Irvine, CA 92619-1697
USA
Phone: +1 973 829 5174
Email:
EMail: yohba@tari.toshiba.com
Full Copyright Statement
Copyright (C) The Internet Society (2003). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph
are included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assignees.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.