RFC 4851
The Flexible Authentication via Secure Tunneling Extensible Authentication Protocol Method (EAP-FAST)
Network Working Group N. Cam-Winget Request for Comments: 4851 D. McGrew Category: Informational J. Salowey H. Zhou Cisco Systems May 2007 The Flexible Authentication via Secure Tunneling Extensible Authentication Protocol Method (EAP-FAST) Status of This Memo This memo provides information for the Internet community. It does not specify an Internet standard of any kind. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The IETF Trust (2007).Abstract
This document defines the Extensible Authentication Protocol (EAP) based Flexible Authentication via Secure Tunneling (EAP-FAST) protocol. EAP-FAST is an EAP method that enables secure communication between a peer and a server by using the Transport Layer Security (TLS) to establish a mutually authenticated tunnel. Within the tunnel, Type-Length-Value (TLV) objects are used to convey authentication related data between the peer and the EAP server.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1. Specification Requirements . . . . . . . . . . . . . . . . 5 1.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 2. Protocol Overview . . . . . . . . . . . . . . . . . . . . . . 6 2.1. Architectural Model . . . . . . . . . . . . . . . . . . . 6 2.2. Protocol Layering Model . . . . . . . . . . . . . . . . . 7 3. EAP-FAST Protocol . . . . . . . . . . . . . . . . . . . . . . 8 3.1. Version Negotiation . . . . . . . . . . . . . . . . . . . 8 3.2. EAP-FAST Authentication Phase 1: Tunnel Establishment . . 9 3.2.1. TLS Session Resume Using Server State . . . . . . . . 10 3.2.2. TLS Session Resume Using a PAC . . . . . . . . . . . . 10 3.2.3. Transition between Abbreviated and Full TLS Handshake . . . . . . . . . . . . . . . . . . . . . . 12 3.3. EAP-FAST Authentication Phase 2: Tunneled Authentication . . . . . . . . . . . . . . . . . . . . . . 12 3.3.1. EAP Sequences . . . . . . . . . . . . . . . . . . . . 13 3.3.2. Protected Termination and Acknowledged Result Indication . . . . . . . . . . . . . . . . . . . . . . 13 3.4. Determining Peer-Id and Server-Id . . . . . . . . . . . . 14 3.5. EAP-FAST Session Identifier . . . . . . . . . . . . . . . 15 3.6. Error Handling . . . . . . . . . . . . . . . . . . . . . . 15 3.6.1. TLS Layer Errors . . . . . . . . . . . . . . . . . . . 15 3.6.2. Phase 2 Errors . . . . . . . . . . . . . . . . . . . . 16 3.7. Fragmentation . . . . . . . . . . . . . . . . . . . . . . 16 4. Message Formats . . . . . . . . . . . . . . . . . . . . . . . 18 4.1. EAP-FAST Message Format . . . . . . . . . . . . . . . . . 18 4.1.1. Authority ID Data . . . . . . . . . . . . . . . . . . 20 4.2. EAP-FAST TLV Format and Support . . . . . . . . . . . . . 20 4.2.1. General TLV Format . . . . . . . . . . . . . . . . . . 21 4.2.2. Result TLV . . . . . . . . . . . . . . . . . . . . . . 22 4.2.3. NAK TLV . . . . . . . . . . . . . . . . . . . . . . . 23 4.2.4. Error TLV . . . . . . . . . . . . . . . . . . . . . . 24 4.2.5. Vendor-Specific TLV . . . . . . . . . . . . . . . . . 25 4.2.6. EAP-Payload TLV . . . . . . . . . . . . . . . . . . . 26 4.2.7. Intermediate-Result TLV . . . . . . . . . . . . . . . 28 4.2.8. Crypto-Binding TLV . . . . . . . . . . . . . . . . . . 29 4.2.9. Request-Action TLV . . . . . . . . . . . . . . . . . . 31 4.3. Table of TLVs . . . . . . . . . . . . . . . . . . . . . . 32 5. Cryptographic Calculations . . . . . . . . . . . . . . . . . . 32 5.1. EAP-FAST Authentication Phase 1: Key Derivations . . . . . 32 5.2. Intermediate Compound Key Derivations . . . . . . . . . . 33 5.3. Computing the Compound MAC . . . . . . . . . . . . . . . . 34 5.4. EAP Master Session Key Generation . . . . . . . . . . . . 35 5.5. T-PRF . . . . . . . . . . . . . . . . . . . . . . . . . . 35 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 36
7. Security Considerations . . . . . . . . . . . . . . . . . . . 37 7.1. Mutual Authentication and Integrity Protection . . . . . . 37 7.2. Method Negotiation . . . . . . . . . . . . . . . . . . . . 38 7.3. Separation of Phase 1 and Phase 2 Servers . . . . . . . . 38 7.4. Mitigation of Known Vulnerabilities and Protocol Deficiencies . . . . . . . . . . . . . . . . . . . . . . . 39 7.4.1. User Identity Protection and Verification . . . . . . 39 7.4.2. Dictionary Attack Resistance . . . . . . . . . . . . . 40 7.4.3. Protection against Man-in-the-Middle Attacks . . . . . 40 7.4.4. PAC Binding to User Identity . . . . . . . . . . . . . 41 7.5. Protecting against Forged Clear Text EAP Packets . . . . . 41 7.6. Server Certificate Validation . . . . . . . . . . . . . . 42 7.7. Tunnel PAC Considerations . . . . . . . . . . . . . . . . 42 7.8. Security Claims . . . . . . . . . . . . . . . . . . . . . 43 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 44 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 44 9.1. Normative References . . . . . . . . . . . . . . . . . . . 44 9.2. Informative References . . . . . . . . . . . . . . . . . . 45 Appendix A. Examples . . . . . . . . . . . . . . . . . . . . . . 46 A.1. Successful Authentication . . . . . . . . . . . . . . . . 46 A.2. Failed Authentication . . . . . . . . . . . . . . . . . . 47 A.3. Full TLS Handshake using Certificate-based Ciphersuite . . 48 A.4. Client Authentication during Phase 1 with Identity Privacy . . . . . . . . . . . . . . . . . . . . . . . . . 50 A.5. Fragmentation and Reassembly . . . . . . . . . . . . . . . 52 A.6. Sequence of EAP Methods . . . . . . . . . . . . . . . . . 53 A.7. Failed Crypto-Binding . . . . . . . . . . . . . . . . . . 56 A.8. Sequence of EAP Method with Vendor-Specific TLV Exchange . . . . . . . . . . . . . . . . . . . . . . . . . 57 Appendix B. Test Vectors . . . . . . . . . . . . . . . . . . . . 60 B.1. Key Derivation . . . . . . . . . . . . . . . . . . . . . . 60 B.2. Crypto-Binding MIC . . . . . . . . . . . . . . . . . . . . 62
1. Introduction
Network access solutions requiring user friendly and easily deployable secure authentication mechanisms highlight the need for strong mutual authentication protocols that enable the use of weaker user credentials. This document defines an Extensible Authentication Protocol (EAP), which consists of establishing a Transport Layer Security (TLS) tunnel using TLS 1.0 [RFC2246], TLS 1.1 [RFC4346], or a successor version of TLS, using the latest version supported by both parties. Once the tunnel is established, the protocol further exchanges data in the form of type, length, and value objects (TLV) to perform further authentication. EAP-FAST supports the TLS extension defined in [RFC4507] to support fast re-establishment of the secure tunnel without having to maintain per-session state on the server. [EAP-PROV] defines EAP-FAST-based mechanisms to provision the credential for this extension which is called a Protected Access Credential (PAC). EAP-FAST's design motivations included: o Mutual authentication: an EAP server must be able to verify the identity and authenticity of the peer, and the peer must be able to verify the authenticity of the EAP server. o Immunity to passive dictionary attacks: many authentication protocols require a password to be explicitly provided (either as cleartext or hashed) by the peer to the EAP server; at minimum, the communication of the weak credential (e.g., password) must be immune from eavesdropping. o Immunity to man-in-the-middle (MitM) attacks: in establishing a mutually authenticated protected tunnel, the protocol must prevent adversaries from successfully interjecting information into the conversation between the peer and the EAP server. o Flexibility to enable support for most password authentication interfaces: as many different password interfaces (e.g., Microsoft Challenge Handshake Authentication Protocol (MS-CHAP), Lightweight Directory Access Protocol (LDAP), One-Time Password (OTP), etc.) exist to authenticate a peer, the protocol must provide this support seamlessly. o Efficiency: specifically when using wireless media, peers will be limited in computational and power resources. The protocol must enable the network access communication to be computationally lightweight.
With these motivational goals defined, further secondary design
criteria are imposed:
o Flexibility to extend the communications inside the tunnel: with
the growing complexity in network infrastructures, the need to
gain authentication, authorization, and accounting is also
evolving. For instance, there may be instances in which multiple
existing authentication protocols are required to achieve mutual
authentication. Similarly, different protected conversations may
be required to achieve the proper authorization once a peer has
successfully authenticated.
o Minimize the authentication server's per user authentication state
requirements: with large deployments, it is typical to have many
servers acting as the authentication servers for many peers. It
is also highly desirable for a peer to use the same shared secret
to secure a tunnel much the same way it uses the username and
password to gain access to the network. The protocol must
facilitate the use of a single strong shared secret by the peer
while enabling the servers to minimize the per user and device
state it must cache and manage.
1.1. Specification Requirements
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] .
1.2. Terminology
Much of the terminology in this document comes from [RFC3748].
Additional terms are defined below:
Protected Access Credential (PAC)
Credentials distributed to a peer for future optimized network
authentication. The PAC consists of, at most, three components: a
shared secret, an opaque element, and optionally other
information. The shared secret component contains the pre-shared
key between the peer and the authentication server. The opaque
part is provided to the peer and is presented to the
authentication server when the peer wishes to obtain access to
network resources. Finally, a PAC may optionally include other
information that may be useful to the peer. The opaque part of
the PAC is the same type of data as the ticket in [RFC4507] and
the shared secret is used to derive the TLS master secret.
2. Protocol Overview
EAP-FAST is an authentication protocol similar to EAP-TLS [RFC2716] that enables mutual authentication and cryptographic context establishment by using the TLS handshake protocol. EAP-FAST allows for the established TLS tunnel to be used for further authentication exchanges. EAP-FAST makes use of TLVs to carry out the inner authentication exchanges. The tunnel is then used to protect weaker inner authentication methods, which may be based on passwords, and to communicate the results of the authentication. EAP-FAST makes use of the TLS enhancements in [RFC4507] to enable an optimized TLS tunnel session resume while minimizing server state. The secret key used in EAP-FAST is referred to as the Protected Access Credential key (or PAC-Key); the PAC-Key is used to mutually authenticate the peer and the server when securing a tunnel. The ticket is referred to as the Protected Access Credential opaque data (or PAC-Opaque). The secret key and ticket used to establish the tunnel may be provisioned through mechanisms that do not involve the TLS handshake. It is RECOMMENDED that implementations support the capability to distribute the ticket and secret key within the EAP- FAST tunnel as specified in [EAP-PROV]. The EAP-FAST conversation is used to establish or resume an existing session to typically establish network connectivity between a peer and the network. Upon successful execution of EAP-FAST, both EAP peer and EAP server derive strong session key material that can then be communicated to the network access server (NAS) for use in establishing a link layer security association.2.1. Architectural Model
The network architectural model for EAP-FAST usage is shown below: +----------+ +----------+ +----------+ +----------+ | | | | | | | Inner | | Peer |<---->| Authen- |<---->| EAP-FAST |<---->| Method | | | | ticator | | server | | server | | | | | | | | | +----------+ +----------+ +----------+ +----------+ EAP-FAST Architectural Model The entities depicted above are logical entities and may or may not correspond to separate network components. For example, the EAP-FAST server and inner method server might be a single entity; the authenticator and EAP-FAST server might be a single entity; or the functions of the authenticator, EAP-FAST server, and inner method
server might be combined into a single physical device. For example, typical 802.11 deployments place the Authenticator in an access point (AP) while a Radius server may provide the EAP-FAST and inner method server components. The above diagram illustrates the division of labor among entities in a general manner and shows how a distributed system might be constructed; however, actual systems might be realized more simply. The security considerations Section 7.3 provides an additional discussion of the implications of separating the EAP-FAST server from the inner method server.2.2. Protocol Layering Model
EAP-FAST packets are encapsulated within EAP; EAP in turn requires a carrier protocol for transport. EAP-FAST packets encapsulate TLS, which is then used to encapsulate user authentication information. Thus, EAP-FAST messaging can be described using a layered model, where each layer encapsulates the layer above it. The following diagram clarifies the relationship between protocols: +---------------------------------------------------------------+ | Inner EAP Method | Other TLV information | |---------------------------------------------------------------| | TLV Encapsulation (TLVs) | |---------------------------------------------------------------| | TLS | |---------------------------------------------------------------| | EAP-FAST | |---------------------------------------------------------------| | EAP | |---------------------------------------------------------------| | Carrier Protocol (EAP over LAN, RADIUS, Diameter, etc.) | +---------------------------------------------------------------+ Protocol Layering Model The TLV layer is a payload with Type-Length-Value (TLV) Objects defined in Section 4.2. The TLV objects are used to carry arbitrary parameters between an EAP peer and an EAP server. All conversations in the EAP-FAST protected tunnel must be encapsulated in a TLV layer. Methods for encapsulating EAP within carrier protocols are already defined. For example, IEEE 802.1X [IEEE.802-1X.2004] may be used to transport EAP between the peer and the authenticator; RADIUS [RFC3579] or Diameter [RFC4072] may be used to transport EAP between the authenticator and the EAP-FAST server.
3. EAP-FAST Protocol
EAP-FAST authentication occurs in two phases. In the first phase, EAP-FAST employs the TLS handshake to provide an authenticated key exchange and to establish a protected tunnel. Once the tunnel is established the second phase begins with the peer and server engaging in further conversations to establish the required authentication and authorization policies. The operation of the protocol, including Phase 1 and Phase 2, are the topic of this section. The format of EAP-FAST messages is given in Section 4 and the cryptographic calculations are given in Section 5.3.1. Version Negotiation
EAP-FAST packets contain a 3-bit version field, following the TLS Flags field, which enables EAP-FAST implementations to be backward compatible with previous versions of the protocol. This specification documents the EAP-FAST version 1 protocol; implementations of this specification MUST use a version field set to 1. Version negotiation proceeds as follows: In the first EAP-Request sent with EAP type=EAP-FAST, the EAP server must set the version field to the highest supported version number. If the EAP peer supports this version of the protocol, it MUST respond with an EAP-Response of EAP type=EAP-FAST, and the version number proposed by the EAP-FAST server. If the EAP-FAST peer does not support this version, it responds with an EAP-Response of EAP type=EAP-FAST and the highest supported version number. If the EAP-FAST server does not support the version number proposed by the EAP-FAST peer, it terminates the conversation. Otherwise the EAP-FAST conversation continues. The version negotiation procedure guarantees that the EAP-FAST peer and server will agree to the latest version supported by both parties. If version negotiation fails, then use of EAP-FAST will not be possible, and another mutually acceptable EAP method will need to be negotiated if authentication is to proceed. The EAP-FAST version is not protected by TLS; and hence can be modified in transit. In order to detect a modification of the EAP- FAST version, the peers MUST exchange the EAP-FAST version number
received during version negotiation using the Crypto-Binding TLV described in Section 4.2.8. The receiver of the Crypto-Binding TLV MUST verify that the version received in the Crypto-Binding TLV matches the version sent by the receiver in the EAP-FAST version negotiation.3.2. EAP-FAST Authentication Phase 1: Tunnel Establishment
EAP-FAST is based on the TLS handshake [RFC2246] to establish an authenticated and protected tunnel. The TLS version offered by the peer and server MUST be TLS v1.0 or later. This version of the EAP- FAST implementation MUST support the following TLS ciphersuites: TLS_RSA_WITH_RC4_128_SHA TLS_RSA_WITH_AES_128_CBC_SHA [RFC3268] TLS_DHE_RSA_WITH_AES_128_CBC_SHA [RFC3268] Other ciphersuites MAY be supported. It is RECOMMENDED that anonymous ciphersuites such as TLS_DH_anon_WITH_AES_128_CBC_SHA only be used in the context of the provisioning described in [EAP-PROV]. Care must be taken to address potential man-in-the-middle attacks when ciphersuites that do not provide authenticated tunnel establishment are used. During the EAP-FAST Phase 1 conversation the EAP-FAST endpoints MAY negotiate TLS compression. The EAP server initiates the EAP-FAST conversation with an EAP request containing an EAP-FAST/Start packet. This packet includes a set Start (S) bit, the EAP-FAST version as specified in Section 3.1, and an authority identity. The TLS payload in the initial packet is empty. The authority identity (A-ID) is used to provide the peer a hint of the server's identity that may be useful in helping the peer select the appropriate credential to use. Assuming that the peer supports EAP-FAST the conversation continues with the peer sending an EAP-Response packet with EAP type of EAP-FAST with the Start (S) bit clear and the version as specified in Section 3.1. This message encapsulates one or more TLS records containing the TLS handshake messages. If the EAP-FAST version negotiation is successful then the EAP-FAST conversation continues until the EAP server and EAP peer are ready to enter Phase 2. When the full TLS handshake is performed, then the first payload of EAP-FAST Phase 2 MAY be sent along with server-finished handshake message to reduce the number of round trips. After the TLS session is established, another EAP exchange MAY occur within the tunnel to authenticate the EAP peer. EAP-FAST implementations MUST support client authentication during tunnel
establishment using the TLS ciphersuites specified in Section 3.2. EAP-FAST implementations SHOULD also support the immediate renegotiation of a TLS session to initiate a new handshake message exchange under the protection of the current ciphersuite. This allows support for protection of the peer's identity. Note that the EAP peer does not need to authenticate as part of the TLS exchange, but can alternatively be authenticated through additional EAP exchanges carried out in Phase 2. The EAP-FAST tunnel protects peer identity information from disclosure outside the tunnel. Implementations that wish to provide identity privacy for the peer identity must carefully consider what information is disclosed outside the tunnel. The following sections describe resuming a TLS session based on server-side or client-side state.3.2.1. TLS Session Resume Using Server State
EAP-FAST session resumption is achieved in the same manner TLS achieves session resume. To support session resumption, the server and peer must minimally cache the SessionID, master secret, and ciphersuite. The peer attempts to resume a session by including a valid SessionID from a previous handshake in its ClientHello message. If the server finds a match for the SessionID and is willing to establish a new connection using the specified session state, the server will respond with the same SessionID and proceed with the EAP- FAST Authentication Phase 1 tunnel establishment based on a TLS abbreviated handshake. After a successful conclusion of the EAP-FAST Authentication Phase 1 conversation, the conversation then continues on to Phase 2.3.2.2. TLS Session Resume Using a PAC
EAP-FAST supports the resumption of sessions based on client-side state using techniques described in [RFC4507]. This version of EAP- FAST does not support the provisioning of a ticket through the use of the SessionTicket handshake message. Instead it supports the provisioning of a ticket called a Protected Access Credential (PAC) as described in [EAP-PROV]. Implementations may provide additional ways to provision the PAC, such as manual configuration. Since the PAC mentioned here is used for establishing the TLS Tunnel, it is more specifically referred to as the Tunnel PAC. The Tunnel PAC is a security credential provided by the EAP server to a peer and comprised of:
1. PAC-Key: this is a 32-octet key used by the peer to establish the
EAP-FAST Phase 1 tunnel. This key is used to derive the TLS
premaster secret as described in Section 5.1. The PAC-Key is
randomly generated by the EAP server to produce a strong entropy
32-octet key. The PAC-Key is a secret and MUST be treated
accordingly. For example, as the PAC-Key is a separate component
provisioned by the server to establish a secure tunnel, the
server may deliver this component protected by a secure channel,
and it must be stored securely by the peer.
2. PAC-Opaque: this is a variable length field that is sent to the
EAP server during the EAP-FAST Phase 1 tunnel establishment. The
PAC-Opaque can only be interpreted by the EAP server to recover
the required information for the server to validate the peer's
identity and authentication. For example, the PAC-Opaque
includes the PAC-Key and may contain the PAC's peer identity.
The PAC-Opaque format and contents are specific to the PAC
issuing server. The PAC-Opaque may be presented in the clear, so
an attacker MUST NOT be able to gain useful information from the
PAC-Opaque itself. The server issuing the PAC-Opaque must ensure
it is protected with strong cryptographic keys and algorithms.
3. PAC-Info: this is a variable length field used to provide, at a
minimum, the authority identity of the PAC issuer. Other useful
but not mandatory information, such as the PAC-Key lifetime, may
also be conveyed by the PAC issuing server to the peer during PAC
provisioning or refreshment.
The use of the PAC is based on the SessionTicket extension defined in
[RFC4507]. The EAP server initiates the EAP-FAST conversation as
normal. Upon receiving the A-ID from the server, the peer checks to
see if it has an existing valid PAC-Key and PAC-Opaque for the
server. If it does, then it obtains the PAC-Opaque and puts it in
the SessionTicket extension in the ClientHello. It is RECOMMENDED in
EAP-FAST that the peer include an empty Session ID in a ClientHello
containing a PAC-Opaque. EAP-FAST does not currently support the
SessionTicket Handshake message so an empty SessionTicket extension
MUST NOT be included in the ClientHello. If the PAC-Opaque included
in the SessionTicket extension is valid and the EAP server permits
the abbreviated TLS handshake, it will select the ciphersuite allowed
to be used from information within the PAC and finish with the
abbreviated TLS handshake. If the server receives a Session ID and a
PAC-Opaque in the SessionTicket extension in a ClientHello, it should
place the same Session ID in the ServerHello if it is resuming a
session based on the PAC-Opaque. The conversation then proceeds as
described in [RFC4507] until the handshake completes or a fatal error
occurs. After the abbreviated handshake completes, the peer and
server are ready to commence Phase 2. Note that when a PAC is used,
the TLS master secret is calculated from the PAC-Key, client random, and server random as described in Section 5.1. Specific details for the Tunnel PAC format, provisioning and security considerations are best described in [EAP-PROV]3.2.3. Transition between Abbreviated and Full TLS Handshake
If session resumption based on server-side or client-side state fails, the server can gracefully fall back to a full TLS handshake. If the ServerHello received by the peer contains a empty Session ID or a Session ID that is different than in the ClientHello, the server may be falling back to a full handshake. The peer can distinguish the server's intent of negotiating full or abbreviated TLS handshake by checking the next TLS handshake messages in the server response to the ClientHello. If ChangeCipherSpec follows the ServerHello in response to the ClientHello, then the server has accepted the session resumption and intends to negotiate the abbreviated handshake. Otherwise, the server intends to negotiate the full TLS handshake. A peer can request for a new PAC to be provisioned after the full TLS handshake and mutual authentication of the peer and the server. In order to facilitate the fallback to a full handshake, the peer SHOULD include ciphersuites that allow for a full handshake and possibly PAC provisioning so the server can select one of these in case session resumption fails. An example of the transition is shown in Appendix A.3.3. EAP-FAST Authentication Phase 2: Tunneled Authentication
The second portion of the EAP-FAST Authentication occurs immediately after successful completion of Phase 1. Phase 2 occurs even if both peer and authenticator are authenticated in the Phase 1 TLS negotiation. Phase 2 MUST NOT occur if the Phase 1 TLS handshake fails. Phase 2 consists of a series of requests and responses encapsulated in TLV objects defined in Section 4.2. Phase 2 MUST always end with a protected termination exchange described in Section 3.3.2. The TLV exchange may include the execution of zero or more EAP methods within the protected tunnel as described in Section 3.3.1. A server MAY proceed directly to the protected termination exchange if it does not wish to request further authentication from the peer. However, the peer and server must not assume that either will skip inner EAP methods or other TLV exchanges. The peer may have roamed to a network that requires conformance with a different authentication policy or the peer may request the server take additional action through the use of the Request-Action TLV.
3.3.1. EAP Sequences
EAP [RFC3748] prohibits use of multiple authentication methods within a single EAP conversation in order to limit vulnerabilities to man- in-the-middle attacks. EAP-FAST addresses man-in-the-middle attacks through support for cryptographic protection of the inner EAP exchange and cryptographic binding of the inner authentication method(s) to the protected tunnel. EAP methods are executed serially in a sequence. This version of EAP-FAST does not support initiating multiple EAP methods simultaneously in parallel. The methods need not be distinct. For example, EAP-TLS could be run twice as an inner method, first using machine credentials followed by a second instance using user credentials. EAP method messages are carried within EAP-Payload TLVs defined in Section 4.2.6. If more than one method is going to be executed in the tunnel then, upon completion of a method, a server MUST send an Intermediate-Result TLV indicating the result. The peer MUST respond to the Intermediate-Result TLV indicating its result. If the result indicates success, the Intermediate-Result TLV MUST be accompanied by a Crypto-Binding TLV. The Crypto-Binding TLV is further discussed in Section 4.2.8 and Section 5.3. The Intermediate-Result TLVs can be included with other TLVs such as EAP-Payload TLVs starting a new EAP conversation or with the Result TLV used in the protected termination exchange. In the case where only one EAP method is executed in the tunnel, the Intermediate-Result TLV MUST NOT be sent with the Result TLV. In this case, the status of the inner EAP method is represented by the final Result TLV, which also represents the result of the whole EAP-FAST conversation. This is to maintain backward compatibility with existing implementations. If both peer and server indicate success, then the method is considered complete. If either indicates failure. then the method is considered failed. The result of failure of an EAP method does not always imply a failure of the overall authentication. If one authentication method fails, the server may attempt to authenticate the peer with a different method.3.3.2. Protected Termination and Acknowledged Result Indication
A successful EAP-FAST Phase 2 conversation MUST always end in a successful Result TLV exchange. An EAP-FAST server may initiate the Result TLV exchange without initiating any EAP conversation in EAP- FAST Phase 2. After the final Result TLV exchange, the TLS tunnel is terminated and a clear text EAP-Success or EAP-Failure is sent by the server. The format of the Result TLV is described in Section 4.2.2.
A server initiates a successful protected termination exchange by sending a Result TLV indicating success. The server may send the Result TLV along with an Intermediate-Result TLV and a Crypto-Binding TLV. If the peer requires nothing more from the server it will respond with a Result TLV indicating success accompanied by an Intermediate-Result TLV and Crypto-Binding TLV if necessary. The server then tears down the tunnel and sends a clear text EAP-Success. If the peer receives a Result TLV indicating success from the server, but its authentication policies are not satisfied (for example it requires a particular authentication mechanism be run or it wants to request a PAC), it may request further action from the server using the Request-Action TLV. The Request-Action TLV is sent along with the Result TLV indicating what EAP Success/Failure result the peer would expect if the requested action is not granted. The value of the Request-Action TLV indicates what the peer would like to do next. The format and values for the Request-Action TLV are defined in Section 4.2.9. Upon receiving the Request-Action TLV the server may process the request or ignore it, based on its policy. If the server ignores the request, it proceeds with termination of the tunnel and send the clear text EAP Success or Failure message based on the value of the peer's result TLV. If the server honors and processes the request, it continues with the requested action. The conversation completes with a Result TLV exchange. The Result TLV may be included with the TLV that completes the requested action. Error handling for Phase 2 is discussed in Section 3.6.2.3.4. Determining Peer-Id and Server-Id
The Peer-Id and Server-Id may be determined based on the types of credentials used during either the EAP-FAST tunnel creation or authentication. When X.509 certificates are used for peer authentication, the Peer-Id is determined by the subject or subjectAltName fields in the peer certificate. As noted in [RFC3280] (updated by [RFC4630]): The subject field identifies the entity associated with the public key stored in the subject public key field. The subject name MAY be carried in the subject field and/or the subjectAltName extension.... If subject naming information is present only in the subjectAltName extension (e.g., a key bound only to an email address or URI), then the subject name MUST be an empty sequence and the subjectAltName extension MUST be critical.
Where it is non-empty, the subject field MUST contain an X.500
distinguished name (DN).
If an inner EAP method is run, then the Peer-Id is obtained from the
inner method.
When the server uses an X.509 certificate to establish the TLS
tunnel, the Server-Id is determined in a similar fashion as stated
above for the Peer-Id; e.g., the subject or subjectAltName field in
the server certificate defines the Server-Id.
3.5. EAP-FAST Session Identifier
The EAP session identifier is constructed using the random values
provided by the peer and server during the TLS tunnel establishment.
The Session-Id is defined as follows:
Session-Id = 0x2B || client_random || server_random)
client_random = 32 byte nonce generated by the peer
server_random = 32 byte nonce generated by the server
3.6. Error Handling
EAP-FAST uses the following error handling rules summarized below:
1. Errors in the TLS layer are communicated via TLS alert messages
in all phases of EAP-FAST.
2. The Intermediate-Result TLVs carry success or failure indications
of the individual EAP methods in EAP-FAST Phase 2. Errors within
the EAP conversation in Phase 2 are expected to be handled by
individual EAP methods.
3. Violations of the TLV rules are handled using Result TLVs
together with Error TLVs.
4. Tunnel compromised errors (errors caused by Crypto-Binding failed
or missing) are handled using Result TLVs and Error TLVs.
3.6.1. TLS Layer Errors
If the EAP-FAST server detects an error at any point in the TLS
Handshake or the TLS layer, the server SHOULD send an EAP-FAST
request encapsulating a TLS record containing the appropriate TLS
alert message rather than immediately terminating the conversation so
as to allow the peer to inform the user of the cause of the failure
and possibly allow for a restart of the conversation. The peer MUST
send an EAP-FAST response to an alert message. The EAP-Response
packet sent by the peer may encapsulate a TLS ClientHello handshake message, in which case the EAP-FAST server MAY allow the EAP-FAST conversation to be restarted, or it MAY contain an EAP-FAST response with a zero-length message, in which case the server MUST terminate the conversation with an EAP-Failure packet. It is up to the EAP- FAST server whether to allow restarts, and if so, how many times the conversation can be restarted. An EAP-FAST Server implementing restart capability SHOULD impose a limit on the number of restarts, so as to protect against denial-of-service attacks. If the EAP-FAST peer detects an error at any point in the TLS layer, the EAP-FAST peer should send an EAP-FAST response encapsulating a TLS record containing the appropriate TLS alert message. The server may restart the conversation by sending an EAP-FAST request packet encapsulating the TLS HelloRequest handshake message. The peer may allow the EAP-FAST conversation to be restarted or it may terminate the conversation by sending an EAP-FAST response with an zero-length message.3.6.2. Phase 2 Errors
Any time the peer or the server finds a fatal error outside of the TLS layer during Phase 2 TLV processing, it MUST send a Result TLV of failure and an Error TLV with the appropriate error code. For errors involving the processing of the sequence of exchanges, such as a violation of TLV rules (e.g., multiple EAP-Payload TLVs), the error code is Unexpected_TLVs_Exchanged. For errors involving a tunnel compromise, the error-code is Tunnel_Compromise_Error. Upon sending a Result TLV with a fatal Error TLV the sender terminates the TLS tunnel. Note that a server will still wait for a message from the peer after it sends a failure, however the server does not need to process the contents of the response message. If a server receives a Result TLV of failure with a fatal Error TLV, it SHOULD send a clear text EAP-Failure. If a peer receives a Result TLV of failure, it MUST respond with a Result TLV indicating failure. If the server has sent a Result TLV of failure, it ignores the peer response, and it SHOULD send a clear text EAP-Failure.3.7. Fragmentation
A single TLS record may be up to 16384 octets in length, but a TLS message may span multiple TLS records, and a TLS certificate message may in principle be as long as 16 MB. This is larger than the maximum size for a message on most media types, therefore it is desirable to support fragmentation. Note that in order to protect against reassembly lockup and denial-of-service attacks, it may be desirable for an implementation to set a maximum size for one such
group of TLS messages. Since a typical certificate chain is rarely longer than a few thousand octets, and no other field is likely to be anywhere near as long, a reasonable choice of maximum acceptable message length might be 64 KB. This is still a fairly large message packet size so an EAP-FAST implementation MUST provide its own support for fragmentation and reassembly. Since EAP is an lock-step protocol, fragmentation support can be added in a simple manner. In EAP, fragments that are lost or damaged in transit will be retransmitted, and since sequencing information is provided by the Identifier field in EAP, there is no need for a fragment offset field. EAP-FAST fragmentation support is provided through the addition of flag bits within the EAP-Response and EAP-Request packets, as well as a TLS Message Length field of four octets. Flags include the Length included (L), More fragments (M), and EAP-FAST Start (S) bits. The L flag is set to indicate the presence of the four-octet TLS Message Length field, and MUST be set for the first fragment of a fragmented TLS message or set of messages. The M flag is set on all but the last fragment. The S flag is set only within the EAP-FAST start message sent from the EAP server to the peer. The TLS Message Length field is four octets, and provides the total length of the TLS message or set of messages that is being fragmented; this simplifies buffer allocation. When an EAP-FAST peer receives an EAP-Request packet with the M bit set, it MUST respond with an EAP-Response with EAP-Type of EAP-FAST and no data. This serves as a fragment ACK. The EAP server must wait until it receives the EAP-Response before sending another fragment. In order to prevent errors in processing of fragments, the EAP server MUST increment the Identifier field for each fragment contained within an EAP-Request, and the peer must include this Identifier value in the fragment ACK contained within the EAP- Response. Retransmitted fragments will contain the same Identifier value. Similarly, when the EAP-FAST server receives an EAP-Response with the M bit set, it must respond with an EAP-Request with EAP-Type of EAP- FAST and no data. This serves as a fragment ACK. The EAP peer MUST wait until it receives the EAP-Request before sending another fragment. In order to prevent errors in the processing of fragments, the EAP server MUST increment the Identifier value for each fragment ACK contained within an EAP-Request, and the peer MUST include this Identifier value in the subsequent fragment contained within an EAP- Response.
(next page on part 2)
