RFC 3748






Network Working Group                                           B. Aboba
Request for Comments: 3748                                     Microsoft
Obsoletes: 2284                                                 L. Blunk
Category: Standards Track                             Merit Network, Inc
                                                           J. Vollbrecht
                                               Vollbrecht Consulting LLC
                                                              J. Carlson
                                                                     Sun
                                                       H. Levkowetz, Ed.
                                                             ipUnplugged
                                                               June 2004


                Extensible Authentication Protocol (EAP)

Status of this Memo



   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice



   Copyright (C) The Internet Society (2004).

Abstract



   This document defines the Extensible Authentication Protocol (EAP),
   an authentication framework which supports multiple authentication
   methods.  EAP typically runs directly over data link layers such as
   Point-to-Point Protocol (PPP) or IEEE 802, without requiring IP.  EAP
   provides its own support for duplicate elimination and
   retransmission, but is reliant on lower layer ordering guarantees.
   Fragmentation is not supported within EAP itself; however, individual
   EAP methods may support this.

   This document obsoletes RFC 2284.  A summary of the changes between
   this document and RFC 2284 is available in Appendix A.











Aboba, et al.               Standards Track                     [Page 1]

RFC 3748                          EAP                          June 2004


Table of Contents



   1.   Introduction. . . . . . . . . . . . . . . . . . . . . . . . .  3
        1.1.  Specification of Requirements . . . . . . . . . . . . .  4
        1.2.  Terminology . . . . . . . . . . . . . . . . . . . . . .  4
        1.3.  Applicability . . . . . . . . . . . . . . . . . . . . .  6
   2.   Extensible Authentication Protocol (EAP). . . . . . . . . . .  7
        2.1.  Support for Sequences . . . . . . . . . . . . . . . . .  9
        2.2.  EAP Multiplexing Model. . . . . . . . . . . . . . . . . 10
        2.3.  Pass-Through Behavior . . . . . . . . . . . . . . . . . 12
        2.4.  Peer-to-Peer Operation. . . . . . . . . . . . . . . . . 14
   3.   Lower Layer Behavior. . . . . . . . . . . . . . . . . . . . . 15
        3.1.  Lower Layer Requirements. . . . . . . . . . . . . . . . 15
        3.2.  EAP Usage Within PPP. . . . . . . . . . . . . . . . . . 18
              3.2.1. PPP Configuration Option Format. . . . . . . . . 18
        3.3.  EAP Usage Within IEEE 802 . . . . . . . . . . . . . . . 19
        3.4.  Lower Layer Indications . . . . . . . . . . . . . . . . 19
   4.   EAP Packet Format . . . . . . . . . . . . . . . . . . . . . . 20
        4.1.  Request and Response. . . . . . . . . . . . . . . . . . 21
        4.2.  Success and Failure . . . . . . . . . . . . . . . . . . 23
        4.3.  Retransmission Behavior . . . . . . . . . . . . . . . . 26
   5.   Initial EAP Request/Response Types. . . . . . . . . . . . . . 27
        5.1.  Identity. . . . . . . . . . . . . . . . . . . . . . . . 28
        5.2.  Notification. . . . . . . . . . . . . . . . . . . . . . 29
        5.3.  Nak . . . . . . . . . . . . . . . . . . . . . . . . . . 31
              5.3.1. Legacy Nak . . . . . . . . . . . . . . . . . . . 31
              5.3.2. Expanded Nak . . . . . . . . . . . . . . . . . . 32
        5.4.  MD5-Challenge . . . . . . . . . . . . . . . . . . . . . 35
        5.5.  One-Time Password (OTP) . . . . . . . . . . . . . . . . 36
        5.6.  Generic Token Card (GTC). . . . . . . . . . . . . . . . 37
        5.7.  Expanded Types. . . . . . . . . . . . . . . . . . . . . 38
        5.8.  Experimental. . . . . . . . . . . . . . . . . . . . . . 40
   6.   IANA Considerations . . . . . . . . . . . . . . . . . . . . . 40
        6.1.  Packet Codes. . . . . . . . . . . . . . . . . . . . . . 41
        6.2.  Method Types. . . . . . . . . . . . . . . . . . . . . . 41
   7.   Security Considerations . . . . . . . . . . . . . . . . . . . 42
        7.1.  Threat Model. . . . . . . . . . . . . . . . . . . . . . 42
        7.2.  Security Claims . . . . . . . . . . . . . . . . . . . . 43
              7.2.1. Security Claims Terminology for EAP Methods. . . 44
        7.3.  Identity Protection . . . . . . . . . . . . . . . . . . 46
        7.4.  Man-in-the-Middle Attacks . . . . . . . . . . . . . . . 47
        7.5.  Packet Modification Attacks . . . . . . . . . . . . . . 48
        7.6.  Dictionary Attacks. . . . . . . . . . . . . . . . . . . 49
        7.7.  Connection to an Untrusted Network. . . . . . . . . . . 49
        7.8.  Negotiation Attacks . . . . . . . . . . . . . . . . . . 50
        7.9.  Implementation Idiosyncrasies . . . . . . . . . . . . . 50
        7.10. Key Derivation. . . . . . . . . . . . . . . . . . . . . 51
        7.11. Weak Ciphersuites . . . . . . . . . . . . . . . . . . . 53



Aboba, et al.               Standards Track                     [Page 2]

RFC 3748                          EAP                          June 2004


        7.12. Link Layer. . . . . . . . . . . . . . . . . . . . . . . 53
        7.13. Separation of Authenticator and Backend Authentication
              Server. . . . . . . . . . . . . . . . . . . . . . . . . 54
        7.14. Cleartext Passwords . . . . . . . . . . . . . . . . . . 55
        7.15. Channel Binding . . . . . . . . . . . . . . . . . . . . 55
        7.16. Protected Result Indications. . . . . . . . . . . . . . 56
   8.   Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . 58
   9.   References. . . . . . . . . . . . . . . . . . . . . . . . . . 59
        9.1.  Normative References. . . . . . . . . . . . . . . . . . 59
        9.2.  Informative References. . . . . . . . . . . . . . . . . 60
   Appendix A. Changes from RFC 2284. . . . . . . . . . . . . . . . . 64
   Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 66
   Full Copyright Statement . . . . . . . . . . . . . . . . . . . . . 67

1.  Introduction



   This document defines the Extensible Authentication Protocol (EAP),
   an authentication framework which supports multiple authentication
   methods.  EAP typically runs directly over data link layers such as
   Point-to-Point Protocol (PPP) or IEEE 802, without requiring IP.  EAP
   provides its own support for duplicate elimination and
   retransmission, but is reliant on lower layer ordering guarantees.
   Fragmentation is not supported within EAP itself; however, individual
   EAP methods may support this.

   EAP may be used on dedicated links, as well as switched circuits, and
   wired as well as wireless links.  To date, EAP has been implemented
   with hosts and routers that connect via switched circuits or dial-up
   lines using PPP [RFC1661].  It has also been implemented with
   switches and access points using IEEE 802 [IEEE-802].  EAP
   encapsulation on IEEE 802 wired media is described in [IEEE-802.1X],
   and encapsulation on IEEE wireless LANs in [IEEE-802.11i].

   One of the advantages of the EAP architecture is its flexibility.
   EAP is used to select a specific authentication mechanism, typically
   after the authenticator requests more information in order to
   determine the specific authentication method to be used.  Rather than
   requiring the authenticator to be updated to support each new
   authentication method, EAP permits the use of a backend
   authentication server, which may implement some or all authentication
   methods, with the authenticator acting as a pass-through for some or
   all methods and peers.

   Within this document, authenticator requirements apply regardless of
   whether the authenticator is operating as a pass-through or not.
   Where the requirement is meant to apply to either the authenticator
   or backend authentication server, depending on where the EAP
   authentication is terminated, the term "EAP server" will be used.



Aboba, et al.               Standards Track                     [Page 3]

RFC 3748                          EAP                          June 2004


1.1.  Specification of Requirements



   In this document, several words are used to signify the requirements
   of the specification.  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



   This document frequently uses the following terms:

   authenticator
      The end of the link initiating EAP authentication.  The term
      authenticator is used in [IEEE-802.1X], and has the same meaning
      in this document.

   peer
      The end of the link that responds to the authenticator.  In
      [IEEE-802.1X], this end is known as the Supplicant.

   Supplicant
      The end of the link that responds to the authenticator in [IEEE-
      802.1X].  In this document, this end of the link is called the
      peer.

   backend authentication server
      A backend authentication server is an entity that provides an
      authentication service to an authenticator.  When used, this
      server typically executes EAP methods for the authenticator.  This
      terminology is also used in [IEEE-802.1X].

   AAA
      Authentication, Authorization, and Accounting.  AAA protocols with
      EAP support include RADIUS [RFC3579] and Diameter [DIAM-EAP].  In
      this document, the terms "AAA server" and "backend authentication
      server" are used interchangeably.

   Displayable Message
      This is interpreted to be a human readable string of characters.
      The message encoding MUST follow the UTF-8 transformation format
      [RFC2279].









Aboba, et al.               Standards Track                     [Page 4]

RFC 3748                          EAP                          June 2004


   EAP server
      The entity that terminates the EAP authentication method with the
      peer.  In the case where no backend authentication server is used,
      the EAP server is part of the authenticator.  In the case where
      the authenticator operates in pass-through mode, the EAP server is
      located on the backend authentication server.

   Silently Discard
      This means the implementation discards the packet without further
      processing.  The implementation SHOULD provide the capability of
      logging the event, including the contents of the silently
      discarded packet, and SHOULD record the event in a statistics
      counter.

   Successful Authentication
      In the context of this document, "successful authentication" is an
      exchange of EAP messages, as a result of which the authenticator
      decides to allow access by the peer, and the peer decides to use
      this access.  The authenticator's decision typically involves both
      authentication and authorization aspects; the peer may
      successfully authenticate to the authenticator, but access may be
      denied by the authenticator due to policy reasons.

   Message Integrity Check (MIC)
      A keyed hash function used for authentication and integrity
      protection of data.  This is usually called a Message
      Authentication Code (MAC), but IEEE 802 specifications (and this
      document) use the acronym MIC to avoid confusion with Medium
      Access Control.

   Cryptographic Separation
      Two keys (x and y) are "cryptographically separate" if an
      adversary that knows all messages exchanged in the protocol cannot
      compute x from y or y from x without "breaking" some cryptographic
      assumption.  In particular, this definition allows that the
      adversary has the knowledge of all nonces sent in cleartext, as
      well as all predictable counter values used in the protocol.
      Breaking a cryptographic assumption would typically require
      inverting a one-way function or predicting the outcome of a
      cryptographic pseudo-random number generator without knowledge of
      the secret state.  In other words, if the keys are
      cryptographically separate, there is no shortcut to compute x from
      y or y from x, but the work an adversary must do to perform this
      computation is equivalent to performing an exhaustive search for
      the secret state value.






Aboba, et al.               Standards Track                     [Page 5]

RFC 3748                          EAP                          June 2004


   Master Session Key (MSK)
      Keying material that is derived between the EAP peer and server
      and exported by the EAP method.  The MSK is at least 64 octets in
      length.  In existing implementations, a AAA server acting as an
      EAP server transports the MSK to the authenticator.

   Extended Master Session Key (EMSK)
      Additional keying material derived between the EAP client and
      server that is exported by the EAP method.  The EMSK is at least
      64 octets in length.  The EMSK is not shared with the
      authenticator or any other third party.  The EMSK is reserved for
      future uses that are not defined yet.

   Result indications
      A method provides result indications if after the method's last
      message is sent and received:

      1) The peer is aware of whether it has authenticated the server,
         as well as whether the server has authenticated it.

      2) The server is aware of whether it has authenticated the peer,
         as well as whether the peer has authenticated it.

   In the case where successful authentication is sufficient to
   authorize access, then the peer and authenticator will also know if
   the other party is willing to provide or accept access.  This may not
   always be the case.  An authenticated peer may be denied access due
   to lack of authorization (e.g., session limit) or other reasons.
   Since the EAP exchange is run between the peer and the server, other
   nodes (such as AAA proxies) may also affect the authorization
   decision.  This is discussed in more detail in Section 7.16.

1.3.  Applicability



   EAP was designed for use in network access authentication, where IP
   layer connectivity may not be available.  Use of EAP for other
   purposes, such as bulk data transport, is NOT RECOMMENDED.

   Since EAP does not require IP connectivity, it provides just enough
   support for the reliable transport of authentication protocols, and
   no more.

   EAP is a lock-step protocol which only supports a single packet in
   flight.  As a result, EAP cannot efficiently transport bulk data,
   unlike transport protocols such as TCP [RFC793] or SCTP [RFC2960].






Aboba, et al.               Standards Track                     [Page 6]

RFC 3748                          EAP                          June 2004


   While EAP provides support for retransmission, it assumes ordering
   guarantees provided by the lower layer, so out of order reception is
   not supported.

   Since EAP does not support fragmentation and reassembly, EAP
   authentication methods generating payloads larger than the minimum
   EAP MTU need to provide fragmentation support.

   While authentication methods such as EAP-TLS [RFC2716] provide
   support for fragmentation and reassembly, the EAP methods defined in
   this document do not.  As a result, if the EAP packet size exceeds
   the EAP MTU of the link, these methods will encounter difficulties.

   EAP authentication is initiated by the server (authenticator),
   whereas many authentication protocols are initiated by the client
   (peer).  As a result, it may be necessary for an authentication
   algorithm to add one or two additional messages (at most one
   roundtrip) in order to run over EAP.

   Where certificate-based authentication is supported, the number of
   additional roundtrips may be much larger due to fragmentation of
   certificate chains.  In general, a fragmented EAP packet will require
   as many round-trips to send as there are fragments.  For example, a
   certificate chain 14960 octets in size would require ten round-trips
   to send with a 1496 octet EAP MTU.

   Where EAP runs over a lower layer in which significant packet loss is
   experienced, or where the connection between the authenticator and
   authentication server experiences significant packet loss, EAP
   methods requiring many round-trips can experience difficulties.  In
   these situations, use of EAP methods with fewer roundtrips is
   advisable.

2.  Extensible Authentication Protocol (EAP)



   The EAP authentication exchange proceeds as follows:

   [1] The authenticator sends a Request to authenticate the peer.  The
       Request has a Type field to indicate what is being requested.
       Examples of Request Types include Identity, MD5-challenge, etc.
       The MD5-challenge Type corresponds closely to the CHAP
       authentication protocol [RFC1994].  Typically, the authenticator
       will send an initial Identity Request; however, an initial
       Identity Request is not required, and MAY be bypassed.  For
       example, the identity may not be required where it is determined
       by the port to which the peer has connected (leased lines,





Aboba, et al.               Standards Track                     [Page 7]

RFC 3748                          EAP                          June 2004


       dedicated switch or dial-up ports), or where the identity is
       obtained in another fashion (via calling station identity or MAC
       address, in the Name field of the MD5-Challenge Response, etc.).

   [2] The peer sends a Response packet in reply to a valid Request.  As
       with the Request packet, the Response packet contains a Type
       field, which corresponds to the Type field of the Request.

   [3] The authenticator sends an additional Request packet, and the
       peer replies with a Response.  The sequence of Requests and
       Responses continues as long as needed.  EAP is a 'lock step'
       protocol, so that other than the initial Request, a new Request
       cannot be sent prior to receiving a valid Response.  The
       authenticator is responsible for retransmitting requests as
       described in Section 4.1.  After a suitable number of
       retransmissions, the authenticator SHOULD end the EAP
       conversation.  The authenticator MUST NOT send a Success or
       Failure packet when retransmitting or when it fails to get a
       response from the peer.

   [4] The conversation continues until the authenticator cannot
       authenticate the peer (unacceptable Responses to one or more
       Requests), in which case the authenticator implementation MUST
       transmit an EAP Failure (Code 4).  Alternatively, the
       authentication conversation can continue until the authenticator
       determines that successful authentication has occurred, in which
       case the authenticator MUST transmit an EAP Success (Code 3).

   Advantages:

   o  The EAP protocol can support multiple authentication mechanisms
      without having to pre-negotiate a particular one.

   o  Network Access Server (NAS) devices (e.g., a switch or access
      point) do not have to understand each authentication method and
      MAY act as a pass-through agent for a backend authentication
      server.  Support for pass-through is optional.  An authenticator
      MAY authenticate local peers, while at the same time acting as a
      pass-through for non-local peers and authentication methods it
      does not implement locally.

   o  Separation of the authenticator from the backend authentication
      server simplifies credentials management and policy decision
      making.







Aboba, et al.               Standards Track                     [Page 8]

RFC 3748                          EAP                          June 2004


   Disadvantages:

   o  For use in PPP, EAP requires the addition of a new authentication
      Type to PPP LCP and thus PPP implementations will need to be
      modified to use it.  It also strays from the previous PPP
      authentication model of negotiating a specific authentication
      mechanism during LCP.  Similarly, switch or access point
      implementations need to support [IEEE-802.1X] in order to use EAP.

   o  Where the authenticator is separate from the backend
      authentication server, this complicates the security analysis and,
      if needed, key distribution.

2.1.  Support for Sequences



   An EAP conversation MAY utilize a sequence of methods.  A common
   example of this is an Identity request followed by a single EAP
   authentication method such as an MD5-Challenge.  However, the peer
   and authenticator MUST utilize only one authentication method (Type 4
   or greater) within an EAP conversation, after which the authenticator
   MUST send a Success or Failure packet.

   Once a peer has sent a Response of the same Type as the initial
   Request, an authenticator MUST NOT send a Request of a different Type
   prior to completion of the final round of a given method (with the
   exception of a Notification-Request) and MUST NOT send a Request for
   an additional method of any Type after completion of the initial
   authentication method; a peer receiving such Requests MUST treat them
   as invalid, and silently discard them.  As a result, Identity Requery
   is not supported.

   A peer MUST NOT send a Nak (legacy or expanded) in reply to a Request
   after an initial non-Nak Response has been sent.  Since spoofed EAP
   Request packets may be sent by an attacker, an authenticator
   receiving an unexpected Nak SHOULD discard it and log the event.

   Multiple authentication methods within an EAP conversation are not
   supported due to their vulnerability to man-in-the-middle attacks
   (see Section 7.4) and incompatibility with existing implementations.

   Where a single EAP authentication method is utilized, but other
   methods are run within it (a "tunneled" method), the prohibition
   against multiple authentication methods does not apply.  Such
   "tunneled" methods appear as a single authentication method to EAP.
   Backward compatibility can be provided, since a peer not supporting a
   "tunneled" method can reply to the initial EAP-Request with a Nak





Aboba, et al.               Standards Track                     [Page 9]

RFC 3748                          EAP                          June 2004


   (legacy or expanded).  To address security vulnerabilities,
   "tunneled" methods MUST support protection against man-in-the-middle
   attacks.

2.2.  EAP Multiplexing Model



   Conceptually, EAP implementations consist of the following
   components:

   [a] Lower layer.  The lower layer is responsible for transmitting and
       receiving EAP frames between the peer and authenticator.  EAP has
       been run over a variety of lower layers including PPP, wired IEEE
       802 LANs [IEEE-802.1X], IEEE 802.11 wireless LANs [IEEE-802.11],
       UDP (L2TP [RFC2661] and IKEv2 [IKEv2]), and TCP [PIC].  Lower
       layer behavior is discussed in Section 3.

   [b] EAP layer.  The EAP layer receives and transmits EAP packets via
       the lower layer, implements duplicate detection and
       retransmission, and delivers and receives EAP messages to and
       from the EAP peer and authenticator layers.

   [c] EAP peer and authenticator layers.  Based on the Code field, the
       EAP layer demultiplexes incoming EAP packets to the EAP peer and
       authenticator layers.  Typically, an EAP implementation on a
       given host will support either peer or authenticator
       functionality, but it is possible for a host to act as both an
       EAP peer and authenticator.  In such an implementation both EAP
       peer and authenticator layers will be present.

   [d] EAP method layers.  EAP methods implement the authentication
       algorithms and receive and transmit EAP messages via the EAP peer
       and authenticator layers.  Since fragmentation support is not
       provided by EAP itself, this is the responsibility of EAP
       methods, which are discussed in Section 5.

   The EAP multiplexing model is illustrated in Figure 1 below.  Note
   that there is no requirement that an implementation conform to this
   model, as long as the on-the-wire behavior is consistent with it.













Aboba, et al.               Standards Track                    [Page 10]

RFC 3748                          EAP                          June 2004


         +-+-+-+-+-+-+-+-+-+-+-+-+  +-+-+-+-+-+-+-+-+-+-+-+-+
         |           |           |  |           |           |
         | EAP method| EAP method|  | EAP method| EAP method|
         | Type = X  | Type = Y  |  | Type = X  | Type = Y  |
         |       V   |           |  |       ^   |           |
         +-+-+-+-!-+-+-+-+-+-+-+-+  +-+-+-+-!-+-+-+-+-+-+-+-+
         |       !               |  |       !               |
         |  EAP  ! Peer layer    |  |  EAP  ! Auth. layer   |
         |       !               |  |       !               |
         +-+-+-+-!-+-+-+-+-+-+-+-+  +-+-+-+-!-+-+-+-+-+-+-+-+
         |       !               |  |       !               |
         |  EAP  ! layer         |  |  EAP  ! layer         |
         |       !               |  |       !               |
         +-+-+-+-!-+-+-+-+-+-+-+-+  +-+-+-+-!-+-+-+-+-+-+-+-+
         |       !               |  |       !               |
         | Lower ! layer         |  | Lower ! layer         |
         |       !               |  |       !               |
         +-+-+-+-!-+-+-+-+-+-+-+-+  +-+-+-+-!-+-+-+-+-+-+-+-+
                 !                          !
                 !   Peer                   ! Authenticator
                 +------------>-------------+

                     Figure 1: EAP Multiplexing Model

   Within EAP, the Code field functions much like a protocol number in
   IP.  It is assumed that the EAP layer demultiplexes incoming EAP
   packets according to the Code field.  Received EAP packets with
   Code=1 (Request), 3 (Success), and 4 (Failure) are delivered by the
   EAP layer to the EAP peer layer, if implemented.  EAP packets with
   Code=2 (Response) are delivered to the EAP authenticator layer, if
   implemented.

   Within EAP, the Type field functions much like a port number in UDP
   or TCP.  It is assumed that the EAP peer and authenticator layers
   demultiplex incoming EAP packets according to their Type, and deliver
   them only to the EAP method corresponding to that Type.  An EAP
   method implementation on a host may register to receive packets from
   the peer or authenticator layers, or both, depending on which role(s)
   it supports.

   Since EAP authentication methods may wish to access the Identity,
   implementations SHOULD make the Identity Request and Response
   accessible to authentication methods (Types 4 or greater), in
   addition to the Identity method.  The Identity Type is discussed in
   Section 5.1.






Aboba, et al.               Standards Track                    [Page 11]

RFC 3748                          EAP                          June 2004


   A Notification Response is only used as confirmation that the peer
   received the Notification Request, not that it has processed it, or
   displayed the message to the user.  It cannot be assumed that the
   contents of the Notification Request or Response are available to
   another method.  The Notification Type is discussed in Section 5.2.

   Nak (Type 3) or Expanded Nak (Type 254) are utilized for the purposes
   of method negotiation.  Peers respond to an initial EAP Request for
   an unacceptable Type with a Nak Response (Type 3) or Expanded Nak
   Response (Type 254).  It cannot be assumed that the contents of the
   Nak Response(s) are available to another method.  The Nak Type(s) are
   discussed in Section 5.3.

   EAP packets with Codes of Success or Failure do not include a Type
   field, and are not delivered to an EAP method.  Success and Failure
   are discussed in Section 4.2.

   Given these considerations, the Success, Failure, Nak Response(s),
   and Notification Request/Response messages MUST NOT be used to carry
   data destined for delivery to other EAP methods.

2.3.  Pass-Through Behavior



   When operating as a "pass-through authenticator", an authenticator
   performs checks on the Code, Identifier, and Length fields as
   described in Section 4.1.  It forwards EAP packets received from the
   peer and destined to its authenticator layer to the backend
   authentication server; packets received from the backend
   authentication server destined to the peer are forwarded to it.

   A host receiving an EAP packet may only do one of three things with
   it: act on it, drop it, or forward it.  The forwarding decision is
   typically based only on examination of the Code, Identifier, and
   Length fields.  A pass-through authenticator implementation MUST be
   capable of forwarding EAP packets received from the peer with Code=2
   (Response) to the backend authentication server. It also MUST be
   capable of receiving EAP packets from the backend authentication
   server and forwarding EAP packets of Code=1 (Request), Code=3
   (Success), and Code=4 (Failure) to the peer.

   Unless the authenticator implements one or more authentication
   methods locally which support the authenticator role, the EAP method
   layer header fields (Type, Type-Data) are not examined as part of the
   forwarding decision.  Where the authenticator supports local
   authentication methods, it MAY examine the Type field to determine
   whether to act on the packet itself or forward it.  Compliant pass-
   through authenticator implementations MUST by default forward EAP
   packets of any Type.



Aboba, et al.               Standards Track                    [Page 12]

RFC 3748                          EAP                          June 2004


   EAP packets received with Code=1 (Request), Code=3 (Success), and
   Code=4 (Failure) are demultiplexed by the EAP layer and delivered to
   the peer layer.  Therefore, unless a host implements an EAP peer
   layer, these packets will be silently discarded.  Similarly, EAP
   packets received with Code=2 (Response) are demultiplexed by the EAP
   layer and delivered to the authenticator layer.  Therefore, unless a
   host implements an EAP authenticator layer, these packets will be
   silently discarded.  The behavior of a "pass-through peer" is
   undefined within this specification, and is unsupported by AAA
   protocols such as RADIUS [RFC3579] and Diameter [DIAM-EAP].

   The forwarding model is illustrated in Figure 2.

        Peer         Pass-through Authenticator   Authentication
                                                      Server

   +-+-+-+-+-+-+                                   +-+-+-+-+-+-+
   |           |                                   |           |
   |EAP method |                                   |EAP method |
   |     V     |                                   |     ^     |
   +-+-+-!-+-+-+   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+   +-+-+-!-+-+-+
   |     !     |   |EAP  |  EAP  |             |   |     !     |
   |     !     |   |Peer |  Auth.| EAP Auth.   |   |     !     |
   |EAP  ! peer|   |     | +-----------+       |   |EAP  !Auth.|
   |     !     |   |     | !     |     !       |   |     !     |
   +-+-+-!-+-+-+   +-+-+-+-!-+-+-+-+-+-!-+-+-+-+   +-+-+-!-+-+-+
   |     !     |   |       !     |     !       |   |     !     |
   |EAP  !layer|   |   EAP !layer| EAP !layer  |   |EAP  !layer|
   |     !     |   |       !     |     !       |   |     !     |
   +-+-+-!-+-+-+   +-+-+-+-!-+-+-+-+-+-!-+-+-+-+   +-+-+-!-+-+-+
   |     !     |   |       !     |     !       |   |     !     |
   |Lower!layer|   |  Lower!layer| AAA ! /IP   |   | AAA ! /IP |
   |     !     |   |       !     |     !       |   |     !     |
   +-+-+-!-+-+-+   +-+-+-+-!-+-+-+-+-+-!-+-+-+-+   +-+-+-!-+-+-+
         !                 !           !                 !
         !                 !           !                 !
         +-------->--------+           +--------->-------+


                   Figure 2: Pass-through Authenticator

   For sessions in which the authenticator acts as a pass-through, it
   MUST determine the outcome of the authentication solely based on the
   Accept/Reject indication sent by the backend authentication server;
   the outcome MUST NOT be determined by the contents of an EAP packet
   sent along with the Accept/Reject indication, or the absence of such
   an encapsulated EAP packet.




Aboba, et al.               Standards Track                    [Page 13]

RFC 3748                          EAP                          June 2004


2.4.  Peer-to-Peer Operation



   Since EAP is a peer-to-peer protocol, an independent and simultaneous
   authentication may take place in the reverse direction (depending on
   the capabilities of the lower layer).  Both ends of the link may act
   as authenticators and peers at the same time.  In this case, it is
   necessary for both ends to implement EAP authenticator and peer
   layers.  In addition, the EAP method implementations on both peers
   must support both authenticator and peer functionality.

   Although EAP supports peer-to-peer operation, some EAP
   implementations, methods, AAA protocols, and link layers may not
   support this.  Some EAP methods may support asymmetric
   authentication, with one type of credential being required for the
   peer and another type for the authenticator.  Hosts supporting peer-
   to-peer operation with such a method would need to be provisioned
   with both types of credentials.

   For example, EAP-TLS [RFC2716] is a client-server protocol in which
   distinct certificate profiles are typically utilized for the client
   and server.  This implies that a host supporting peer-to-peer
   authentication with EAP-TLS would need to implement both the EAP peer
   and authenticator layers, support both peer and authenticator roles
   in the EAP-TLS implementation, and provision certificates appropriate
   for each role.

   AAA protocols such as RADIUS/EAP [RFC3579] and Diameter EAP [DIAM-
   EAP] only support "pass-through authenticator" operation.  As noted
   in [RFC3579] Section 2.6.2, a RADIUS server responds to an Access-
   Request encapsulating an EAP-Request, Success, or Failure packet with
   an Access-Reject.  There is therefore no support for "pass-through
   peer" operation.

   Even where a method is used which supports mutual authentication and
   result indications, several considerations may dictate that two EAP
   authentications (one in each direction) are required.  These include:

   [1] Support for bi-directional session key derivation in the lower
       layer.  Lower layers such as IEEE 802.11 may only support uni-
       directional derivation and transport of transient session keys.
       For example, the group-key handshake defined in [IEEE-802.11i] is
       uni-directional, since in IEEE 802.11 infrastructure mode, only
       the Access Point (AP) sends multicast/broadcast traffic.  In IEEE
       802.11 ad hoc mode, where either peer may send
       multicast/broadcast traffic, two uni-directional group-key






Aboba, et al.               Standards Track                    [Page 14]

RFC 3748                          EAP                          June 2004


       exchanges are required.  Due to limitations of the design, this
       also implies the need for unicast key derivations and EAP method
       exchanges to occur in each direction.

   [2] Support for tie-breaking in the lower layer.  Lower layers such
       as IEEE 802.11 ad hoc do not support "tie breaking" wherein two
       hosts initiating authentication with each other will only go
       forward with a single authentication.  This implies that even if
       802.11 were to support a bi-directional group-key handshake, then
       two authentications, one in each direction, might still occur.

   [3] Peer policy satisfaction.  EAP methods may support result
       indications, enabling the peer to indicate to the EAP server
       within the method that it successfully authenticated the EAP
       server, as well as for the server to indicate that it has
       authenticated the peer.  However, a pass-through authenticator
       will not be aware that the peer has accepted the credentials
       offered by the EAP server, unless this information is provided to
       the authenticator via the AAA protocol.  The authenticator SHOULD
       interpret the receipt of a key attribute within an Accept packet
       as an indication that the peer has successfully authenticated the
       server.

   However, it is possible that the EAP peer's access policy was not
   satisfied during the initial EAP exchange, even though mutual
   authentication occurred.  For example, the EAP authenticator may not
   have demonstrated authorization to act in both peer and authenticator
   roles.  As a result, the peer may require an additional
   authentication in the reverse direction, even if the peer provided an
   indication that the EAP server had successfully authenticated to it.

3.  Lower Layer Behavior



3.1.  Lower Layer Requirements



   EAP makes the following assumptions about lower layers:

   [1] Unreliable transport.  In EAP, the authenticator retransmits
       Requests that have not yet received Responses so that EAP does
       not assume that lower layers are reliable.  Since EAP defines its
       own retransmission behavior, it is possible (though undesirable)
       for retransmission to occur both in the lower layer and the EAP
       layer when EAP is run over a reliable lower layer.








Aboba, et al.               Standards Track                    [Page 15]

RFC 3748                          EAP                          June 2004


   Note that EAP Success and Failure packets are not retransmitted.
   Without a reliable lower layer, and with a non-negligible error rate,
   these packets can be lost, resulting in timeouts.  It is therefore
   desirable for implementations to improve their resilience to loss of
   EAP Success or Failure packets, as described in Section 4.2.

   [2] Lower layer error detection.  While EAP does not assume that the
       lower layer is reliable, it does rely on lower layer error
       detection (e.g., CRC, Checksum, MIC, etc.).  EAP methods may not
       include a MIC, or if they do, it may not be computed over all the
       fields in the EAP packet, such as the Code, Identifier, Length,
       or Type fields.  As a result, without lower layer error
       detection, undetected errors could creep into the EAP layer or
       EAP method layer header fields, resulting in authentication
       failures.

       For example, EAP TLS [RFC2716], which computes its MIC over the
       Type-Data field only, regards MIC validation failures as a fatal
       error.  Without lower layer error detection, this method, and
       others like it, will not perform reliably.

   [3] Lower layer security.  EAP does not require lower layers to
       provide security services such as per-packet confidentiality,
       authentication, integrity, and replay protection.  However, where
       these security services are available, EAP methods supporting Key
       Derivation (see Section 7.2.1) can be used to provide dynamic
       keying material.  This makes it possible to bind the EAP
       authentication to subsequent data and protect against data
       modification, spoofing, or replay.  See Section 7.1 for details.

   [4] Minimum MTU.  EAP is capable of functioning on lower layers that
       provide an EAP MTU size of 1020 octets or greater.

       EAP does not support path MTU discovery, and fragmentation and
       reassembly is not supported by EAP, nor by the methods defined in
       this specification: Identity (1), Notification (2), Nak Response
       (3), MD5-Challenge (4), One Time Password (5), Generic Token Card
       (6), and expanded Nak Response (254) Types.

       Typically, the EAP peer obtains information on the EAP MTU from
       the lower layers and sets the EAP frame size to an appropriate
       value.  Where the authenticator operates in pass-through mode,
       the authentication server does not have a direct way of
       determining the EAP MTU, and therefore relies on the
       authenticator to provide it with this information, such as via
       the Framed-MTU attribute, as described in [RFC3579], Section 2.4.





Aboba, et al.               Standards Track                    [Page 16]

RFC 3748                          EAP                          June 2004


       While methods such as EAP-TLS [RFC2716] support fragmentation and
       reassembly, EAP methods originally designed for use within PPP
       where a 1500 octet MTU is guaranteed for control frames (see
       [RFC1661], Section 6.1) may lack fragmentation and reassembly
       features.

       EAP methods can assume a minimum EAP MTU of 1020 octets in the
       absence of other information.  EAP methods SHOULD include support
       for fragmentation and reassembly if their payloads can be larger
       than this minimum EAP MTU.

       EAP is a lock-step protocol, which implies a certain inefficiency
       when handling fragmentation and reassembly.  Therefore, if the
       lower layer supports fragmentation and reassembly (such as where
       EAP is transported over IP), it may be preferable for
       fragmentation and reassembly to occur in the lower layer rather
       than in EAP.  This can be accomplished by providing an
       artificially large EAP MTU to EAP, causing fragmentation and
       reassembly to be handled within the lower layer.

   [5] Possible duplication.  Where the lower layer is reliable, it will
       provide the EAP layer with a non-duplicated stream of packets.
       However,  while it is desirable that lower layers provide for
       non-duplication, this is not a requirement.  The Identifier field
       provides both the peer and authenticator with the ability to
       detect duplicates.

   [6] Ordering guarantees.  EAP does not require the Identifier to be
       monotonically increasing, and so is reliant on lower layer
       ordering guarantees for correct operation.  EAP was originally
       defined to run on PPP, and [RFC1661] Section 1 has an ordering
       requirement:

           "The Point-to-Point Protocol is designed for simple links
           which transport packets between two peers.  These links
           provide full-duplex simultaneous bi-directional operation,
           and are assumed to deliver packets in order."

       Lower layer transports for EAP MUST preserve ordering between a
       source and destination at a given priority level (the ordering
       guarantee provided by [IEEE-802]).

       Reordering, if it occurs, will typically result in an EAP
       authentication failure, causing EAP authentication to be re-run.
       In an environment in which reordering is likely, it is therefore
       expected that EAP authentication failures will be common.  It is
       RECOMMENDED that EAP only be run over lower layers that provide
       ordering guarantees; running EAP over raw IP or UDP transport is



Aboba, et al.               Standards Track                    [Page 17]

RFC 3748                          EAP                          June 2004


       NOT RECOMMENDED.  Encapsulation of EAP within RADIUS [RFC3579]
       satisfies ordering requirements, since RADIUS is a "lockstep"
       protocol that delivers packets in order.

3.2.  EAP Usage Within PPP



   In order to establish communications over a point-to-point link, each
   end of the PPP link first sends LCP packets to configure the data
   link during the Link Establishment phase.  After the link has been
   established, PPP provides for an optional Authentication phase before
   proceeding to the Network-Layer Protocol phase.

   By default, authentication is not mandatory.  If authentication of
   the link is desired, an implementation MUST specify the
   Authentication Protocol Configuration Option during the Link
   Establishment phase.

   If the identity of the peer has been established in the
   Authentication phase, the server can use that identity in the
   selection of options for the following network layer negotiations.

   When implemented within PPP, EAP does not select a specific
   authentication mechanism at the PPP Link Control Phase, but rather
   postpones this until the Authentication Phase.  This allows the
   authenticator to request more information before determining the
   specific authentication mechanism.  This also permits the use of a
   "backend" server which actually implements the various mechanisms
   while the PPP authenticator merely passes through the authentication
   exchange.  The PPP Link Establishment and Authentication phases, and
   the Authentication Protocol Configuration Option, are defined in The
   Point-to-Point Protocol (PPP) [RFC1661].

3.2.1.  PPP Configuration Option Format



   A summary of the PPP Authentication Protocol Configuration Option
   format to negotiate EAP follows.  The fields are transmitted from
   left to right.

   Exactly one EAP packet is encapsulated in the Information field of a
   PPP Data Link Layer frame where the protocol field indicates type hex
   C227 (PPP EAP).










Aboba, et al.               Standards Track                    [Page 18]

RFC 3748                          EAP                          June 2004


    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
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |     Type      |    Length     |     Authentication Protocol   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Type

      3

   Length

      4

   Authentication Protocol

      C227 (Hex) for Extensible Authentication Protocol (EAP)

3.3.  EAP Usage Within IEEE 802



   The encapsulation of EAP over IEEE 802 is defined in [IEEE-802.1X].
   The IEEE 802 encapsulation of EAP does not involve PPP, and IEEE
   802.1X does not include support for link or network layer
   negotiations.  As a result, within IEEE 802.1X, it is not possible to
   negotiate non-EAP authentication mechanisms, such as PAP or CHAP
   [RFC1994].

3.4.  Lower Layer Indications



   The reliability and security of lower layer indications is dependent
   on the lower layer.  Since EAP is media independent, the presence or
   absence of lower layer security is not taken into account in the
   processing of EAP messages.

   To improve reliability, if a peer receives a lower layer success
   indication as defined in Section 7.2, it MAY conclude that a Success
   packet has been lost, and behave as if it had actually received a
   Success packet.  This includes choosing to ignore the Success in some
   circumstances as described in Section 4.2.

   A discussion of some reliability and security issues with lower layer
   indications in PPP, IEEE 802 wired networks, and IEEE 802.11 wireless
   LANs can be found in the Security Considerations, Section 7.12.

   After EAP authentication is complete, the peer will typically
   transmit and receive data via the authenticator.  It is desirable to
   provide assurance that the entities transmitting data are the same
   ones that successfully completed EAP authentication.  To accomplish



Aboba, et al.               Standards Track                    [Page 19]

RFC 3748                          EAP                          June 2004


   this, it is necessary for the lower layer to provide per-packet
   integrity, authentication and replay protection, and to bind these
   per-packet services to the keys derived during EAP authentication.
   Otherwise, it is possible for subsequent data traffic to be modified,
   spoofed, or replayed.

   Where keying material for the lower layer ciphersuite is itself
   provided by EAP, ciphersuite negotiation and key activation are
   controlled by the lower layer.  In PPP, ciphersuites are negotiated
   within ECP so that it is not possible to use keys derived from EAP
   authentication until the completion of ECP.  Therefore, an initial
   EAP exchange cannot be protected by a PPP ciphersuite, although EAP
   re-authentication can be protected.

   In IEEE 802 media, initial key activation also typically occurs after
   completion of EAP authentication.  Therefore an initial EAP exchange
   typically cannot be protected by the lower layer ciphersuite,
   although an EAP re-authentication or pre-authentication exchange can
   be protected.

4.  EAP Packet Format



   A summary of the EAP packet format is shown below.  The fields are
   transmitted from left to right.

    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             |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |    Data ...
   +-+-+-+-+

   Code

      The Code field is one octet and identifies the Type of EAP packet.
      EAP Codes are assigned as follows:

         1       Request
         2       Response
         3       Success
         4       Failure

      Since EAP only defines Codes 1-4, EAP packets with other codes
      MUST be silently discarded by both authenticators and peers.






Aboba, et al.               Standards Track                    [Page 20]

RFC 3748                          EAP                          June 2004


   Identifier

      The Identifier field is one octet and aids in matching Responses
      with Requests.

   Length

      The Length field is two octets and indicates the length, in
      octets, of the EAP packet including the Code, Identifier, Length,
      and Data fields.  Octets outside the range of the Length field
      should be treated as Data Link Layer padding and MUST be ignored
      upon reception.  A message with the Length field set to a value
      larger than the number of received octets MUST be silently
      discarded.

   Data

      The Data field is zero or more octets.  The format of the Data
      field is determined by the Code field.

4.1.  Request and Response



   Description

      The Request packet (Code field set to 1) is sent by the
      authenticator to the peer.  Each Request has a Type field which
      serves to indicate what is being requested.  Additional Request
      packets MUST be sent until a valid Response packet is received, an
      optional retry counter expires, or a lower layer failure
      indication is received.

      Retransmitted Requests MUST be sent with the same Identifier value
      in order to distinguish them from new Requests.  The content of
      the data field is dependent on the Request Type.  The peer MUST
      send a Response packet in reply to a valid Request packet.
      Responses MUST only be sent in reply to a valid Request and never
      be retransmitted on a timer.

      If a peer receives a valid duplicate Request for which it has
      already sent a Response, it MUST resend its original Response
      without reprocessing the Request.  Requests MUST be processed in
      the order that they are received, and MUST be processed to their
      completion before inspecting the next Request.

   A summary of the Request and Response packet format follows.  The
   fields are transmitted from left to right.





Aboba, et al.               Standards Track                    [Page 21]

RFC 3748                          EAP                          June 2004


    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      |  Type-Data ...
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-

   Code

      1 for Request
      2 for Response

   Identifier

      The Identifier field is one octet.  The Identifier field MUST be
      the same if a Request packet is retransmitted due to a timeout
      while waiting for a Response.  Any new (non-retransmission)
      Requests MUST modify the Identifier field.

      The Identifier field of the Response MUST match that of the
      currently outstanding Request.  An authenticator receiving a
      Response whose Identifier value does not match that of the
      currently outstanding Request MUST silently discard the Response.

      In order to avoid confusion between new Requests and
      retransmissions, the Identifier value chosen for each new Request
      need only be different from the previous Request, but need not be
      unique within the conversation.  One way to achieve this is to
      start the Identifier at an initial value and increment it for each
      new Request.  Initializing the first Identifier with a random
      number rather than starting from zero is recommended, since it
      makes sequence attacks somewhat more difficult.

      Since the Identifier space is unique to each session,
      authenticators are not restricted to only 256 simultaneous
      authentication conversations.  Similarly, with re-authentication,
      an EAP conversation might continue over a long period of time, and
      is not limited to only 256 roundtrips.

   Implementation Note: The authenticator is responsible for
   retransmitting Request messages.  If the Request message is obtained
   from elsewhere (such as from a backend authentication server), then
   the authenticator will need to save a copy of the Request in order to
   accomplish this.  The peer is responsible for detecting and handling
   duplicate Request messages before processing them in any way,
   including passing them on to an outside party.  The authenticator is
   also responsible for discarding Response messages with a non-matching



Aboba, et al.               Standards Track                    [Page 22]

RFC 3748                          EAP                          June 2004


   Identifier value before acting on them in any way, including passing
   them on to the backend authentication server for verification.  Since
   the authenticator can retransmit before receiving a Response from the
   peer, the authenticator can receive multiple Responses, each with a
   matching Identifier.  Until a new Request is received by the
   authenticator, the Identifier value is not updated, so that the
   authenticator forwards Responses to the backend authentication
   server, one at a time.

   Length

      The Length field is two octets and indicates the length of the EAP
      packet including the Code, Identifier, Length, Type, and Type-Data
      fields.  Octets outside the range of the Length field should be
      treated as Data Link Layer padding and MUST be ignored upon
      reception.  A message with the Length field set to a value larger
      than the number of received octets MUST be silently discarded.

   Type

      The Type field is one octet.  This field indicates the Type of
      Request or Response.  A single Type MUST be specified for each EAP
      Request or Response.  An initial specification of Types follows in
      Section 5 of this document.

      The Type field of a Response MUST either match that of the
      Request, or correspond to a legacy or Expanded Nak (see Section
      5.3) indicating that a Request Type is unacceptable to the peer.
      A peer MUST NOT send a Nak (legacy or expanded) in response to a
      Request, after an initial non-Nak Response has been sent.  An EAP
      server receiving a Response not meeting these requirements MUST
      silently discard it.

   Type-Data

      The Type-Data field varies with the Type of Request and the
      associated Response.

4.2.  Success and Failure



   The Success packet is sent by the authenticator to the peer after
   completion of an EAP authentication method (Type 4 or greater) to
   indicate that the peer has authenticated successfully to the
   authenticator.  The authenticator MUST transmit an EAP packet with
   the Code field set to 3 (Success).  If the authenticator cannot
   authenticate the peer (unacceptable Responses to one or more
   Requests), then after unsuccessful completion of the EAP method in
   progress, the implementation MUST transmit an EAP packet with the



Aboba, et al.               Standards Track                    [Page 23]

RFC 3748                          EAP                          June 2004


   Code field set to 4 (Failure).  An authenticator MAY wish to issue
   multiple Requests before sending a Failure response in order to allow
   for human typing mistakes.  Success and Failure packets MUST NOT
   contain additional data.

   Success and Failure packets MUST NOT be sent by an EAP authenticator
   if the specification of the given method does not explicitly permit
   the method to finish at that point.  A peer EAP implementation
   receiving a Success or Failure packet where sending one is not
   explicitly permitted MUST silently discard it.  By default, an EAP
   peer MUST silently discard a "canned" Success packet (a Success
   packet sent immediately upon connection).  This ensures that a rogue
   authenticator will not be able to bypass mutual authentication by
   sending a Success packet prior to conclusion of the EAP method
   conversation.

   Implementation Note: Because the Success and Failure packets are not
   acknowledged, they are not retransmitted by the authenticator, and
   may be potentially lost.  A peer MUST allow for this circumstance as
   described in this note.  See also Section 3.4 for guidance on the
   processing of lower layer success and failure indications.

   As described in Section 2.1, only a single EAP authentication method
   is allowed within an EAP conversation.  EAP methods may implement
   result indications.  After the authenticator sends a failure result
   indication to the peer, regardless of the response from the peer, it
   MUST subsequently send a Failure packet.  After the authenticator
   sends a success result indication to the peer and receives a success
   result indication from the peer, it MUST subsequently send a Success
   packet.

   On the peer, once the method completes unsuccessfully (that is,
   either the authenticator sends a failure result indication, or the
   peer decides that it does not want to continue the conversation,
   possibly after sending a failure result indication), the peer MUST
   terminate the conversation and indicate failure to the lower layer.
   The peer MUST silently discard Success packets and MAY silently
   discard Failure packets.  As a result, loss of a Failure packet need
   not result in a timeout.

   On the peer, after success result indications have been exchanged by
   both sides, a Failure packet MUST be silently discarded.  The peer
   MAY, in the event that an EAP Success is not received, conclude that
   the EAP Success packet was lost and that authentication concluded
   successfully.






Aboba, et al.               Standards Track                    [Page 24]

RFC 3748                          EAP                          June 2004


   If the authenticator has not sent a result indication, and the peer
   is willing to continue the conversation, the peer waits for a Success
   or Failure packet once the method completes, and MUST NOT silently
   discard either of them.  In the event that neither a Success nor
   Failure packet is received, the peer SHOULD terminate the
   conversation to avoid lengthy timeouts in case the lost packet was an
   EAP Failure.

   If the peer attempts to authenticate to the authenticator and fails
   to do so, the authenticator MUST send a Failure packet and MUST NOT
   grant access by sending a Success packet.  However, an authenticator
   MAY omit having the peer authenticate to it in situations where
   limited access is offered (e.g., guest access).  In this case, the
   authenticator MUST send a Success packet.

   Where the peer authenticates successfully to the authenticator, but
   the authenticator does not send a result indication, the
   authenticator MAY deny access by sending a Failure packet where the
   peer is not currently authorized for network access.

   A summary of the Success and Failure packet format is shown below.
   The fields are transmitted from left to right.

    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             |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Code

      3 for Success
      4 for Failure

   Identifier

      The Identifier field is one octet and aids in matching replies to
      Responses.  The Identifier field MUST match the Identifier field
      of the Response packet that it is sent in response to.

   Length

      4








Aboba, et al.               Standards Track                    [Page 25]

RFC 3748                          EAP                          June 2004


4.3.  Retransmission Behavior



   Because the authentication process will often involve user input,
   some care must be taken when deciding upon retransmission strategies
   and authentication timeouts.  By default, where EAP is run over an
   unreliable lower layer, the EAP retransmission timer SHOULD be
   dynamically estimated.  A maximum of 3-5 retransmissions is
   suggested.

   When run over a reliable lower layer (e.g., EAP over ISAKMP/TCP, as
   within [PIC]), the authenticator retransmission timer SHOULD be set
   to an infinite value, so that retransmissions do not occur at the EAP
   layer.  The peer may still maintain a timeout value so as to avoid
   waiting indefinitely for a Request.

   Where the authentication process requires user input, the measured
   round trip times may be determined by user responsiveness rather than
   network characteristics, so that dynamic RTO estimation may not be
   helpful.  Instead, the retransmission timer SHOULD be set so as to
   provide sufficient time for the user to respond, with longer timeouts
   required in certain cases, such as where Token Cards (see Section
   5.6) are involved.

   In order to provide the EAP authenticator with guidance as to the
   appropriate timeout value, a hint can be communicated to the
   authenticator by the backend authentication server (such as via the
   RADIUS Session-Timeout attribute).

   In order to dynamically estimate the EAP retransmission timer, the
   algorithms for the estimation of SRTT, RTTVAR, and RTO described in
   [RFC2988] are RECOMMENDED, including use of Karn's algorithm, with
   the following potential modifications:

   [a] In order to avoid synchronization behaviors that can occur with
       fixed timers among distributed systems, the retransmission timer
       is calculated with a jitter by using the RTO value and randomly
       adding a value drawn between -RTOmin/2 and RTOmin/2.  Alternative
       calculations to create jitter MAY be used.  These MUST be
       pseudo-random.  For a discussion of pseudo-random number
       generation, see [RFC1750].

   [b] When EAP is transported over a single link (as opposed to over
       the Internet), smaller values of RTOinitial, RTOmin, and RTOmax
       MAY be used.  Recommended values are RTOinitial=1 second,
       RTOmin=200ms, and RTOmax=20 seconds.






Aboba, et al.               Standards Track                    [Page 26]

RFC 3748                          EAP                          June 2004


   [c] When EAP is transported over a single link (as opposed to over
       the Internet), estimates MAY be done on a per-authenticator
       basis, rather than a per-session basis.  This enables the
       retransmission estimate to make the most use of information on
       link-layer behavior.

   [d] An EAP implementation MAY clear SRTT and RTTVAR after backing off
       the timer multiple times, as it is likely that the current SRTT
       and RTTVAR are bogus in this situation.  Once SRTT and RTTVAR are
       cleared, they should be initialized with the next RTT sample
       taken as described in [RFC2988] equation 2.2.

5.  Initial EAP Request/Response Types



   This section defines the initial set of EAP Types used in Request/
   Response exchanges.  More Types may be defined in future documents.
   The Type field is one octet and identifies the structure of an EAP
   Request or Response packet.  The first 3 Types are considered special
   case Types.

   The remaining Types define authentication exchanges.  Nak (Type 3) or
   Expanded Nak (Type 254) are valid only for Response packets, they
   MUST NOT be sent in a Request.

   All EAP implementations MUST support Types 1-4, which are defined in
   this document, and SHOULD support Type 254.  Implementations MAY
   support other Types defined here or in future RFCs.

             1       Identity
             2       Notification
             3       Nak (Response only)
             4       MD5-Challenge
             5       One Time Password (OTP)
             6       Generic Token Card (GTC)
           254       Expanded Types
           255       Experimental use

   EAP methods MAY support authentication based on shared secrets.  If
   the shared secret is a passphrase entered by the user,
   implementations MAY support entering passphrases with non-ASCII
   characters.  In this case, the input should be processed using an
   appropriate stringprep [RFC3454] profile, and encoded in octets using
   UTF-8 encoding [RFC2279].  A preliminary version of a possible
   stringprep profile is described in [SASLPREP].







Aboba, et al.               Standards Track                    [Page 27]

RFC 3748                          EAP                          June 2004


5.1.  Identity



   Description

      The Identity Type is used to query the identity of the peer.
      Generally, the authenticator will issue this as the initial
      Request.  An optional displayable message MAY be included to
      prompt the peer in the case where there is an expectation of
      interaction with a user.  A Response of Type 1 (Identity) SHOULD
      be sent in Response to a Request with a Type of 1 (Identity).

      Some EAP implementations piggy-back various options into the
      Identity Request after a NUL-character.  By default, an EAP
      implementation SHOULD NOT assume that an Identity Request or
      Response can be larger than 1020 octets.

      It is RECOMMENDED that the Identity Response be used primarily for
      routing purposes and selecting which EAP method to use.  EAP
      Methods SHOULD include a method-specific mechanism for obtaining
      the identity, so that they do not have to rely on the Identity
      Response.  Identity Requests and Responses are sent in cleartext,
      so an attacker may snoop on the identity, or even modify or spoof
      identity exchanges.  To address these threats, it is preferable
      for an EAP method to include an identity exchange that supports
      per-packet authentication, integrity and replay protection, and
      confidentiality.  The Identity Response may not be the appropriate
      identity for the method; it may have been truncated or obfuscated
      so as to provide privacy, or it may have been decorated for
      routing purposes.  Where the peer is configured to only accept
      authentication methods supporting protected identity exchanges,
      the peer MAY provide an abbreviated Identity Response (such as
      omitting the peer-name portion of the NAI [RFC2486]).  For further
      discussion of identity protection, see Section 7.3.

   Implementation Note: The peer MAY obtain the Identity via user input.
   It is suggested that the authenticator retry the Identity Request in
   the case of an invalid Identity or authentication failure to allow
   for potential typos on the part of the user.  It is suggested that
   the Identity Request be retried a minimum of 3 times before
   terminating the authentication.  The Notification Request MAY be used
   to indicate an invalid authentication attempt prior to transmitting a
   new Identity Request (optionally, the failure MAY be indicated within
   the message of the new Identity Request itself).








Aboba, et al.               Standards Track                    [Page 28]

RFC 3748                          EAP                          June 2004


   Type

      1

   Type-Data

      This field MAY contain a displayable message in the Request,
      containing UTF-8 encoded ISO 10646 characters [RFC2279].  Where
      the Request contains a null, only the portion of the field prior
      to the null is displayed.  If the Identity is unknown, the
      Identity Response field should be zero bytes in length.  The
      Identity Response field MUST NOT be null terminated.  In all
      cases, the length of the Type-Data field is derived from the
      Length field of the Request/Response packet.

   Security Claims (see Section 7.2):

      Auth. mechanism:           None
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   N/A
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No

5.2.  Notification



   Description

      The Notification Type is optionally used to convey a displayable
      message from the authenticator to the peer.  An authenticator MAY
      send a Notification Request to the peer at any time when there is
      no outstanding Request, prior to completion of an EAP
      authentication method.  The peer MUST respond to a Notification
      Request with a Notification Response unless the EAP authentication
      method specification prohibits the use of Notification messages.
      In any case, a Nak Response MUST NOT be sent in response to a
      Notification Request.  Note that the default maximum length of a
      Notification Request is 1020 octets.  By default, this leaves at
      most 1015 octets for the human readable message.




Aboba, et al.               Standards Track                    [Page 29]

RFC 3748                          EAP                          June 2004


      An EAP method MAY indicate within its specification that
      Notification messages must not be sent during that method.  In
      this case, the peer MUST silently discard Notification Requests
      from the point where an initial Request for that Type is answered
      with a Response of the same Type.

      The peer SHOULD display this message to the user or log it if it
      cannot be displayed.  The Notification Type is intended to provide
      an acknowledged notification of some imperative nature, but it is
      not an error indication, and therefore does not change the state
      of the peer.  Examples include a password with an expiration time
      that is about to expire, an OTP sequence integer which is nearing
      0, an authentication failure warning, etc.  In most circumstances,
      Notification should not be required.

   Type

      2

   Type-Data

      The Type-Data field in the Request contains a displayable message
      greater than zero octets in length, containing UTF-8 encoded ISO
      10646 characters [RFC2279].  The length of the message is
      determined by the Length field of the Request packet.  The message
      MUST NOT be null terminated.  A Response MUST be sent in reply to
      the Request with a Type field of 2 (Notification).  The Type-Data
      field of the Response is zero octets in length.  The Response
      should be sent immediately (independent of how the message is
      displayed or logged).

   Security Claims (see Section 7.2):

      Auth. mechanism:           None
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   N/A
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No




Aboba, et al.               Standards Track                    [Page 30]

RFC 3748                          EAP                          June 2004


5.3.  Nak



5.3.1.  Legacy Nak



   Description

      The legacy Nak Type is valid only in Response messages.  It is
      sent in reply to a Request where the desired authentication Type
      is unacceptable.  Authentication Types are numbered 4 and above.
      The Response contains one or more authentication Types desired by
      the Peer.  Type zero (0) is used to indicate that the sender has
      no viable alternatives, and therefore the authenticator SHOULD NOT
      send another Request after receiving a Nak Response containing a
      zero value.

      Since the legacy Nak Type is valid only in Responses and has very
      limited functionality, it MUST NOT be used as a general purpose
      error indication, such as for communication of error messages, or
      negotiation of parameters specific to a particular EAP method.

   Code

      2 for Response.

   Identifier

      The Identifier field is one octet and aids in matching Responses
      with Requests.  The Identifier field of a legacy Nak Response MUST
      match the Identifier field of the Request packet that it is sent
      in response to.

   Length

      >=6

   Type

      3

   Type-Data

      Where a peer receives a Request for an unacceptable authentication
      Type (4-253,255), or a peer lacking support for Expanded Types
      receives a Request for Type 254, a Nak Response (Type 3) MUST be
      sent.  The Type-Data field of the Nak Response (Type 3) MUST
      contain one or more octets indicating the desired authentication
      Type(s), one octet per Type, or the value zero (0) to indicate no
      proposed alternative.  A peer supporting Expanded Types that



Aboba, et al.               Standards Track                    [Page 31]

RFC 3748                          EAP                          June 2004


      receives a Request for an unacceptable authentication Type (4-253,
      255) MAY include the value 254 in the Nak Response (Type 3) to
      indicate the desire for an Expanded authentication Type. If the
      authenticator can accommodate this preference, it will respond
      with an Expanded Type Request (Type 254).

   Security Claims (see Section 7.2):

      Auth. mechanism:           None
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   N/A
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No


5.3.2.  Expanded Nak



   Description

      The Expanded Nak Type is valid only in Response messages.  It MUST
      be sent only in reply to a Request of Type 254 (Expanded Type)
      where the authentication Type is unacceptable.  The Expanded Nak
      Type uses the Expanded Type format itself, and the Response
      contains one or more authentication Types desired by the peer, all
      in Expanded Type format.  Type zero (0) is used to indicate that
      the sender has no viable alternatives.  The general format of the
      Expanded Type is described in Section 5.7.

      Since the Expanded Nak Type is valid only in Responses and has
      very limited functionality, it MUST NOT be used as a general
      purpose error indication, such as for communication of error
      messages, or negotiation of parameters specific to a particular
      EAP method.

   Code

      2 for Response.





Aboba, et al.               Standards Track                    [Page 32]

RFC 3748                          EAP                          June 2004


   Identifier

      The Identifier field is one octet and aids in matching Responses
      with Requests.  The Identifier field of an Expanded Nak Response
      MUST match the Identifier field of the Request packet that it is
      sent in response to.

   Length

      >=20

   Type

      254

   Vendor-Id

      0 (IETF)



   Vendor-Type

      3 (Nak)

   Vendor-Data

      The Expanded Nak Type is only sent when the Request contains an
      Expanded Type (254) as defined in Section 5.7.  The Vendor-Data
      field of the Nak Response MUST contain one or more authentication
      Types (4 or greater), all in expanded format, 8 octets per Type,
      or the value zero (0), also in Expanded Type format, to indicate
      no proposed alternative.  The desired authentication Types may
      include a mixture of Vendor-Specific and IETF Types.  For example,
      an Expanded Nak Response indicating a preference for OTP (Type 5),
      and an MIT (Vendor-Id=20) Expanded Type of 6 would appear as
      follows:
















Aboba, et al.               Standards Track                    [Page 33]

RFC 3748                          EAP                          June 2004


    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
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |     2         |  Identifier   |           Length=28           |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   Type=254    |                0 (IETF)                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                3 (Nak)                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   Type=254    |                0 (IETF)                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                5 (OTP)                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   Type=254    |                20 (MIT)                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                6                              |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   An Expanded Nak Response indicating a no desired alternative would
   appear as follows:

    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
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |     2         |  Identifier   |           Length=20           |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   Type=254    |                0 (IETF)                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                3 (Nak)                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |   Type=254    |                0 (IETF)                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                0 (No alternative)             |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Security Claims (see Section 7.2):

      Auth. mechanism:           None
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   N/A
      Fast reconnect:            No
      Crypt. binding:            N/A



Aboba, et al.               Standards Track                    [Page 34]

RFC 3748                          EAP                          June 2004


      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No


5.4.  MD5-Challenge



   Description

      The MD5-Challenge Type is analogous to the PPP CHAP protocol
      [RFC1994] (with MD5 as the specified algorithm).  The Request
      contains a "challenge" message to the peer.  A Response MUST be
      sent in reply to the Request.  The Response MAY be either of Type
      4 (MD5-Challenge), Nak (Type 3), or Expanded Nak (Type 254).  The
      Nak reply indicates the peer's desired authentication Type(s).
      EAP peer and EAP server implementations MUST support the MD5-
      Challenge mechanism.  An authenticator that supports only pass-
      through MUST allow communication with a backend authentication
      server that is capable of supporting MD5-Challenge, although the
      EAP authenticator implementation need not support MD5-Challenge
      itself.  However, if the EAP authenticator can be configured to
      authenticate peers locally (e.g., not operate in pass-through),
      then the requirement for support of the MD5-Challenge mechanism
      applies.

      Note that the use of the Identifier field in the MD5-Challenge
      Type is different from that described in [RFC1994].  EAP allows
      for retransmission of MD5-Challenge Request packets, while
      [RFC1994] states that both the Identifier and Challenge fields
      MUST change each time a Challenge (the CHAP equivalent of the
      MD5-Challenge Request packet) is sent.

      Note: [RFC1994] treats the shared secret as an octet string, and
      does not specify how it is entered into the system (or if it is
      handled by the user at all).  EAP MD5-Challenge implementations
      MAY support entering passphrases with non-ASCII characters.  See
      Section 5 for instructions how the input should be processed and
      encoded into octets.

   Type

      4

   Type-Data

      The contents of the Type-Data field is summarized below.  For
      reference on the use of these fields, see the PPP Challenge
      Handshake Authentication Protocol [RFC1994].



Aboba, et al.               Standards Track                    [Page 35]

RFC 3748                          EAP                          June 2004


    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
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Value-Size   |  Value ...
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Name ...
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Security Claims (see Section 7.2):

      Auth. mechanism:           Password or pre-shared key.
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   No
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No

5.5.  One-Time Password (OTP)



   Description

      The One-Time Password system is defined in "A One-Time Password
      System" [RFC2289] and "OTP Extended Responses" [RFC2243].  The
      Request contains an OTP challenge in the format described in
      [RFC2289].  A Response MUST be sent in reply to the Request.  The
      Response MUST be of Type 5 (OTP), Nak (Type 3), or Expanded Nak
      (Type 254).  The Nak Response indicates the peer's desired
      authentication Type(s).  The EAP OTP method is intended for use
      with the One-Time Password system only, and MUST NOT be used to
      provide support for cleartext passwords.

   Type

      5









Aboba, et al.               Standards Track                    [Page 36]

RFC 3748                          EAP                          June 2004


   Type-Data

      The Type-Data field contains the OTP "challenge" as a displayable
      message in the Request.  In the Response, this field is used for
      the 6 words from the OTP dictionary [RFC2289].  The messages MUST
      NOT
be null terminated.  The length of the field is derived from
      the Length field of the Request/Reply packet.

      Note: [RFC2289] does not specify how the secret pass-phrase is
      entered by the user, or how the pass-phrase is converted into
      octets.  EAP OTP implementations MAY support entering passphrases
      with non-ASCII characters.  See Section 5 for instructions on how
      the input should be processed and encoded into octets.

   Security Claims (see Section 7.2):

      Auth. mechanism:           One-Time Password
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         Yes
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   No
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No


5.6.  Generic Token Card (GTC)



   Description

      The Generic Token Card Type is defined for use with various Token
      Card implementations which require user input.  The Request
      contains a displayable message and the Response contains the Token
      Card information necessary for authentication.  Typically, this
      would be information read by a user from the Token card device and
      entered as ASCII text.  A Response MUST be sent in reply to the
      Request.  The Response MUST be of Type 6 (GTC), Nak (Type 3), or
      Expanded Nak (Type 254).  The Nak Response indicates the peer's
      desired authentication Type(s).  The EAP GTC method is intended
      for use with the Token Cards supporting challenge/response





Aboba, et al.               Standards Track                    [Page 37]

RFC 3748                          EAP                          June 2004


      authentication and MUST NOT be used to provide support for
      cleartext passwords in the absence of a protected tunnel with
      server authentication.

   Type

      6

   Type-Data

      The Type-Data field in the Request contains a displayable message
      greater than zero octets in length.  The length of the message is
      determined by the Length field of the Request packet.  The message
      MUST NOT be null terminated.  A Response MUST be sent in reply to
      the Request with a Type field of 6 (Generic Token Card).  The
      Response contains data from the Token Card required for
      authentication.  The length of the data is determined by the
      Length field of the Response packet.

      EAP GTC implementations MAY support entering a response with non-
      ASCII characters.  See Section 5 for instructions how the input
      should be processed and encoded into octets.

   Security Claims (see Section 7.2):

      Auth. mechanism:           Hardware token.
      Ciphersuite negotiation:   No
      Mutual authentication:     No
      Integrity protection:      No
      Replay protection:         No
      Confidentiality:           No
      Key derivation:            No
      Key strength:              N/A
      Dictionary attack prot.:   No
      Fast reconnect:            No
      Crypt. binding:            N/A
      Session independence:      N/A
      Fragmentation:             No
      Channel binding:           No


5.7.  Expanded Types



   Description

      Since many of the existing uses of EAP are vendor-specific, the
      Expanded method Type is available to allow vendors to support
      their own Expanded Types not suitable for general usage.



Aboba, et al.               Standards Track                    [Page 38]

RFC 3748                          EAP                          June 2004


      The Expanded Type is also used to expand the global Method Type
      space beyond the original 255 values.  A Vendor-Id of 0 maps the
      original 255 possible Types onto a space of 2^32-1 possible Types.
      (Type 0 is only used in a Nak Response to indicate no acceptable
      alternative).

      An implementation that supports the Expanded attribute MUST treat
      EAP Types that are less than 256 equivalently, whether they appear
      as a single octet or as the 32-bit Vendor-Type within an Expanded
      Type where Vendor-Id is 0.  Peers not equipped to interpret the
      Expanded Type MUST send a Nak as described in Section 5.3.1, and
      negotiate a more suitable authentication method.

      A summary of the Expanded Type format is shown below.  The fields
      are transmitted from left to right.

    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
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |     Type      |               Vendor-Id                       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                          Vendor-Type                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |              Vendor data...
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Type

      254 for Expanded Type



   Vendor-Id

      The Vendor-Id is 3 octets and represents the SMI Network
      Management Private Enterprise Code of the Vendor in network byte
      order, as allocated by IANA.  A Vendor-Id of zero is reserved for
      use by the IETF in providing an expanded global EAP Type space.

   Vendor-Type

      The Vendor-Type field is four octets and represents the vendor-
      specific method Type.

      If the Vendor-Id is zero, the Vendor-Type field is an extension
      and superset of the existing namespace for EAP Types.  The first
      256 Types are reserved for compatibility with single-octet EAP
      Types that have already been assigned or may be assigned in the
      future.  Thus, EAP Types from 0 through 255 are semantically
      identical, whether they appear as single octet EAP Types or as



Aboba, et al.               Standards Track                    [Page 39]

RFC 3748                          EAP                          June 2004


      Vendor-Types when Vendor-Id is zero.  There is one exception to
      this rule: Expanded Nak and Legacy Nak packets share the same
      Type, but must be treated differently because they have a
      different format.

   Vendor-Data

      The Vendor-Data field is defined by the vendor.  Where a Vendor-Id
      of zero is present, the Vendor-Data field will be used for
      transporting the contents of EAP methods of Types defined by the
      IETF.

5.8.  Experimental



   Description

      The Experimental Type has no fixed format or content.  It is
      intended for use when experimenting with new EAP Types.  This Type
      is intended for experimental and testing purposes.  No guarantee
      is made for interoperability between peers using this Type, as
      outlined in [RFC3692].

   Type

      255

   Type-Data

      Undefined

6.  IANA Considerations



   This section provides guidance to the Internet Assigned Numbers
   Authority (IANA) regarding registration of values related to the EAP
   protocol, in accordance with BCP 26, [RFC2434].

   There are two name spaces in EAP that require registration: Packet
   Codes and method Types.

   EAP is not intended as a general-purpose protocol, and allocations
   SHOULD NOT be made for purposes unrelated to authentication.

   The following terms are used here with the meanings defined in BCP
   26: "name space", "assigned value", "registration".

   The following policies are used here with the meanings defined in BCP
   26: "Private Use", "First Come First Served", "Expert Review",
   "Specification Required", "IETF Consensus", "Standards Action".



Aboba, et al.               Standards Track                    [Page 40]

RFC 3748                          EAP                          June 2004


   For registration requests where a Designated Expert should be
   consulted, the responsible IESG area director should appoint the
   Designated Expert.  The intention is that any allocation will be
   accompanied by a published RFC.  But in order to allow for the
   allocation of values prior to the RFC being approved for publication,
   the Designated Expert can approve allocations once it seems clear
   that an RFC will be published.  The Designated expert will post a
   request to the EAP WG mailing list (or a successor designated by the
   Area Director) for comment and review, including an Internet-Draft.
   Before a period of 30 days has passed, the Designated Expert will
   either approve or deny the registration request and publish a notice
   of the decision to the EAP WG mailing list or its successor, as well
   as informing IANA.  A denial notice must be justified by an
   explanation, and in the cases where it is possible, concrete
   suggestions on how the request can be modified so as to become
   acceptable should be provided.

6.1.  Packet Codes



   Packet Codes have a range from 1 to 255, of which 1-4 have been
   allocated.  Because a new Packet Code has considerable impact on
   interoperability, a new Packet Code requires Standards Action, and
   should be allocated starting at 5.

6.2.  Method Types



   The original EAP method Type space has a range from 1 to 255, and is
   the scarcest resource in EAP, and thus must be allocated with care.
   Method Types 1-45 have been allocated, with 20 available for re-use.
   Method Types 20 and 46-191 may be allocated on the advice of a
   Designated Expert, with Specification Required.

   Allocation of blocks of method Types (more than one for a given
   purpose) should require IETF Consensus.  EAP Type Values 192-253 are
   reserved and allocation requires Standards Action.

   Method Type 254 is allocated for the Expanded Type.  Where the
   Vendor-Id field is non-zero, the Expanded Type is used for functions
   specific only to one vendor's implementation of EAP, where no
   interoperability is deemed useful.  When used with a Vendor-Id of
   zero, method Type 254 can also be used to provide for an expanded
   IETF method Type space.  Method Type values 256-4294967295 may be
   allocated after Type values 1-191 have been allocated, on the advice
   of a Designated Expert, with Specification Required.

   Method Type 255 is allocated for Experimental use, such as testing of
   new EAP methods before a permanent Type is allocated.




Aboba, et al.               Standards Track                    [Page 41]

RFC 3748                          EAP                          June 2004


7.  Security Considerations



   This section defines a generic threat model as well as the EAP method
   security claims mitigating those threats.

   It is expected that the generic threat model and corresponding
   security claims will used to define EAP method requirements for use
   in specific environments.  An example of such a requirements analysis
   is provided in [IEEE-802.11i-req].  A security claims section is
   required in EAP method specifications, so that EAP methods can be
   evaluated against the requirements.

7.1.  Threat Model



   EAP was developed for use with PPP [RFC1661] and was later adapted
   for use in wired IEEE 802 networks [IEEE-802] in [IEEE-802.1X].
   Subsequently, EAP has been proposed for use on wireless LAN networks
   and over the Internet.  In all these situations, it is possible for
   an attacker to gain access to links over which EAP packets are
   transmitted.  For example, attacks on telephone infrastructure are
   documented in [DECEPTION].

   An attacker with access to the link may carry out a number of
   attacks, including:

   [1]  An attacker may try to discover user identities by snooping
        authentication traffic.

   [2]  An attacker may try to modify or spoof EAP packets.

   [3]  An attacker may launch denial of service attacks by spoofing
        lower layer indications or Success/Failure packets, by replaying
        EAP packets, or by generating packets with overlapping
        Identifiers.

   [4]  An attacker may attempt to recover the pass-phrase by mounting
        an offline dictionary attack.

   [5]  An attacker may attempt to convince the peer to connect to an
        untrusted network by mounting a man-in-the-middle attack.

   [6]  An attacker may attempt to disrupt the EAP negotiation in order
        cause a weak authentication method to be selected.

   [7]  An attacker may attempt to recover keys by taking advantage of
        weak key derivation techniques used within EAP methods.





Aboba, et al.               Standards Track                    [Page 42]

RFC 3748                          EAP                          June 2004


   [8]  An attacker may attempt to take advantage of weak ciphersuites
        subsequently used after the EAP conversation is complete.

   [9]  An attacker may attempt to perform downgrading attacks on lower
        layer ciphersuite negotiation in order to ensure that a weaker
        ciphersuite is used subsequently to EAP authentication.

   [10] An attacker acting as an authenticator may provide incorrect
        information to the EAP peer and/or server via out-of-band
        mechanisms (such as via a AAA or lower layer protocol).  This
        includes impersonating another authenticator, or providing
        inconsistent information to the peer and EAP server.

   Depending on the lower layer, these attacks may be carried out
   without requiring physical proximity.  Where EAP is used over
   wireless networks, EAP packets may be forwarded by authenticators
   (e.g., pre-authentication) so that the attacker need not be within
   the coverage area of an authenticator in order to carry out an attack
   on it or its peers.  Where EAP is used over the Internet, attacks may
   be carried out at an even greater distance.

7.2.  Security Claims



   In order to clearly articulate the security provided by an EAP
   method, EAP method specifications MUST include a Security Claims
   section, including the following declarations:

   [a] Mechanism.  This is a statement of the authentication technology:
       certificates, pre-shared keys, passwords, token cards, etc.

   [b] Security claims.  This is a statement of the claimed security
       properties of the method, using terms defined in Section 7.2.1:
       mutual authentication, integrity protection, replay protection,
       confidentiality, key derivation, dictionary attack resistance,
       fast reconnect, cryptographic binding.  The Security Claims
       section of an EAP method specification SHOULD provide
       justification for the claims that are made.  This can be
       accomplished by including a proof in an Appendix, or including a
       reference to a proof.

   [c] Key strength.  If the method derives keys, then the effective key
       strength MUST be estimated.  This estimate is meant for potential
       users of the method to determine if the keys produced are strong
       enough for the intended application.







Aboba, et al.               Standards Track                    [Page 43]

RFC 3748                          EAP                          June 2004


       The effective key strength SHOULD be stated as a number of bits,
       defined as follows: If the effective key strength is N bits, the
       best currently known methods to recover the key (with non-
       negligible probability) require, on average, an effort comparable
       to 2^(N-1) operations of a typical block cipher.  The statement
       SHOULD be accompanied by a short rationale, explaining how this
       number was derived.  This explanation SHOULD include the
       parameters required to achieve the stated key strength based on
       current knowledge of the algorithms.

       (Note: Although it is difficult to define what "comparable
       effort" and "typical block cipher" exactly mean, reasonable
       approximations are sufficient here.  Refer to e.g. [SILVERMAN]
       for more discussion.)

       The key strength depends on the methods used to derive the keys.
       For instance, if keys are derived from a shared secret (such as a
       password or a long-term secret), and possibly some public
       information such as nonces, the effective key strength is limited
       by the strength of the long-term secret (assuming that the
       derivation procedure is computationally simple).  To take another
       example, when using public key algorithms, the strength of the
       symmetric key depends on the strength of the public keys used.

   [d] Description of key hierarchy.  EAP methods deriving keys MUST
       either provide a reference to a key hierarchy specification, or
       describe how Master Session Keys (MSKs) and Extended Master
       Session Keys (EMSKs) are to be derived.

   [e] Indication of vulnerabilities.  In addition to the security
       claims that are made, the specification MUST indicate which of
       the security claims detailed in Section 7.2.1 are NOT being made.

7.2.1.  Security Claims Terminology for EAP Methods



   These terms are used to describe the security properties of EAP
   methods:

   Protected ciphersuite negotiation
      This refers to the ability of an EAP method to negotiate the
      ciphersuite used to protect the EAP conversation, as well as to
      integrity protect the negotiation.  It does not refer to the
      ability to negotiate the ciphersuite used to protect data.








Aboba, et al.               Standards Track                    [Page 44]

RFC 3748                          EAP                          June 2004


   Mutual authentication
      This refers to an EAP method in which, within an interlocked
      exchange, the authenticator authenticates the peer and the peer
      authenticates the authenticator.  Two independent one-way methods,
      running in opposite directions do not provide mutual
      authentication as defined here.

   Integrity protection
      This refers to providing data origin authentication and protection
      against unauthorized modification of information for EAP packets
      (including EAP Requests and Responses).  When making this claim, a
      method specification MUST describe the EAP packets and fields
      within the EAP packet that are protected.

   Replay protection
      This refers to protection against replay of an EAP method or its
      messages, including success and failure result indications.

   Confidentiality
      This refers to encryption of EAP messages, including EAP Requests
      and Responses, and success and failure result indications.  A
      method making this claim MUST support identity protection (see
      Section 7.3).

   Key derivation
      This refers to the ability of the EAP method to derive exportable
      keying material, such as the Master Session Key (MSK), and
      Extended Master Session Key (EMSK).  The MSK is used only for
      further key derivation, not directly for protection of the EAP
      conversation or subsequent data.  Use of the EMSK is reserved.

   Key strength
      If the effective key strength is N bits, the best currently known
      methods to recover the key (with non-negligible probability)
      require, on average, an effort comparable to 2^(N-1) operations of
      a typical block cipher.

   Dictionary attack resistance
      Where password authentication is used, passwords are commonly
      selected from a small set (as compared to a set of N-bit keys),
      which raises a concern about dictionary attacks.  A method may be
      said to provide protection against dictionary attacks if, when it
      uses a password as a secret, the method does not allow an offline
      attack that has a work factor based on the number of passwords in
      an attacker's dictionary.






Aboba, et al.               Standards Track                    [Page 45]

RFC 3748                          EAP                          June 2004


   Fast reconnect
      The ability, in the case where a security association has been
      previously established, to create a new or refreshed security
      association more efficiently or in a smaller number of round-
      trips.

   Cryptographic binding
      The demonstration of the EAP peer to the EAP server that a single
      entity has acted as the EAP peer for all methods executed within a
      tunnel method.  Binding MAY also imply that the EAP server
      demonstrates to the peer that a single entity has acted as the EAP
      server for all methods executed within a tunnel method.  If
      executed correctly, binding serves to mitigate man-in-the-middle
      vulnerabilities.

   Session independence
      The demonstration that passive attacks (such as capture of the EAP
      conversation) or active attacks (including compromise of the MSK
      or EMSK) does not enable compromise of subsequent or prior MSKs or
      EMSKs.

   Fragmentation
      This refers to whether an EAP method supports fragmentation and
      reassembly.  As noted in Section 3.1, EAP methods should support
      fragmentation and reassembly if EAP packets can exceed the minimum
      MTU of 1020 octets.

   Channel binding
      The communication within an EAP method of integrity-protected
      channel properties such as endpoint identifiers which can be
      compared to values communicated via out of band mechanisms (such
      as via a AAA or lower layer protocol).

   Note: This list of security claims is not exhaustive.  Additional
   properties, such as additional denial-of-service protection, may be
   relevant as well.

7.3.  Identity Protection



   An Identity exchange is optional within the EAP conversation.
   Therefore, it is possible to omit the Identity exchange entirely, or
   to use a method-specific identity exchange once a protected channel
   has been established.

   However, where roaming is supported as described in [RFC2607], it may
   be necessary to locate the appropriate backend authentication server
   before the authentication conversation can proceed.  The realm
   portion of the Network Access Identifier (NAI) [RFC2486] is typically



Aboba, et al.               Standards Track                    [Page 46]

RFC 3748                          EAP                          June 2004


   included within the EAP-Response/Identity in order to enable the
   authentication exchange to be routed to the appropriate backend
   authentication server.  Therefore, while the peer-name portion of the
   NAI may be omitted in the EAP-Response/Identity where proxies or
   relays are present, the realm portion may be required.

   It is possible for the identity in the identity response to be
   different from the identity authenticated by the EAP method.  This
   may be intentional in the case of identity privacy.  An EAP method
   SHOULD use the authenticated identity when making access control
   decisions.

7.4.  Man-in-the-Middle Attacks



   Where EAP is tunneled within another protocol that omits peer
   authentication, there exists a potential vulnerability to a man-in-
   the-middle attack.  For details, see [BINDING] and [MITM].

   As noted in Section 2.1, EAP does not permit untunneled sequences of
   authentication methods.  Were a sequence of EAP authentication
   methods to be permitted, the peer might not have proof that a single
   entity has acted as the authenticator for all EAP methods within the
   sequence.  For example, an authenticator might terminate one EAP
   method, then forward the next method in the sequence to another party
   without the peer's knowledge or consent.  Similarly, the
   authenticator might not have proof that a single entity has acted as
   the peer for all EAP methods within the sequence.

   Tunneling EAP within another protocol enables an attack by a rogue
   EAP authenticator tunneling EAP to a legitimate server.  Where the
   tunneling protocol is used for key establishment but does not require
   peer authentication, an attacker convincing a legitimate peer to
   connect to it will be able to tunnel EAP packets to a legitimate
   server, successfully authenticating and obtaining the key.  This
   allows the attacker to successfully establish itself as a man-in-
   the-middle, gaining access to the network, as well as the ability to
   decrypt data traffic between the legitimate peer and server.

   This attack may be mitigated by the following measures:

   [a] Requiring mutual authentication within EAP tunneling mechanisms.

   [b] Requiring cryptographic binding between the EAP tunneling
       protocol and the tunneled EAP methods.  Where cryptographic
       binding is supported, a mechanism is also needed to protect
       against downgrade attacks that would bypass it.  For further
       details on cryptographic binding, see [BINDING].




Aboba, et al.               Standards Track                    [Page 47]

RFC 3748                          EAP                          June 2004


   [c] Limiting the EAP methods authorized for use without protection,
       based on peer and authenticator policy.

   [d] Avoiding the use of tunnels when a single, strong method is
       available.

7.5.  Packet Modification Attacks



   While EAP methods may support per-packet data origin authentication,
   integrity, and replay protection, support is not provided within the
   EAP layer.

   Since the Identifier is only a single octet, it is easy to guess,
   allowing an attacker to successfully inject or replay EAP packets.
   An attacker may also modify EAP headers (Code, Identifier, Length,
   Type) within EAP packets where the header is unprotected.  This could
   cause packets to be inappropriately discarded or misinterpreted.

   To protect EAP packets against modification, spoofing, or replay,
   methods supporting protected ciphersuite negotiation, mutual
   authentication, and key derivation, as well as integrity and replay
   protection, are recommended.  See Section 7.2.1 for definitions of
   these security claims.

   Method-specific MICs may be used to provide protection.  If a per-
   packet MIC is employed within an EAP method, then peers,
   authentication servers, and authenticators not operating in pass-
   through mode MUST validate the MIC.  MIC validation failures SHOULD
   be logged.  Whether a MIC validation failure is considered a fatal
   error or not is determined by the EAP method specification.

   It is RECOMMENDED that methods providing integrity protection of EAP
   packets include coverage of all the EAP header fields, including the
   Code, Identifier, Length, Type, and Type-Data fields.

   Since EAP messages of Types Identity, Notification, and Nak do not
   include their own MIC, it may be desirable for the EAP method MIC to
   cover information contained within these messages, as well as the
   header of each EAP message.

   To provide protection, EAP also may be encapsulated within a
   protected channel created by protocols such as ISAKMP [RFC2408], as
   is done in [IKEv2] or within TLS [RFC2246].  However, as noted in
   Section 7.4, EAP tunneling may result in a man-in-the-middle
   vulnerability.






Aboba, et al.               Standards Track                    [Page 48]

RFC 3748                          EAP                          June 2004


   Existing EAP methods define message integrity checks (MICs) that
   cover more than one EAP packet.  For example, EAP-TLS [RFC2716]
   defines a MIC over a TLS record that could be split into multiple
   fragments; within the FINISHED message, the MIC is computed over
   previous messages.  Where the MIC covers more than one EAP packet, a
   MIC validation failure is typically considered a fatal error.

   Within EAP-TLS [RFC2716], a MIC validation failure is treated as a
   fatal error, since that is what is specified in TLS [RFC2246].
   However, it is also possible to develop EAP methods that support
   per-packet MICs, and respond to verification failures by silently
   discarding the offending packet.

   In this document, descriptions of EAP message handling assume that
   per-packet MIC validation, where it occurs, is effectively performed
   as though it occurs before sending any responses or changing the
   state of the host which received the packet.

7.6.  Dictionary Attacks



   Password authentication algorithms such as EAP-MD5, MS-CHAPv1
   [RFC2433], and Kerberos V [RFC1510] are known to be vulnerable to
   dictionary attacks.  MS-CHAPv1 vulnerabilities are documented in
   [PPTPv1]; MS-CHAPv2 vulnerabilities are documented in [PPTPv2];
   Kerberos vulnerabilities are described in [KRBATTACK], [KRBLIM], and
   [KERB4WEAK].

   In order to protect against dictionary attacks, authentication
   methods resistant to dictionary attacks (as defined in Section 7.2.1)
   are recommended.

   If an authentication algorithm is used that is known to be vulnerable
   to dictionary attacks, then the conversation may be tunneled within a
   protected channel in order to provide additional protection.
   However, as noted in Section 7.4, EAP tunneling may result in a man-
   in-the-middle vulnerability, and therefore dictionary attack
   resistant methods are preferred.

7.7.  Connection to an Untrusted Network



   With EAP methods supporting one-way authentication, such as EAP-MD5,
   the peer does not authenticate the authenticator, making the peer
   vulnerable to attack by a rogue authenticator.  Methods supporting
   mutual authentication (as defined in Section 7.2.1) address this
   vulnerability.

   In EAP there is no requirement that authentication be full duplex or
   that the same protocol be used in both directions.  It is perfectly



Aboba, et al.               Standards Track                    [Page 49]

RFC 3748                          EAP                          June 2004


   acceptable for different protocols to be used in each direction.
   This will, of course, depend on the specific protocols negotiated.
   However, in general, completing a single unitary mutual
   authentication is preferable to two one-way authentications, one in
   each direction.  This is because separate authentications that are
   not bound cryptographically so as to demonstrate they are part of the
   same session are subject to man-in-the-middle attacks, as discussed
   in Section 7.4.

7.8.  Negotiation Attacks



   In a negotiation attack, the attacker attempts to convince the peer
   and authenticator to negotiate a less secure EAP method.  EAP does
   not provide protection for Nak Response packets, although it is
   possible for a method to include coverage of Nak Responses within a
   method-specific MIC.

   Within or associated with each authenticator, it is not anticipated
   that a particular named peer will support a choice of methods.  This
   would make the peer vulnerable to attacks that negotiate the least
   secure method from among a set.  Instead, for each named peer, there
   SHOULD be an indication of exactly one method used to authenticate
   that peer name.  If a peer needs to make use of different
   authentication methods under different circumstances, then distinct
   identities SHOULD be employed, each of which identifies exactly one
   authentication method.

7.9.  Implementation Idiosyncrasies



   The interaction of EAP with lower layers such as PPP and IEEE 802 are
   highly implementation dependent.

   For example, upon failure of authentication, some PPP implementations
   do not terminate the link, instead limiting traffic in Network-Layer
   Protocols to a filtered subset, which in turn allows the peer the
   opportunity to update secrets or send mail to the network
   administrator indicating a problem.  Similarly, while an
   authentication failure will result in denied access to the controlled
   port in [IEEE-802.1X], limited traffic may be permitted on the
   uncontrolled port.

   In EAP there is no provision for retries of failed authentication.
   However, in PPP the LCP state machine can renegotiate the
   authentication protocol at any time, thus allowing a new attempt.
   Similarly, in IEEE 802.1X the Supplicant or Authenticator can re-
   authenticate at any time.  It is recommended that any counters used
   for authentication failure not be reset until after successful
   authentication, or subsequent termination of the failed link.



Aboba, et al.               Standards Track                    [Page 50]

RFC 3748                          EAP                          June 2004


7.10.  Key Derivation



   It is possible for the peer and EAP server to mutually authenticate
   and derive keys.  In order to provide keying material for use in a
   subsequently negotiated ciphersuite, an EAP method supporting key
   derivation MUST export a Master Session Key (MSK) of at least 64
   octets, and an Extended Master Session Key (EMSK) of at least 64
   octets.  EAP Methods deriving keys MUST provide for mutual
   authentication between the EAP peer and the EAP Server.

   The MSK and EMSK MUST NOT be used directly to protect data; however,
   they are of sufficient size to enable derivation of a AAA-Key
   subsequently used to derive Transient Session Keys (TSKs) for use
   with the selected ciphersuite.  Each ciphersuite is responsible for
   specifying how to derive the TSKs from the AAA-Key.

   The AAA-Key is derived from the keying material exported by the EAP
   method (MSK and EMSK).  This derivation occurs on the AAA server.  In
   many existing protocols that use EAP, the AAA-Key and MSK are
   equivalent, but more complicated mechanisms are possible (see
   [KEYFRAME] for details).

   EAP methods SHOULD ensure the freshness of the MSK and EMSK, even in
   cases where one party may not have a high quality random number
   generator.  A RECOMMENDED method is for each party to provide a nonce
   of at least 128 bits, used in the derivation of the MSK and EMSK.

   EAP methods export the MSK and EMSK, but not Transient Session Keys
   so as to allow EAP methods to be ciphersuite and media independent.
   Keying material exported by EAP methods MUST be independent of the
   ciphersuite negotiated to protect data.

   Depending on the lower layer, EAP methods may run before or after
   ciphersuite negotiation, so that the selected ciphersuite may not be
   known to the EAP method.  By providing keying material usable with
   any ciphersuite, EAP methods can used with a wide range of
   ciphersuites and media.

   In order to preserve algorithm independence, EAP methods deriving
   keys SHOULD support (and document) the protected negotiation of the
   ciphersuite used to protect the EAP conversation between the peer and
   server.  This is distinct from the ciphersuite negotiated between the
   peer and authenticator, used to protect data.

   The strength of Transient Session Keys (TSKs) used to protect data is
   ultimately dependent on the strength of keys generated by the EAP
   method.  If an EAP method cannot produce keying material of
   sufficient strength, then the TSKs may be subject to a brute force



Aboba, et al.               Standards Track                    [Page 51]

RFC 3748                          EAP                          June 2004


   attack.  In order to enable deployments requiring strong keys, EAP
   methods supporting key derivation SHOULD be capable of generating an
   MSK and EMSK, each with an effective key strength of at least 128
   bits.

   Methods supporting key derivation MUST demonstrate cryptographic
   separation between the MSK and EMSK branches of the EAP key
   hierarchy.  Without violating a fundamental cryptographic assumption
   (such as the non-invertibility of a one-way function), an attacker
   recovering the MSK or EMSK MUST NOT be able to recover the other
   quantity with a level of effort less than brute force.

   Non-overlapping substrings of the MSK MUST be cryptographically
   separate from each other, as defined in Section 7.2.1.  That is,
   knowledge of one substring MUST NOT help in recovering some other
   substring without breaking some hard cryptographic assumption.  This
   is required because some existing ciphersuites form TSKs by simply
   splitting the AAA-Key to pieces of appropriate length.  Likewise,
   non-overlapping substrings of the EMSK MUST be cryptographically
   separate from each other, and from substrings of the MSK.

   The EMSK is reserved for future use and MUST remain on the EAP peer
   and EAP server where it is derived; it MUST NOT be transported to, or
   shared with, additional parties, or used to derive any other keys.
   (This restriction will be relaxed in a future document that specifies
   how the EMSK can be used.)

   Since EAP does not provide for explicit key lifetime negotiation, EAP
   peers, authenticators, and authentication servers MUST be prepared
   for situations in which one of the parties discards the key state,
   which remains valid on another party.

   This specification does not provide detailed guidance on how EAP
   methods derive the MSK and EMSK, how the AAA-Key is derived from the
   MSK and/or EMSK, or how the TSKs are derived from the AAA-Key.

   The development and validation of key derivation algorithms is
   difficult, and as a result, EAP methods SHOULD re-use well
   established and analyzed mechanisms for key derivation (such as those
   specified in IKE [RFC2409] or TLS [RFC2246]), rather than inventing
   new ones. EAP methods SHOULD also utilize well established and
   analyzed mechanisms for MSK and EMSK derivation.  Further details on
   EAP Key Derivation are provided within [KEYFRAME].








Aboba, et al.               Standards Track                    [Page 52]

RFC 3748                          EAP                          June 2004


7.11.  Weak Ciphersuites



   If after the initial EAP authentication, data packets are sent
   without per-packet authentication, integrity, and replay protection,
   an attacker with access to the media can inject packets, "flip bits"
   within existing packets, replay packets, or even hijack the session
   completely.  Without per-packet confidentiality, it is possible to
   snoop data packets.

   To protect against data modification, spoofing, or snooping, it is
   recommended that EAP methods supporting mutual authentication and key
   derivation (as defined by Section 7.2.1) be used, along with lower
   layers providing per-packet confidentiality, authentication,
   integrity, and replay protection.

   Additionally, if the lower layer performs ciphersuite negotiation, it
   should be understood that EAP does not provide by itself integrity
   protection of that negotiation.  Therefore, in order to avoid
   downgrading attacks which would lead to weaker ciphersuites being
   used, clients implementing lower layer ciphersuite negotiation SHOULD
   protect against negotiation downgrading.

   This can be done by enabling users to configure which ciphersuites
   are acceptable as a matter of security policy, or the ciphersuite
   negotiation MAY be authenticated using keying material derived from
   the EAP authentication and a MIC algorithm agreed upon in advance by
   lower-layer peers.

7.12.  Link Layer



   There are reliability and security issues with link layer indications
   in PPP, IEEE 802 LANs, and IEEE 802.11 wireless LANs:

   [a] PPP.  In PPP, link layer indications such as LCP-Terminate (a
       link failure indication) and NCP (a link success indication) are
       not authenticated or integrity protected.  They can therefore be
       spoofed by an attacker with access to the link.

   [b] IEEE 802.  IEEE 802.1X EAPOL-Start and EAPOL-Logoff frames are
       not authenticated or integrity protected.  They can therefore be
       spoofed by an attacker with access to the link.

   [c] IEEE 802.11.  In IEEE 802.11, link layer indications include
       Disassociate and Deauthenticate frames (link failure
       indications), and the first message of the 4-way handshake (link
       success indication).  These messages are not authenticated or
       integrity protected, and although they are not forwardable, they
       are spoofable by an attacker within range.



Aboba, et al.               Standards Track                    [Page 53]

RFC 3748                          EAP                          June 2004


   In IEEE 802.11, IEEE 802.1X data frames may be sent as Class 3
   unicast data frames, and are therefore forwardable.  This implies
   that while EAPOL-Start and EAPOL-Logoff messages may be authenticated
   and integrity protected, they can be spoofed by an authenticated
   attacker far from the target when "pre-authentication" is enabled.

   In IEEE 802.11, a "link down" indication is an unreliable indication
   of link failure, since wireless signal strength can come and go and
   may be influenced by radio frequency interference generated by an
   attacker.  To avoid unnecessary resets, it is advisable to damp these
   indications, rather than passing them directly to the EAP.  Since EAP
   supports retransmission, it is robust against transient connectivity
   losses.

7.13.  Separation of Authenticator and Backend Authentication Server



   It is possible for the EAP peer and EAP server to mutually
   authenticate and derive a AAA-Key for a ciphersuite used to protect
   subsequent data traffic.  This does not present an issue on the peer,
   since the peer and EAP client reside on the same machine; all that is
   required is for the client to derive the AAA-Key from the MSK and
   EMSK exported by the EAP method, and to subsequently pass a Transient
   Session Key (TSK) to the ciphersuite module.

   However, in the case where the authenticator and authentication
   server reside on different machines, there are several implications
   for security.

   [a] Authentication will occur between the peer and the authentication
       server, not between the peer and the authenticator.  This means
       that it is not possible for the peer to validate the identity of
       the authenticator that it is speaking to, using EAP alone.

   [b] As discussed in [RFC3579], the authenticator is dependent on the
       AAA protocol in order to know the outcome of an authentication
       conversation, and does not look at the encapsulated EAP packet
       (if one is present) to determine the outcome.  In practice, this
       implies that the AAA protocol spoken between the authenticator
       and authentication server MUST support per-packet authentication,
       integrity, and replay protection.

   [c] After completion of the EAP conversation, where lower layer
       security services such as per-packet confidentiality,
       authentication, integrity, and replay protection will be enabled,
       a secure association protocol SHOULD be run between the peer and
       authenticator in order to provide mutual authentication between





Aboba, et al.               Standards Track                    [Page 54]

RFC 3748                          EAP                          June 2004


       the peer and authenticator, guarantee liveness of transient
       session keys, provide protected ciphersuite and capabilities
       negotiation for subsequent data, and synchronize key usage.

   [d] A AAA-Key derived from the MSK and/or EMSK negotiated between the
       peer and authentication server MAY be transmitted to the
       authenticator.  Therefore, a mechanism needs to be provided to
       transmit the AAA-Key from the authentication server to the
       authenticator that needs it.  The specification of the AAA-key
       derivation, transport, and wrapping mechanisms is outside the
       scope of this document.  Further details on AAA-Key Derivation
       are provided within [KEYFRAME].

7.14.  Cleartext Passwords



   This specification does not define a mechanism for cleartext password
   authentication.  The omission is intentional.  Use of cleartext
   passwords would allow the password to be captured by an attacker with
   access to a link over which EAP packets are transmitted.

   Since protocols encapsulating EAP, such as RADIUS [RFC3579], may not
   provide confidentiality, EAP packets may be subsequently encapsulated
   for transport over the Internet where they may be captured by an
   attacker.

   As a result, cleartext passwords cannot be securely used within EAP,
   except where encapsulated within a protected tunnel with server
   authentication.  Some of the same risks apply to EAP methods without
   dictionary attack resistance, as defined in Section 7.2.1.  For
   details, see Section 7.6.

7.15.  Channel Binding



   It is possible for a compromised or poorly implemented EAP
   authenticator to communicate incorrect information to the EAP peer
   and/or server.  This may enable an authenticator to impersonate
   another authenticator or communicate incorrect information via out-
   of-band mechanisms (such as via a AAA or lower layer protocol).

   Where EAP is used in pass-through mode, the EAP peer typically does
   not verify the identity of the pass-through authenticator, it only
   verifies that the pass-through authenticator is trusted by the EAP
   server.  This creates a potential security vulnerability.

   Section 4.3.7 of [RFC3579] describes how an EAP pass-through
   authenticator acting as a AAA client can be detected if it attempts
   to impersonate another authenticator (such by sending incorrect NAS-
   Identifier [RFC2865], NAS-IP-Address [RFC2865] or NAS-IPv6-Address



Aboba, et al.               Standards Track                    [Page 55]

RFC 3748                          EAP                          June 2004


   [RFC3162] attributes via the AAA protocol).  However, it is possible
   for a pass-through authenticator acting as a AAA client to provide
   correct information to the AAA server while communicating misleading
   information to the EAP peer via a lower layer protocol.

   For example, it is possible for a compromised authenticator to
   utilize another authenticator's Called-Station-Id or NAS-Identifier
   in communicating with the EAP peer via a lower layer protocol, or for
   a pass-through authenticator acting as a AAA client to provide an
   incorrect peer Calling-Station-Id [RFC2865][RFC3580] to the AAA
   server via the AAA protocol.

   In order to address this vulnerability, EAP methods may support a
   protected exchange of channel properties such as endpoint
   identifiers, including (but not limited to): Called-Station-Id
   [RFC2865][RFC3580], Calling-Station-Id [RFC2865][RFC3580], NAS-
   Identifier [RFC2865], NAS-IP-Address [RFC2865], and NAS-IPv6-Address
   [RFC3162].

   Using such a protected exchange, it is possible to match the channel
   properties provided by the authenticator via out-of-band mechanisms
   against those exchanged within the EAP method.  Where discrepancies
   are found, these SHOULD be logged; additional actions MAY also be
   taken, such as denying access.

7.16.  Protected Result Indications



   Within EAP, Success and Failure packets are neither acknowledged nor
   integrity protected.  Result indications improve resilience to loss
   of Success and Failure packets when EAP is run over lower layers
   which do not support retransmission or synchronization of the
   authentication state.  In media such as IEEE 802.11, which provides
   for retransmission, as well as synchronization of authentication
   state via the 4-way handshake defined in [IEEE-802.11i], additional
   resilience is typically of marginal benefit.

   Depending on the method and circumstances, result indications can be
   spoofable by an attacker.  A method is said to provide protected
   result indications if it supports result indications, as well as the
   "integrity protection" and "replay protection" claims.  A method
   supporting protected result indications MUST indicate which result
   indications are protected, and which are not.

   Protected result indications are not required to protect against
   rogue authenticators.  Within a mutually authenticating method,
   requiring that the server authenticate to the peer before the peer
   will accept a Success packet prevents an attacker from acting as a
   rogue authenticator.



Aboba, et al.               Standards Track                    [Page 56]

RFC 3748                          EAP                          June 2004


   However, it is possible for an attacker to forge a Success packet
   after the server has authenticated to the peer, but before the peer
   has authenticated to the server.  If the peer were to accept the
   forged Success packet and attempt to access the network when it had
   not yet successfully authenticated to the server, a denial of service
   attack could be mounted against the peer.  After such an attack, if
   the lower layer supports failure indications, the authenticator can
   synchronize state with the peer by providing a lower layer failure
   indication.  See Section 7.12 for details.

   If a server were to authenticate the peer and send a Success packet
   prior to determining whether the peer has authenticated the
   authenticator, an idle timeout can occur if the authenticator is not
   authenticated by the peer.  Where supported by the lower layer, an
   authenticator sensing the absence of the peer can free resources.

   In a method supporting result indications, a peer that has
   authenticated the server does not consider the authentication
   successful until it receives an indication that the server
   successfully authenticated it.  Similarly, a server that has
   successfully authenticated the peer does not consider the
   authentication successful until it receives an indication that the
   peer has authenticated the server.

   In order to avoid synchronization problems, prior to sending a
   success result indication, it is desirable for the sender to verify
   that sufficient authorization exists for granting access, though, as
   discussed below, this is not always possible.

   While result indications may enable synchronization of the
   authentication result between the peer and server, this does not
   guarantee that the peer and authenticator will be synchronized in
   terms of their authorization or that timeouts will not occur.  For
   example, the EAP server may not be aware of an authorization decision
   made by a AAA proxy; the AAA server may check authorization only
   after authentication has completed successfully, to discover that
   authorization cannot be granted, or the AAA server may grant access
   but the authenticator may be unable to provide it due to a temporary
   lack of resources.  In these situations, synchronization may only be
   achieved via lower layer result indications.

   Success indications may be explicit or implicit.  For example, where
   a method supports error messages, an implicit success indication may
   be defined as the reception of a specific message without a preceding
   error message.  Failures are typically indicated explicitly.  As
   described in Section 4.2, a peer silently discards a Failure packet
   received at a point where the method does not explicitly permit this




Aboba, et al.               Standards Track                    [Page 57]

RFC 3748                          EAP                          June 2004


   to be sent.  For example, a method providing its own error messages
   might require the peer to receive an error message prior to accepting
   a Failure packet.

   Per-packet authentication, integrity, and replay protection of result
   indications protects against spoofing.  Since protected result
   indications require use of a key for per-packet authentication and
   integrity protection, methods supporting protected result indications
   MUST also support the "key derivation", "mutual authentication",
   "integrity protection", and "replay protection" claims.

   Protected result indications address some denial-of-service
   vulnerabilities due to spoofing of Success and Failure packets,
   though not all.  EAP methods can typically provide protected result
   indications only in some circumstances.  For example, errors can
   occur prior to key derivation, and so it may not be possible to
   protect all failure indications.  It is also possible that result
   indications may not be supported in both directions or that
   synchronization may not be achieved in all modes of operation.

   For example, within EAP-TLS [RFC2716], in the client authentication
   handshake, the server authenticates the peer, but does not receive a
   protected indication of whether the peer has authenticated it.  In
   contrast, the peer authenticates the server and is aware of whether
   the server has authenticated it.  In the session resumption
   handshake, the peer authenticates the server, but does not receive a
   protected indication of whether the server has authenticated it.  In
   this mode, the server authenticates the peer and is aware of whether
   the peer has authenticated it.

8.  Acknowledgements



   This protocol derives much of its inspiration from Dave Carrel's AHA
   document, as well as the PPP CHAP protocol [RFC1994].  Valuable
   feedback was provided by Yoshihiro Ohba of Toshiba America Research,
   Jari Arkko of Ericsson, Sachin Seth of Microsoft, Glen Zorn of Cisco
   Systems, Jesse Walker of Intel, Bill Arbaugh, Nick Petroni and Bryan
   Payne of the University of Maryland, Steve Bellovin of AT&T Research,
   Paul Funk of Funk Software, Pasi Eronen of Nokia, Joseph Salowey of
   Cisco, Paul Congdon of HP, and members of the EAP working group.

   The use of Security Claims sections for EAP methods, as required by
   Section 7.2 and specified for each EAP method described in this
   document, was inspired by Glen Zorn through [EAP-EVAL].







Aboba, et al.               Standards Track                    [Page 58]

RFC 3748                          EAP                          June 2004


9.  References



9.1.  Normative References



   [RFC1661]          Simpson, W., "The Point-to-Point Protocol (PPP)",
                      STD 51, RFC 1661, July 1994.

   [RFC1994]          Simpson, W., "PPP Challenge Handshake
                      Authentication Protocol (CHAP)", RFC 1994, August
                      1996.

   [RFC2119]          Bradner, S., "Key words for use in RFCs to
                      Indicate Requirement Levels", BCP 14, RFC 2119,
                      March 1997.

   [RFC2243]          Metz, C., "OTP Extended Responses", RFC 2243,
                      November 1997.

   [RFC2279]          Yergeau, F., "UTF-8, a transformation format of
                      ISO 10646", RFC 2279, January 1998.

   [RFC2289]          Haller, N., Metz, C., Nesser, P. and M. Straw, "A
                      One-Time Password System", RFC 2289, February
                      1998.

   [RFC2434]          Narten, T. and H. Alvestrand, "Guidelines for
                      Writing an IANA Considerations Section in RFCs",
                      BCP 26, RFC 2434, October 1998.

   [RFC2988]          Paxson, V. and M. Allman, "Computing TCP's
                      Retransmission Timer", RFC 2988, November 2000.

   [IEEE-802]         Institute of Electrical and Electronics Engineers,
                      "Local and Metropolitan Area Networks: Overview
                      and Architecture", IEEE Standard 802, 1990.

   [IEEE-802.1X]      Institute of Electrical and Electronics Engineers,
                      "Local and Metropolitan Area Networks: Port-Based
                      Network Access Control", IEEE Standard 802.1X,
                      September 2001.











Aboba, et al.               Standards Track                    [Page 59]

RFC 3748                          EAP                          June 2004


9.2.  Informative References



   [RFC793]           Postel, J., "Transmission Control Protocol", STD
                      7, RFC 793, September 1981.

   [RFC1510]          Kohl, J. and B. Neuman, "The Kerberos Network
                      Authentication Service (V5)", RFC 1510, September
                      1993.

   [RFC1750]          Eastlake, D., Crocker, S. and J. Schiller,
                      "Randomness Recommendations for Security", RFC
                      1750, December 1994.

   [RFC2246]          Dierks, T., Allen, C., Treese, W., Karlton, P.,
                      Freier, A. and P. Kocher, "The TLS Protocol
                      Version 1.0", RFC 2246, January 1999.

   [RFC2284]          Blunk, L. and J. Vollbrecht, "PPP Extensible
                      Authentication Protocol (EAP)", RFC 2284, March
                      1998.

   [RFC2486]          Aboba, B. and M. Beadles, "The Network Access
                      Identifier", RFC 2486, January 1999.

   [RFC2408]          Maughan, D., Schneider, M. and M. Schertler,
                      "Internet Security Association and Key Management
                      Protocol (ISAKMP)", RFC 2408, November 1998.

   [RFC2409]          Harkins, D. and D. Carrel, "The Internet Key
                      Exchange (IKE)", RFC 2409, November 1998.

   [RFC2433]          Zorn, G. and S. Cobb, "Microsoft PPP CHAP
                      Extensions", RFC 2433, October 1998.

   [RFC2607]          Aboba, B. and J. Vollbrecht, "Proxy Chaining and
                      Policy Implementation in Roaming", RFC 2607, June
                      1999.

   [RFC2661]          Townsley, W., Valencia, A., Rubens, A., Pall, G.,
                      Zorn, G. and B. Palter, "Layer Two Tunneling
                      Protocol "L2TP"", RFC 2661, August 1999.

   [RFC2716]          Aboba, B. and D. Simon, "PPP EAP TLS
                      Authentication Protocol", RFC 2716, October 1999.

   [RFC2865]          Rigney, C., Willens, S., Rubens, A. and W.
                      Simpson, "Remote Authentication Dial In User
                      Service (RADIUS)", RFC 2865, June 2000.



Aboba, et al.               Standards Track                    [Page 60]

RFC 3748                          EAP                          June 2004


   [RFC2960]          Stewart, R., Xie, Q., Morneault, K., Sharp, C.,
                      Schwarzbauer, H., Taylor, T., Rytina, I., Kalla,
                      M., Zhang, L. and V. Paxson, "Stream Control
                      Transmission Protocol", RFC 2960, October 2000.

   [RFC3162]          Aboba, B., Zorn, G. and D. Mitton, "RADIUS and
                      IPv6", RFC 3162, August 2001.

   [RFC3454]          Hoffman, P. and M. Blanchet, "Preparation of
                      Internationalized Strings ("stringprep")", RFC
                      3454, December 2002.

   [RFC3579]          Aboba, B. and P. Calhoun, "RADIUS (Remote
                      Authentication Dial In User Service) Support For
                      Extensible Authentication Protocol (EAP)", RFC
                      3579, September 2003.

   [RFC3580]          Congdon, P., Aboba, B., Smith, A., Zorn, G. and J.
                      Roese, "IEEE 802.1X Remote Authentication Dial In
                      User Service (RADIUS) Usage Guidelines", RFC 3580,
                      September 2003.

   [RFC3692]          Narten, T., "Assigning Experimental and Testing
                      Numbers Considered Useful", BCP 82, RFC 3692,
                      January 2004.

   [DECEPTION]        Slatalla, M. and J. Quittner, "Masters of
                      Deception", Harper-Collins, New York, 1995.

   [KRBATTACK]        Wu, T., "A Real-World Analysis of Kerberos
                      Password Security", Proceedings of the 1999 ISOC
                      Network and Distributed System Security Symposium,
                      http://www.isoc.org/isoc/conferences/ndss/99/
                      proceedings/papers/wu.pdf.

   [KRBLIM]           Bellovin, S. and M. Merrit, "Limitations of the
                      Kerberos authentication system", Proceedings of
                      the 1991 Winter USENIX Conference, pp. 253-267,
                      1991.

   [KERB4WEAK]        Dole, B., Lodin, S. and E. Spafford, "Misplaced
                      trust:  Kerberos 4 session keys", Proceedings of
                      the Internet Society Network and Distributed
                      System Security Symposium, pp. 60-70, March 1997.







Aboba, et al.               Standards Track                    [Page 61]

RFC 3748                          EAP                          June 2004


   [PIC]              Aboba, B., Krawczyk, H. and Y. Sheffer, "PIC, A
                      Pre-IKE Credential Provisioning Protocol", Work in
                      Progress, October 2002.

   [IKEv2]            Kaufman, C., "Internet Key Exchange (IKEv2)
                      Protocol", Work in Progress, January 2004.

   [PPTPv1]           Schneier, B. and Mudge, "Cryptanalysis of
                      Microsoft's Point-to- Point Tunneling Protocol",
                      Proceedings of the 5th ACM Conference on
                      Communications and Computer Security, ACM Press,
                      November 1998.

   [IEEE-802.11]      Institute of Electrical and Electronics Engineers,
                      "Wireless LAN Medium Access Control (MAC) and
                      Physical Layer (PHY) Specifications", IEEE
                      Standard 802.11, 1999.

   [SILVERMAN]        Silverman, Robert D., "A Cost-Based Security
                      Analysis of Symmetric and Asymmetric Key Lengths",
                      RSA Laboratories Bulletin 13, April 2000 (Revised
                      November 2001),
                      http://www.rsasecurity.com/rsalabs/bulletins/
                      bulletin13.html.

   [KEYFRAME]         Aboba, B., "EAP Key Management Framework", Work in
                      Progress, October 2003.

   [SASLPREP]         Zeilenga, K., "SASLprep: Stringprep profile for
                      user names and passwords", Work in Progress, March
                      2004.

   [IEEE-802.11i]     Institute of Electrical and Electronics Engineers,
                      "Unapproved Draft Supplement to Standard for
                      Telecommunications and Information Exchange
                      Between Systems - LAN/MAN Specific Requirements -
                      Part 11: Wireless LAN Medium Access Control (MAC)
                      and Physical Layer (PHY) Specifications:
                      Specification for Enhanced Security", IEEE Draft
                      802.11i (work in progress), 2003.

   [DIAM-EAP]         Eronen, P., Hiller, T. and G. Zorn, "Diameter
                      Extensible Authentication Protocol (EAP)
                      Application", Work in Progress, February 2004.

   [EAP-EVAL]         Zorn, G., "Specifying Security Claims for EAP
                      Authentication Types", Work in Progress, October
                      2002.



Aboba, et al.               Standards Track                    [Page 62]

RFC 3748                          EAP                          June 2004


   [BINDING]          Puthenkulam, J., "The Compound Authentication
                      Binding Problem", Work in Progress, October 2003.

   [MITM]             Asokan, N., Niemi, V. and K. Nyberg, "Man-in-the-
                      Middle in Tunneled Authentication Protocols", IACR
                      ePrint Archive Report 2002/163, October 2002,
                      <http://eprint.iacr.org/2002/163>.

   [IEEE-802.11i-req] Stanley, D., "EAP Method Requirements for Wireless
                      LANs", Work in Progress, February 2004.

   [PPTPv2]           Schneier, B. and Mudge, "Cryptanalysis of
                      Microsoft's PPTP Authentication Extensions (MS-
                      CHAPv2)", CQRE 99, Springer-Verlag, 1999, pp.
                      192-203.




































Aboba, et al.               Standards Track                    [Page 63]

RFC 3748                          EAP                          June 2004


Appendix A. Changes from RFC 2284



   This section lists the major changes between [RFC2284] and this
   document.  Minor changes, including style, grammar, spelling, and
   editorial changes are not mentioned here.

   o  The Terminology section (Section 1.2) has been expanded, defining
      more concepts and giving more exact definitions.

   o  The concepts of Mutual Authentication, Key Derivation, and Result
      Indications are introduced and discussed throughout the document
      where appropriate.

   o In Section 2, it is explicitly specified that more than one
      exchange of Request and Response packets may occur as part of the
      EAP authentication exchange.  How this may be used and how it may
      not be used is specified in detail in Section 2.1.

   o  Also in Section 2, some requirements have been made explicit for
      the authenticator when acting in pass-through mode.

   o  An EAP multiplexing model (Section 2.2) has been added to
      illustrate a typical implementation of EAP.  There is no
      requirement that an implementation conform to this model, as long
      as the on-the-wire behavior is consistent with it.

   o  As EAP is now in use with a variety of lower layers, not just PPP
      for which it was first designed, Section 3 on lower layer behavior
      has been added.

   o  In the description of the EAP Request and Response interaction
      (Section 4.1), both the behavior on receiving duplicate requests,
      and when packets should be silently discarded has been more
      exactly specified.  The implementation notes in this section have
      been substantially expanded.

   o  In Section 4.2, it has been clarified that Success and Failure
      packets must not contain additional data, and the implementation
      note has been expanded.  A subsection giving requirements on
      processing of success and failure packets has been added.

   o  Section 5 on EAP Request/Response Types lists two new Type values:
      the Expanded Type (Section 5.7), which is used to expand the Type
      value number space, and the Experimental Type.  In the Expanded
      Type number space, the new Expanded Nak (Section 5.3.2) Type has
      been added.  Clarifications have been made in the description of
      most of the existing Types.  Security claims summaries have been
      added for authentication methods.



Aboba, et al.               Standards Track                    [Page 64]

RFC 3748                          EAP                          June 2004


   o  In Sections 5, 5.1, and 5.2, a requirement has been added such
      that fields with displayable messages should contain UTF-8 encoded
      ISO 10646 characters.

   o  It is now required in Section 5.1 that if the Type-Data field of
      an Identity Request contains a NUL-character, only the part before
      the null is displayed.  RFC 2284 prohibits the null termination of
      the Type-Data field of Identity messages.  This rule has been
      relaxed for Identity Request messages and the Identity Request
      Type-Data field may now be null terminated.

   o  In Section 5.5, support for OTP Extended Responses [RFC2243] has
      been added to EAP OTP.

   o  An IANA Considerations section (Section 6) has been added, giving
      registration policies for the numbering spaces defined for EAP.

   o  The Security Considerations (Section 7) have been greatly
      expanded, giving a much more comprehensive coverage of possible
      threats and other security considerations.

   o  In Section 7.5, text has been added on method-specific behavior,
      providing guidance on how EAP method-specific integrity checks
      should be processed.  Where possible, it is desirable for a
      method-specific MIC to be computed over the entire EAP packet,
      including the EAP layer header (Code, Identifier, Length) and EAP
      method layer header (Type, Type-Data).

   o  In Section 7.14 the security risks involved in use of cleartext
      passwords with EAP are described.

   o  In Section 7.15 text has been added relating to detection of rogue
      NAS behavior.


















Aboba, et al.               Standards Track                    [Page 65]

RFC 3748                          EAP                          June 2004


Authors' Addresses



   Bernard Aboba
   Microsoft Corporation
   One Microsoft Way
   Redmond, WA  98052
   USA

   Phone: +1 425 706 6605
   Fax:   +1 425 936 6605
   EMail: bernarda@microsoft.com

   Larry J. Blunk
   Merit Network, Inc
   4251 Plymouth Rd., Suite 2000
   Ann Arbor, MI  48105-2785
   USA

   Phone: +1 734-647-9563
   Fax:   +1 734-647-3185
   EMail: ljb@merit.edu

   John R. Vollbrecht
   Vollbrecht Consulting LLC
   9682 Alice Hill Drive
   Dexter, MI  48130
   USA

   EMail: jrv@umich.edu

   James Carlson
   Sun Microsystems, Inc
   1 Network Drive
   Burlington, MA  01803-2757
   USA

   Phone: +1 781 442 2084
   Fax:   +1 781 442 1677
   EMail: james.d.carlson@sun.com

   Henrik Levkowetz
   ipUnplugged AB
   Arenavagen 33
   Stockholm  S-121 28
   SWEDEN

   Phone: +46 708 32 16 08
   EMail: henrik@levkowetz.com



Aboba, et al.               Standards Track                    [Page 66]

RFC 3748                          EAP                          June 2004


Full Copyright Statement



   Copyright (C) The Internet Society (2004).  This document is subject
   to the rights, licenses and restrictions contained in BCP 78, and
   except as set forth therein, the authors retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
   ENGINEERING TASK FORCE DISCLAIM 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.

Intellectual Property



   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at ietf-
   ipr@ietf.org.

Acknowledgement



   Funding for the RFC Editor function is currently provided by the
   Internet Society.









Aboba, et al.               Standards Track                    [Page 67]