RFC 9140

Internet Engineering Task Force (IETF)                           T. Aura
Request for Comments: 9140                              Aalto University
Category: Standards Track                                       M. Sethi
ISSN: 2070-1721                                                 Ericsson
                                                             A. Peltonen
                                                        Aalto University
                                                           December 2021

          Nimble Out-of-Band Authentication for EAP (EAP-NOOB)


   The Extensible Authentication Protocol (EAP) provides support for
   multiple authentication methods.  This document defines the EAP-NOOB
   authentication method for nimble out-of-band (OOB) authentication and
   key derivation.  The EAP method is intended for bootstrapping all
   kinds of Internet-of-Things (IoT) devices that have no preconfigured
   authentication credentials.  The method makes use of a user-assisted,
   one-directional, out-of-band (OOB) message between the peer device
   and authentication server to authenticate the in-band key exchange.
   The device must have a nonnetwork input or output interface, such as
   a display, microphone, speaker, or blinking light, that can send or
   receive dynamically generated messages of tens of bytes in length.

Status of This Memo

   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at

Copyright Notice

   Copyright (c) 2021 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Revised BSD License text as described in Section 4.e of the
   Trust Legal Provisions and are provided without warranty as described
   in the Revised BSD License.

Table of Contents

   1.  Introduction
   2.  Terminology
   3.  EAP-NOOB Method
     3.1.  Protocol Overview
     3.2.  Protocol Messages and Sequences
       3.2.1.  Common Handshake in All EAP-NOOB Exchanges
       3.2.2.  Initial Exchange
       3.2.3.  OOB Step
       3.2.4.  Completion Exchange
       3.2.5.  Waiting Exchange
     3.3.  Protocol Data Fields
       3.3.1.  Peer Identifier and NAI
       3.3.2.  Message Data Fields
     3.4.  Fast Reconnect and Rekeying
       3.4.1.  Persistent EAP-NOOB Association
       3.4.2.  Reconnect Exchange
       3.4.3.  User Reset
     3.5.  Key Derivation
     3.6.  Error Handling
       3.6.1.  Invalid Messages
       3.6.2.  Unwanted Peer
       3.6.3.  State Mismatch
       3.6.4.  Negotiation Failure
       3.6.5.  Cryptographic Verification Failure
       3.6.6.  Application-Specific Failure
   4.  ServerInfo and PeerInfo Contents
   5.  IANA Considerations
     5.1.  Cryptosuites
     5.2.  Message Types
     5.3.  Error Codes
     5.4.  ServerInfo Data Fields
     5.5.  PeerInfo Data Fields
     5.6.  Domain Name Reservation
     5.7.  Guidance for Designated Experts
   6.  Security Considerations
     6.1.  Authentication Principle
     6.2.  Identifying Correct Endpoints
     6.3.  Trusted Path Issues and Misbinding Attacks
     6.4.  Peer Identifiers and Attributes
     6.5.  Downgrading Threats
     6.6.  Protected Success and Failure Indications
     6.7.  Channel Binding
     6.8.  Denial of Service
     6.9.  Recovery from Loss of Last Message
     6.10. Privacy Considerations
     6.11. EAP Security Claims
   7.  References
     7.1.  Normative References
     7.2.  Informative References
   Appendix A.  Exchanges and Events per State
   Appendix B.  Application-Specific Parameters
   Appendix C.  EAP-NOOB Roaming
   Appendix D.  OOB Message as a URL

   Authors' Addresses

1.  Introduction

   This document describes a method for registration, authentication,
   and key derivation for network-connected smart devices, such as
   consumer and enterprise appliances that are part of the Internet of
   Things (IoT).  These devices may be off-the-shelf hardware that is
   sold and distributed without any prior registration or credential-
   provisioning process, or they may be recycled devices after a hard
   reset.  Thus, the device registration in a server database, ownership
   of the device, and the authentication credentials for both network
   access and application-level security must all be established at the
   time of the device deployment.  Furthermore, many such devices have
   only limited user interfaces that could be used for their
   configuration.  Often, the user interfaces are limited to either only
   input (e.g., a camera) or output (e.g., a display screen).  The
   device configuration is made more challenging by the fact that the
   devices may exist in large numbers and may have to be deployed or
   reconfigured nimbly based on user needs.

   To summarize, devices may have the following characteristics:

   *  no preestablished relation with the intended server or user,

   *  no preprovisioned device identifier or authentication credentials,

   *  an input or output interface that may be capable of only one-
      directional out-of-band communication.

   Many proprietary out-of-band (OOB) configuration methods exist for
   specific IoT devices.  The goal of this specification is to provide
   an open standard and a generic protocol for bootstrapping the
   security of network-connected appliances, such as displays, printers,
   speakers, and cameras.  The security bootstrapping in this
   specification makes use of a user-assisted OOB channel.  The device
   authentication relies on a user having physical access to the device,
   and the key exchange security is based on the assumption that
   attackers are not able to observe or modify the messages conveyed
   through the OOB channel.  We follow the common approach taken in
   pairing protocols: performing a Diffie-Hellman key exchange over the
   insecure network and authenticating the established key with the help
   of the OOB channel in order to prevent impersonation attacks.

   The solution presented here is intended for devices that have either
   a nonnetwork input or output interface, such as a camera, microphone,
   display screen, speaker, or blinking Light Emitting Diode (LED)
   light, that is able to send or receive dynamically generated messages
   of tens of bytes in length.  Naturally, this solution may not be
   appropriate for very small sensors or actuators that have no user
   interface at all or for devices that are inaccessible to the user.
   We also assume that the OOB channel is at least partly automated
   (e.g., a camera scanning a bar code); thus, there is no need to
   absolutely minimize the length of the data transferred through the
   OOB channel.  This differs, for example, from Bluetooth pairing
   [Bluetooth], where it is essential to minimize the length of the
   manually transferred or compared codes.  The OOB messages in this
   specification are dynamically generated.  Thus, we do not support
   static printed registration codes.  One reason for requiring dynamic
   OOB messages is that the receipt of the OOB message authorizes the
   server to take ownership of the device.  Dynamic OOB messages are
   more secure than static printed codes, which could be leaked and
   later misused.

2.  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

   In addition, this document frequently uses the following terms as
   they have been defined in [RFC5216]:

         The entity initiating EAP authentication.

         The entity that responds to the authenticator.  In
         [IEEE-802.1X], this entity is known as the supplicant.  (We use
         the terms peer, device, and peer device interchangeably.)

         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.

3.  EAP-NOOB Method

   This section defines the EAP-NOOB method.  The protocol is a
   generalized version of the original idea presented by Sethi et al.

3.1.  Protocol Overview

   One EAP-NOOB method execution spans two or more EAP conversations,
   called Exchanges in this specification.  Each Exchange consists of
   several EAP request-response pairs.  At least two separate EAP
   conversations are needed to give the human user time to deliver the
   OOB message between them.

   The overall protocol starts with the Initial Exchange, which
   comprises four EAP request-response pairs.  In the Initial Exchange,
   the server allocates an identifier to the peer, and the server and
   peer negotiate the protocol version and cryptosuite (i.e.,
   cryptographic algorithm suite), exchange nonces, and perform an
   Ephemeral Elliptic Curve Diffie-Hellman (ECDHE) key exchange.  The
   user-assisted OOB Step then takes place.  This step requires only one
   out-of-band message, either from the peer to the server or from the
   server to the peer.  While waiting for the OOB Step action, the peer
   MAY probe the server by reconnecting to it with EAP-NOOB.  If the OOB
   Step has already taken place, the probe leads to the Completion
   Exchange, which completes the mutual authentication and key
   confirmation.  On the other hand, if the OOB Step has not yet taken
   place, the probe leads to the Waiting Exchange, and the peer will
   perform another probe after a server-defined minimum waiting time.
   The Initial Exchange and Waiting Exchange always end in EAP-Failure,
   while the Completion Exchange may result in EAP-Success.  Once the
   peer and server have performed a successful Completion Exchange, both
   endpoints store the created association in persistent storage, and
   the OOB Step is not repeated.  Thereafter, creation of new temporal
   keys, ECDHE rekeying, and updates of cryptographic algorithms can be
   achieved with the Reconnect Exchange.

                                       OOB Output/Initial Exchange/
                                                  Waiting Exchange
                                                    |     v
        .------------------.   Initial       .------------------.
        |                  |   Exchange      |                  |
     .->| Unregistered (0) |---------------->|Waiting for OOB(1)|
     |  |   (ephemeral)    |                 |   (ephemeral)    |
     |  |                  |                 |                  |
     |  '------------------'                 '------------------'
     |                                         |      |      ^
    User Reset                 Completion      |      |      |
     |                         Exchange        |     OOB   OOB
     |<-------.      .-------------------------'    Input  Reject/
     |        |      |                                |    Initial
     |        |      |                                |    Exchange
     |        |      v                                v      |
     |  .------------------.   Completion    .------------------.
     |  |                  |   Exchange      |                  |
     |  |  Registered (4)  |<----------------| OOB Received (2) |
     |  |   (persistent)   |                 |   (ephemeral)    |
     |  |                  |                 |                  |
     |  '------------------'                 '------------------'
     |        |      ^
     |  Mobility/    |
     |  Timeout/   Reconnect
     |  Failure    Exchange
     |        |      |
     |        v      |
     |  .------------------.
     |  |                  |
     '--| Reconnecting (3) |
        |   (persistent)   |
        |                  |

          Figure 1: EAP-NOOB Server-Peer Association State Machine

   Figure 1 shows the association state machine, which is the same for
   the server and for the peer.  (For readability, only the main state
   transitions are shown.  The complete table of transitions can be
   found in Appendix A.)  When the peer initiates the EAP-NOOB method,
   the server chooses the ensuing message exchange based on the
   combination of the server and peer states.  The EAP server and peer
   are initially in the Unregistered (0) state, in which no state
   information needs to be stored.  Before a successful Completion
   Exchange, the server-peer association state is ephemeral in both the
   server and peer (ephemeral states 0..2), and a timeout or error may
   cause one or both endpoints to go back to the Unregistered (0) state
   so that the Initial Exchange is repeated.  After the Completion
   Exchange has resulted in EAP-Success, the association state becomes
   persistent (persistent states 3..4).  Only user reset or memory
   failure can cause the return of the server or the peer from the
   persistent states to the ephemeral states and to the Initial

   The server MUST NOT repeat a successful OOB Step with the same peer
   except if the association with the peer is explicitly reset by the
   user or lost due to failure of the persistent storage in the server.
   More specifically, once the association has entered the Registered
   (4) state, the server MUST NOT delete the association or go back to
   the ephemeral states 0..2 without explicit user approval.  Similarly,
   the peer MUST NOT repeat the OOB Step unless the user explicitly
   deletes the association with the server from the peer or resets the
   peer to the Unregistered (0) state.  The server and peer MAY
   implement user reset of the association by deleting the state data
   from that endpoint.  If an endpoint continues to store data about the
   association after the user reset, its behavior MUST be equivalent to
   having deleted the association data.

   It can happen that the peer accidentally (or through user reset)
   loses its persistent state and reconnects to the server without a
   previously allocated peer identifier.  In that case, the server MUST
   treat the peer as a new peer.  The server MAY use auxiliary
   information, such as the PeerInfo field received in the Initial
   Exchange, to detect multiple associations with the same peer.
   However, it MUST NOT delete or merge redundant associations without
   user or application approval because EAP-NOOB internally has no
   secure way of verifying that the two peers are the same physical
   device.  Similarly, the server might lose the association state
   because of a memory failure or user reset.  In that case, the only
   way to recover is that the user also resets the peer.

   A special feature of the EAP-NOOB method is that the server is not
   assumed to have any a priori knowledge of the peer.  Therefore, the
   peer initially uses the generic identity string "noob@eap-noob.arpa"
   as its Network Access Identifier (NAI).  The server then allocates a
   server-specific identifier to the peer.  The generic NAI serves two
   purposes: firstly, it tells the server that the peer supports and
   expects the EAP-NOOB method; secondly, it allows routing of the EAP-
   NOOB sessions to a specific authentication server in an
   Authentication, Authorization, and Accounting (AAA) architecture.

   EAP-NOOB is an unusual EAP method in that the peer has to have
   multiple EAP conversations with the server before it can receive EAP-
   Success.  The reason is that, while EAP allows delays between the
   request-response pairs, e.g., for repeated password entry, the user
   delays in OOB authentication can be much longer than in password
   trials.  Moreover, EAP-NOOB supports peers with no input capability
   in the user interface (e.g., LED light bulbs).  Since users cannot
   initiate the protocol in these devices, the devices have to perform
   the Initial Exchange opportunistically and hope for the OOB Step to
   take place within a timeout period (NoobTimeout), which is why the
   timeout needs to be several minutes rather than seconds.  To support
   such high-latency OOB channels, the peer and server perform the
   Initial Exchange in one EAP conversation, then allow time for the OOB
   message to be delivered, and later perform the Waiting Exchange and
   Completion Exchange in different EAP conversations.

3.2.  Protocol Messages and Sequences

   This section defines the EAP-NOOB exchanges, which correspond to EAP
   conversations.  The exchanges start with a common handshake, which
   determines the type of the following exchange.  The common handshake
   messages and the subsequent messages for each exchange type are
   listed in the diagrams below.  The diagrams also specify the data
   fields present in each message.  Each exchange comprises multiple EAP
   request-response pairs and ends in either EAP-Failure, indicating
   that authentication is not (yet) successful, or in EAP-Success.

3.2.1.  Common Handshake in All EAP-NOOB Exchanges

   All EAP-NOOB exchanges start with common handshake messages.  The
   handshake begins with the identity request and response that are
   common to all EAP methods.  Their purpose is to enable the AAA
   architecture to route the EAP conversation to the EAP server and to
   enable the EAP server to select the EAP method.  The handshake then
   continues with one EAP-NOOB request-response pair in which the server
   discovers the peer identifier used in EAP-NOOB and the peer state.

   In more detail, each EAP-NOOB exchange begins with the authenticator
   sending an EAP-Request/Identity packet to the peer.  From this point
   on, the EAP conversation occurs between the server and the peer, and
   the authenticator acts as a pass-through device.  The peer responds
   to the authenticator with an EAP-Response/Identity packet, which
   contains the Network Access Identifier (NAI).  The authenticator,
   acting as a pass-through device, forwards this response and the
   following EAP conversation between the peer and the AAA architecture.
   The AAA architecture routes the conversation to a specific AAA server
   (called "EAP server" or simply "server" in this specification) based
   on the realm part of the NAI.  The server selects the EAP-NOOB method
   based on the user part of the NAI, as defined in Section 3.3.1.

   After receiving the EAP-Response/Identity message, the server sends
   the first EAP-NOOB request (Type=1) to the peer, which responds with
   the peer identifier (PeerId) and state (PeerState) in the range 0..3.
   However, the peer SHOULD omit the PeerId from the response (Type=1)
   when PeerState=0.  The server then chooses the EAP-NOOB exchange,
   i.e., the ensuing message sequence, as explained below.  The peer
   recognizes the exchange based on the message type field (Type) of the
   next EAP-NOOB request received from the server.

   The server MUST determine the exchange type based on the combination
   of the peer and server states as follows (also summarized in
   Table 14).  If either the peer or server is in the Unregistered (0)
   state and the other is in one of the ephemeral states (0..2), the
   server chooses the Initial Exchange.  If either the peer or server is
   in the OOB Received (2) state and the other is either in the Waiting
   for OOB (1) or OOB Received (2) state, the OOB Step has taken place
   and the server chooses the Completion Exchange.  If both the server
   and peer are in the Waiting for OOB (1) state, the server chooses the
   Waiting Exchange.  If the peer is in the Reconnecting (3) state and
   the server is in the Registered (4) or Reconnecting (3) state, the
   server chooses the Reconnect Exchange.  All other state combinations
   are error situations where user action is required, and the server
   SHOULD indicate such errors to the peer with the error code 2002 (see
   Section 3.6.3).  Note also that the peer MUST NOT initiate EAP-NOOB
   when the peer is in the Registered (4) state.

         EAP Peer                      Authenticator    EAP Server
           |                                   |              |
           |<----------- EAP-Request/Identity -|              |
           |                                                  |
           |                                                  |
           |------------ EAP-Response/Identity -------------->|
           |      (NAI=noob@eap-noob.arpa)                    |
           |                                                  |
           |                                                  |
           |<----------- EAP-Request/EAP-NOOB ----------------|
           |      (Type=1)                                    |
           |                                                  |
           |                                                  |
           |------------ EAP-Response/EAP-NOOB -------------->|
           |      (Type=1,[PeerId],PeerState=1)               |
           |                                                  |
           |  continuing with exchange-specific messages...   |

            Figure 2: Common Handshake in All EAP-NOOB Exchanges

3.2.2.  Initial Exchange

   The Initial Exchange comprises the common handshake and two further
   EAP-NOOB request-response pairs: one for version, cryptosuite, and
   parameter negotiation and the other for the ECDHE key exchange.  The
   first EAP-NOOB request (Type=2) from the server contains a newly
   allocated PeerId for the peer and an optional NewNAI for assigning a
   new NAI to the peer.  The server allocates a new PeerId in the
   Initial Exchange regardless of any old PeerId received in the
   previous response (Type=1).  The server also sends in the request a
   list of the protocol versions (Vers) and cryptosuites (Cryptosuites)
   it supports, an indicator of the OOB channel directions it supports
   (Dirs), and a ServerInfo object.  The peer chooses one of the
   versions and cryptosuites.  The peer sends a response (Type=2) with
   the selected protocol version (Verp), the received PeerId, the
   selected cryptosuite (Cryptosuitep), an indicator of the OOB channel
   direction(s) selected by the peer (Dirp), and a PeerInfo object.  In
   the second EAP-NOOB request and response (Type=3), the server and
   peer exchange the public components of their ECDHE keys and nonces
   (PKs, Ns, PKp, and Np).  The ECDHE keys MUST be based on the
   negotiated cryptosuite, i.e., Cryptosuitep.  The Initial Exchange
   always ends with EAP-Failure from the server because the
   authentication cannot yet be completed.

        EAP Peer                                        EAP Server
          |       ...continuing from common handshake        |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=2,Vers,PeerId,[NewNAI],               |
          |       Cryptosuites,Dirs,ServerInfo)              |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=2,Verp,PeerId,Cryptosuitep,           |
          |        Dirp,PeerInfo)                            |
          |                                                  |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=3,PeerId,PKs,Ns,[SleepTime])          |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=3,PeerId,PKp,Np)                      |
          |                                                  |
          |                                                  |
          |<----------- EAP-Failure -------------------------|
          |                                                  |

                         Figure 3: Initial Exchange

   At the conclusion of the Initial Exchange, both the server and the
   peer move to the Waiting for OOB (1) state.

3.2.3.  OOB Step

   The OOB Step, labeled as OOB Output and OOB Input in Figure 1, takes
   place after the Initial Exchange.  Depending on the negotiated OOB
   channel direction, the peer or the server outputs the OOB message as
   shown in Figures 4 or 5, respectively.  The data fields are the
   PeerId, the secret nonce Noob, and the cryptographic fingerprint
   Hoob.  The contents of the data fields are defined in Section 3.3.2.
   The OOB message is delivered to the other endpoint via a user-
   assisted OOB channel.

   For brevity, we will use the terms OOB sender and OOB receiver in
   addition to the already familiar EAP server and EAP peer.  If the OOB
   message is sent in the server-to-peer direction, the OOB sender is
   the server and the OOB receiver is the peer.  On the other hand, if
   the OOB message is sent in the peer-to-server direction, the OOB
   sender is the peer and the OOB receiver is the server.

        EAP Peer                                        EAP Server
          |                                                  |
          |             (PeerId,Noob,Hoob)                   |
          |                                                  |

                Figure 4: OOB Step, from Peer to EAP Server

        EAP Peer                                        EAP Server
          |                                                  |
          |             (PeerId,Noob,Hoob)                   |
          |                                                  |

                Figure 5: OOB Step, from EAP Server to Peer

   The OOB receiver MUST compare the received value of the fingerprint
   Hoob (see Section 3.3.2) with a value that it computed locally for
   the PeerID received.  This integrity check ensures that the endpoints
   agree on contents of the Initial Exchange.  If the values are equal,
   the receiver moves to the OOB Received (2) state.  Otherwise, the
   receiver MUST reject the OOB message.  For usability reasons, the OOB
   receiver SHOULD indicate the acceptance or rejection of the OOB
   message to the user.  The receiver SHOULD reject invalid OOB messages
   without changing its state in the association state machine until an
   application-specific number of invalid messages (OobRetries) has been
   reached; after which, the receiver SHOULD consider it an error and go
   back to the Unregistered (0) state.

   The server or peer MAY send multiple OOB messages with different Noob
   values while in the Waiting for OOB (1) state.  The OOB sender SHOULD
   remember the Noob values until they expire and accept any one of them
   in the following Completion Exchange.  The Noob values sent by the
   server expire after an application-dependent timeout (NoobTimeout),
   and the server MUST NOT accept Noob values older than that in the
   Completion Exchange.  The RECOMMENDED value for NoobTimeout is 3600
   seconds if there are no application-specific reasons for making it
   shorter or longer.  The Noob values sent by the peer expire, as
   defined in Section 3.2.5.

   The OOB receiver does not accept further OOB messages after it has
   accepted one and moved to the OOB Received (2) state.  However, the
   receiver MAY buffer redundant OOB messages in case an OOB message
   expiry or similar error detected in the Completion Exchange causes it
   to return to the Waiting for OOB (1) state.  It is RECOMMENDED that
   the OOB receiver notifies the user about redundant OOB messages, but
   it MAY instead discard them silently.

   The sender will typically generate a new Noob, and therefore a new
   OOB message, at constant time intervals (NoobInterval).  The
   RECOMMENDED interval is

      NoobInterval = NoobTimeout / 2

   in which case, the receiver of the OOB will at any given time accept
   either of the two latest Noob values.  However, the timing of the
   Noob generation may also be based on user interaction or on
   implementation considerations.

   Even though not recommended (see Section 3.3), this specification
   allows both directions to be negotiated (Dirp=3) for the OOB channel.
   In that case, both sides SHOULD output the OOB message, and it is up
   to the user to deliver at least one of them.

   The details of the OOB channel implementation including the message
   encoding are defined by the application.  Appendix D gives an example
   of how the OOB message can be encoded as a URL that may be embedded
   in a dynamic QR code or NFC (Near Field Communication) tag.

3.2.4.  Completion Exchange

   After the Initial Exchange, if the OOB channel directions selected by
   the peer include the peer-to-server direction, the peer SHOULD
   initiate the EAP-NOOB method again after an applications-specific
   waiting time in order to probe for completion of the OOB Step.  If
   the OOB channel directions selected by the peer include the server-
   to-peer direction and the peer receives the OOB message, it SHOULD
   initiate the EAP-NOOB method immediately.  Depending on the
   combination of the peer and server states, the server continues with
   the Completion Exchange or Waiting Exchange (see Section 3.2.1 on how
   the server makes this decision).

   The Completion Exchange comprises the common handshake and one or two
   further EAP-NOOB request-response pairs.  If the peer is in the
   Waiting for OOB (1) state, the OOB message has been sent in the peer-
   to-server direction.  In that case, only one request-response pair
   (Type=6) takes place.  In the request, the server sends the NoobId
   value (see Section 3.3.2), which the peer uses to identify the exact
   OOB message received by the server.  On the other hand, if the peer
   is in the OOB Received (2) state, the direction of the OOB message is
   from server to peer.  In this case, two request-response pairs
   (Type=5 and Type=6) are needed.  The purpose of the first request-
   response pair (Type=5) is that it enables the server to discover
   NoobId, which identifies the exact OOB message received by the peer.
   The server returns the same NoobId to the peer in the latter request.

   In the last request-response pair (Type=6) of the Completion
   Exchange, the server and peer exchange message authentication codes.
   Both sides MUST compute the keys Kms and Kmp, as defined in
   Section 3.5, and the message authentication codes MACs and MACp, as
   defined in Section 3.3.2.  Both sides MUST compare the received
   message authentication code with a locally computed value.  If the
   peer finds that it has received the correct value of MACs and the
   server finds that it has received the correct value of MACp, the
   Completion Exchange ends in EAP-Success.  Otherwise, the endpoint
   where the comparison fails indicates this with an error message
   (error code 4001, see Section 3.6.5), and the Completion Exchange
   ends in EAP-Failure.

   After the successful Completion Exchange, both the server and the
   peer move to the Registered (4) state.  They also derive the output
   keying material and store the persistent EAP-NOOB association state,
   as defined in Sections 3.4 and 3.5.

   It is possible that the OOB message expires before it is received.
   In that case, the sender of the OOB message no longer recognizes the
   NoobId that it receives in the Completion Exchange.  Another reason
   why the OOB sender might not recognize the NoobId is if the received
   OOB message was spoofed and contained an attacker-generated Noob
   value.  The recipient of an unrecognized NoobId indicates this with
   an error message (error code 2003, see Section 3.6.1), and the
   Completion Exchange ends in EAP-Failure.  The recipient of the error
   message 2003 moves back to the Waiting for OOB (1) state.  This state
   transition is called OOB Reject in Figure 1 (even though it really is
   a specific type of failed Completion Exchange).  On the other hand,
   the sender of the error message stays in its previous state.

   Although it is not expected to occur in practice, poor user interface
   design could lead to two OOB messages delivered simultaneously, one
   from the peer to the server and the other from the server to the
   peer.  The server detects this event in the beginning of the
   Completion Exchange by observing that both the server and peer are in
   the OOB Received (2) state.  In that case, as a tiebreaker, the
   server MUST behave as if only the server-to-peer message had been

        EAP Peer                                        EAP Server
          |       ...continuing from common handshake        |
          |                                                  |
          |<----------- [ EAP-Request/EAP-NOOB ] ------------|
          |      (Type=5,PeerId)                             |
          |                                                  |
          |                                                  |
          |------------ [ EAP-Response/EAP-NOOB ] ---------->|
          |      (Type=5,PeerId,NoobId)                      |
          |                                                  |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=6,PeerId,NoobId,MACs)                 |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=6,PeerId,MACp)                        |
          |                                                  |
          |                                                  |
          |<----------- EAP-Success -------------------------|
          |                                                  |

                       Figure 6: Completion Exchange

3.2.5.  Waiting Exchange

   As explained in Section 3.2.4, the peer SHOULD probe the server for
   completion of the OOB Step.  When the combination of the peer and
   server states indicates that the OOB message has not yet been
   delivered, the server chooses the Waiting Exchange (see Section 3.2.1
   on how the server makes this decision).  The Waiting Exchange
   comprises the common handshake and one further request-response pair,
   and it always ends in EAP-Failure.

   In order to limit the rate at which peers probe the server, the
   server MAY send to the peer either in the Initial Exchange or in the
   Waiting Exchange a minimum time to wait before probing the server
   again.  A peer that has not received an OOB message SHOULD wait at
   least the server-specified minimum waiting time in seconds
   (SleepTime) before initiating EAP again with the same server.  The
   peer uses the latest SleepTime value that it has received in or after
   the Initial Exchange.  If the server has not sent any SleepTime
   value, the peer MUST wait for an application-specified minimum time

   After the Waiting Exchange, the peer MUST discard (from its local
   ephemeral storage) Noob values that it has sent to the server in OOB
   messages that are older than the application-defined timeout
   NoobTimeout (see Section 3.2.3).  The peer SHOULD discard such
   expired Noob values even if the probing failed because of, e.g.,
   failure to connect to the EAP server or an incorrect message
   authentication code.  The timeout of peer-generated Noob values is
   defined like this in order to allow the peer to probe the server once
   after it has waited for the server-specified SleepTime.

   If the server and peer have negotiated to use only the server-to-peer
   direction for the OOB channel (Dirp=2), the peer SHOULD nevertheless
   probe the server.  The purpose of this is to keep the server informed
   about the peers that are still waiting for OOB messages.  The server
   MAY set SleepTime to a high number (e.g., 3600) to prevent the peer
   from probing the server frequently.

        EAP Peer                                        EAP Server
          |       ...continuing from common handshake        |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=4,PeerId,[SleepTime])                 |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=4,PeerId)                             |
          |                                                  |
          |                                                  |
          |<----------- EAP-Failure -------------------------|
          |                                                  |

                         Figure 7: Waiting Exchange

3.3.  Protocol Data Fields

   This section defines the various identifiers and data fields used in
   the EAP-NOOB method.

3.3.1.  Peer Identifier and NAI

   The server allocates a new peer identifier (PeerId) for the peer in
   the Initial Exchange.  The peer identifier MUST follow the syntax of
   the utf8-username specified in [RFC7542].  The server MUST generate
   the identifiers in such a way that they do not repeat and cannot be
   guessed by the peer or third parties before the server sends them to
   the peer in the Initial Exchange.  One way to generate the
   identifiers is to choose a random 16-byte identifier and to base64url
   encode it without padding [RFC4648] into a 22-character ASCII string.
   Another way to generate the identifiers is to choose a random
   22-character alphanumeric ASCII string.  It is RECOMMENDED to not use
   identifiers longer than this because they result in longer OOB

   The peer uses the allocated PeerId to identify itself to the server
   in the subsequent exchanges.  The peer MUST copy the PeerId byte by
   byte from the message where it was allocated, and the server MUST
   perform a byte-by-byte comparison between the received and the
   previously allocated PeerID.  The peer sets the PeerId value in
   response type 1 as follows.  As stated in Section 3.2.1, when the
   peer is in the Unregistered (0) state, it SHOULD omit the PeerId from
   response type 1.  When the peer is in one of the states 1..2, it MUST
   use the PeerId that the server assigned to it in the latest Initial
   Exchange.  When the peer is in one of the persistent states 3..4, it
   MUST use the PeerId from its persistent EAP-NOOB association.  (The
   PeerId is written to the association when the peer moves to the
   Registered (4) state after a Completion Exchange.)

   The default NAI for the peer is "noob@eap-noob.arpa".  The peer
   implementation MAY allow the user or application to configure a
   different NAI, which overrides the default NAI.  Furthermore, the
   server MAY assign a new NAI to the peer in the Initial Exchange or
   Reconnect Exchange in the NewNAI field of request types 2 and 7 to
   override any previous NAI value.  When the peer is in the
   Unregistered (0) state, or when the peer is in one of the states 1..2
   and the server did not send a NewNAI in the latest Initial Exchange,
   the peer MUST use the configured NAI or, if it does not exist, the
   default NAI.  When the peer is in one of the states 1..2 and the
   server sent a NewNAI in the latest Initial Exchange, the peer MUST
   use this server-assigned NAI.  When the peer moves to the Registered
   (4) state after the Completion Exchange, it writes to the persistent
   EAP-NOOB association the same NAI value that it used in the
   Completion Exchange.  When the peer is in the Reconnecting (3) or
   Registered (4) state, it MUST use the NAI from its persistent EAP-
   NOOB association.  When the server sends NewNAI in the Reconnect
   Exchange, the peer writes its value to the persistent EAP-NOOB
   association when it moves from the Reconnecting (3) state to the
   Registered (4) state.  All the NAI values MUST follow the syntax
   specified in [RFC7542].

   The purpose of the server-assigned NAI is to enable more flexible
   routing of the EAP sessions over the AAA infrastructure, including
   roaming scenarios (see Appendix C).  Moreover, some authenticators or
   AAA servers use the realm part of the assigned NAI to determine peer-
   specific connection parameters, such as isolating the peer to a
   specific VLAN.  On the other hand, the user- or application-
   configured NAI enables registration of new devices while roaming.  It
   also enables manufacturers to set up their own AAA servers for
   bootstrapping of new peer devices.

   The peer's PeerId and server-assigned NAI are ephemeral until a
   successful Completion Exchange takes place.  Thereafter, the values
   become parts of the persistent EAP-NOOB association until the user
   resets the peer and server or until a new NAI is assigned in the
   Reconnect Exchange.

3.3.2.  Message Data Fields

   Table 1 defines the data fields in the protocol messages.  The in-
   band messages are formatted as JSON objects [RFC8259] in UTF-8
   encoding.  The JSON member names are in the left-hand column of the

    | Data Field    | Description                                     |
    | Vers, Verp    | EAP-NOOB protocol versions supported by the EAP |
    |               | server and the protocol version chosen by the   |
    |               | peer.  Vers is a JSON array of unsigned         |
    |               | integers, and Verp is an unsigned integer.      |
    |               | Example values are "[1]" and "1", respectively. |
    | PeerId        | Peer identifier, as defined in Section 3.3.1.   |
    | NAI, NewNAI   | Peer NAI and server-assigned new peer NAI, as   |
    |               | defined in Section 3.3.1.                       |
    | Type          | EAP-NOOB message type.  The type is an integer  |
    |               | in the range 0..9.  EAP-NOOB requests and the   |
    |               | corresponding responses share the same type     |
    |               | value.                                          |
    | PeerState     | Peer state is an integer in the range 0..4 (see |
    |               | Figure 1).  However, only values 0..3 are ever  |
    |               | sent in the protocol messages.                  |
    | PKs, PKp      | The public components of the ECDHE keys of the  |
    |               | server and peer.  PKs and PKp are sent in the   |
    |               | JSON Web Key (JWK) format [RFC7517].  The       |
    |               | detailed format of the JWK object is defined by |
    |               | the cryptosuite.                                |
    | Cryptosuites, | The identifiers of cryptosuites supported by    |
    | Cryptosuitep  | the server and of the cryptosuite selected by   |
    |               | the peer.  The server-supported cryptosuites in |
    |               | Cryptosuites are formatted as a JSON array of   |
    |               | the identifier integers.  The server MUST send  |
    |               | a nonempty array with no repeating elements,    |
    |               | ordered by decreasing priority.  The peer MUST  |
    |               | respond with exactly one suite in the           |
    |               | Cryptosuitep value, formatted as an identifier  |
    |               | integer.  Mandatory-to-implement cryptosuites   |
    |               | and the registration procedure for new          |
    |               | cryptosuites are specified in Section 5.1.      |
    |               | Example values are "[1]" and "1", respectively. |
    | Dirs, Dirp    | An integer indicating the OOB channel           |
    |               | directions supported by the server and the      |
    |               | directions selected by the peer.  The possible  |
    |               | values are 1=peer-to-server, 2=server-to-peer,  |
    |               | and 3=both directions.                          |
    | Dir           | The actual direction of the OOB message         |
    |               | (1=peer-to-server, 2=server-to-peer).  This     |
    |               | value is not sent over any communication        |
    |               | channel, but it is included in the computation  |
    |               | of the cryptographic fingerprint Hoob.          |
    | Ns, Np        | 32-byte nonces for the Initial Exchange.        |
    | ServerInfo    | This field contains information about the       |
    |               | server to be passed from the EAP method to the  |
    |               | application layer in the peer.  The information |
    |               | is specific to the application or to the OOB    |
    |               | channel, and it is encoded as a JSON object of  |
    |               | at most 500 bytes.  It could include, for       |
    |               | example, the access-network name and server     |
    |               | name, a Uniform Resource Locator (URL)          |
    |               | [RFC3986], or some other information that helps |
    |               | the user deliver the OOB message to the server  |
    |               | through the out-of-band channel.                |
    | PeerInfo      | This field contains information about the peer  |
    |               | to be passed from the EAP method to the         |
    |               | application layer in the server.  The           |
    |               | information is specific to the application or   |
    |               | to the OOB channel, and it is encoded as a JSON |
    |               | object of at most 500 bytes.  It could include, |
    |               | for example, the peer brand, model, and serial  |
    |               | number, which help the user distinguish between |
    |               | devices and deliver the OOB message to the      |
    |               | correct peer through the out-of-band channel.   |
    | SleepTime     | The number of seconds for which the peer MUST   |
    |               | NOT start a new execution of the EAP-NOOB       |
    |               | method with the authenticator, unless the peer  |
    |               | receives the OOB message or the sending is      |
    |               | triggered by an application-specific user       |
    |               | action.  The server can use this field to limit |
    |               | the rate at which peers probe it.  SleepTime is |
    |               | an unsigned integer in the range 0..3600.       |
    | Noob          | 16-byte secret nonce sent through the OOB       |
    |               | channel and used for the session key            |
    |               | derivation.  The endpoint that received the OOB |
    |               | message uses this secret in the Completion      |
    |               | Exchange to authenticate the exchanged key to   |
    |               | the endpoint that sent the OOB message.         |
    | Hoob          | 16-byte cryptographic fingerprint (i.e., hash   |
    |               | value) computed from all the parameters         |
    |               | exchanged in the Initial Exchange and in the    |
    |               | OOB message.  Receiving this fingerprint over   |
    |               | the OOB channel guarantees the integrity of the |
    |               | key exchange and parameter negotiation.  Hence, |
    |               | it authenticates the exchanged key to the       |
    |               | endpoint that receives the OOB message.         |
    | NoobId        | 16-byte identifier for the OOB message,         |
    |               | computed with a one-way function from the nonce |
    |               | Noob in the message.                            |
    | MACs, MACp    | Message authentication codes (HMAC) for mutual  |
    |               | authentication, key confirmation, and integrity |
    |               | check on the exchanged information.  The input  |
    |               | to the HMAC is defined below, and the key for   |
    |               | the HMAC is defined in Section 3.5.             |
    | Ns2, Np2      | 32-byte nonces for the Reconnect Exchange.      |
    | KeyingMode    | Integer indicating the key derivation method. 0 |
    |               | in the Completion Exchange, and 1..3 in the     |
    |               | Reconnect Exchange.                             |
    | PKs2, PKp2    | The public components of the ECDHE keys of the  |
    |               | server and peer for the Reconnect Exchange.     |
    |               | PKp2 and PKs2 are sent in the JSON Web Key      |
    |               | (JWK) format [RFC7517].  The detailed format of |
    |               | the JWK object is defined by the cryptosuite.   |
    | MACs2, MACp2  | Message authentication codes (HMAC) for mutual  |
    |               | authentication, key confirmation, and integrity |
    |               | check on the Reconnect Exchange.  The input to  |
    |               | the HMAC is defined below, and the key for the  |
    |               | HMAC is defined in Section 3.5.                 |
    | ErrorCode     | Integer indicating an error condition.  Defined |
    |               | in Section 5.3.                                 |
    | ErrorInfo     | Textual error message for logging and debugging |
    |               | purposes.  A UTF-8 string of at most 500 bytes. |

                        Table 1: Message Data Fields

   It is RECOMMENDED for servers to support both OOB channel directions
   (Dirs=3) unless the type of the OOB channel limits them to one
   direction (Dirs=1 or Dirs=2).  On the other hand, it is RECOMMENDED
   that the peer selects only one direction (Dirp=1 or Dirp=2) even when
   both directions (Dirp=3) would be technically possible.  The reason
   is that, if value 3 is negotiated, the user may be presented with two
   OOB messages, one for each direction, even though only one of them
   needs to be delivered.  This can be confusing to the user.
   Nevertheless, the EAP-NOOB protocol is designed to also cope with the
   value 3; in which case, it uses the first delivered OOB message.  In
   the unlikely case of simultaneously delivered OOB messages, the
   protocol prioritizes the server-to-peer direction.

   The nonces in the in-band messages (Ns, Np, Ns2, Np2) are 32-byte
   fresh random byte strings, and the secret nonce Noob is a 16-byte
   fresh random byte string.  All the nonces are generated by the
   endpoint that sends the message.

   The fingerprint Hoob and the identifier NoobId are computed with the
   cryptographic hash function H, which is specified in the negotiated
   cryptosuite and truncated to the 16 leftmost bytes of the output.
   The message authentication codes (MACs, MACp, MACs2, MACp2) are
   computed with the function HMAC, which is the hashed message
   authentication code [RFC2104] based on the cryptographic hash
   function H and truncated to the 32 leftmost bytes of the output.

   The inputs to the hash function for computing the fingerprint Hoob
   and to the HMAC for computing MACs, MACp, MACs2, and MACp2 are JSON
   arrays containing a fixed number (17) of elements.  The array
   elements MUST be copied to the array verbatim from the sent and
   received in-band messages.  When the element is a JSON object, its
   members MUST NOT be reordered or reencoded.  White space MUST NOT be
   added anywhere in the JSON structure.  Implementers should check that
   their JSON library copies the elements as UTF-8 strings, does not
   modify them in any way, and does not add white space to the HMAC

   The inputs for computing the fingerprint and message authentication
   codes are the following:

   Hoob = H(Dir,Vers,Verp,PeerId,Cryptosuites,Dirs,ServerInfo,

   NoobId = H("NoobId",Noob).

   MACs = HMAC(Kms; 2,Vers,Verp,PeerId,Cryptosuites,Dirs,ServerInfo,

   MACp = HMAC(Kmp; 1,Vers,Verp,PeerId,Cryptosuites,Dirs,ServerInfo,

   MACs2 = HMAC(Kms2; 2,Vers,Verp,PeerId,Cryptosuites,"",[ServerInfo],

   MACp2 = HMAC(Kmp2; 1,Vers,Verp,PeerId,Cryptosuites,"",[ServerInfo],

   The inputs denoted with "" above are not present, and the values in
   brackets [] are optional.  Both kinds of missing input values are
   represented by empty strings "" in the HMAC input (JSON array).  The
   NAI included in the inputs is the NAI value that will be in the
   persistent EAP-NOOB association if the Completion Exchange or
   Reconnect Exchange succeeds.  In the Completion Exchange, the NAI is
   the NewNAI value assigned by the server in the preceding Initial
   Exchange or, if no NewNAI was sent, the NAI used by the client in the
   Initial Exchange.  In the Reconnect Exchange, the NAI is the NewNAI
   value assigned by the server in the same Reconnect Exchange or, if no
   NewNAI was sent, the unchanged NAI from the persistent EAP-NOOB
   association.  Each of the values in brackets for the computation of
   Macs2 and Macp2 MUST be included if it was sent or received in the
   same Reconnect Exchange; otherwise, the value is replaced by an empty
   string "".

   The parameter Dir indicates the direction in which the OOB message
   containing the Noob value is being sent (1=peer-to-server, 2=server-
   to-peer).  This field is included in the Hoob input to prevent the
   user from accidentally delivering the OOB message back to its
   originator in the rare cases where both OOB directions have been
   negotiated.  The keys (Kms, Kmp, Kms2, and Kmp2) for the HMACs are
   defined in Section 3.5.

   The nonces (Ns, Np, Ns2, Np2, and Noob) and the hash value (NoobId)
   MUST be base64url encoded [RFC4648] when they are used as input to
   the cryptographic functions H or HMAC.  These values and the message
   authentication codes (MACs, MACp, MACs2, and MACp2) MUST also be
   base64url encoded when they are sent as JSON strings in the in-band
   messages.  The values Noob and Hoob in the OOB channel MAY be
   base64url encoded if that is appropriate for the application and the
   OOB channel.  All base64url encoding is done without padding.  The
   base64url-encoded values will naturally consume more space than the
   number of bytes specified above (e.g., a 22-character string for a
   16-byte nonce and a 43-character string for a 32-byte nonce or
   message authentication code).  In the key derivation in Section 3.5,
   on the other hand, the unencoded nonces (raw bytes) are used as input
   to the key derivation function.

   The ServerInfo and PeerInfo are JSON objects with UTF-8 encoding.
   The length of either encoded object as a byte array MUST NOT exceed
   500 bytes.  The format and semantics of these objects MUST be defined
   by the application that uses the EAP-NOOB method.

3.4.  Fast Reconnect and Rekeying

   EAP-NOOB implements fast reconnect ([RFC3748], Section 7.2.1), which
   avoids repeated use of the user-assisted OOB channel.

   The rekeying and the Reconnect Exchange may be needed for several
   reasons.  New EAP output values Main Session Key (MSK) and Extended
   Main Session Key (EMSK) may be needed because of mobility or timeout
   of session keys.  Software or hardware failure or user action may
   also cause the authenticator, EAP server, or peer to lose its
   nonpersistent state data.  The failure would typically be detected by
   the peer or authenticator when session keys are no longer accepted by
   the other endpoint.  Changes in the supported cryptosuites in the EAP
   server or peer may also cause the need for a new key exchange.  When
   the EAP server or peer detects any one of these events, it MUST
   change from the Registered (4) state to the Reconnecting (3) state.
   These state transitions are labeled Mobility/Timeout/Failure in
   Figure 1.  The EAP-NOOB method will then perform the Reconnect
   Exchange the next time when EAP is triggered.

3.4.1.  Persistent EAP-NOOB Association

   To enable rekeying, the EAP server and peer store the session state
   in persistent memory after a successful Completion Exchange.  This
   state data, called "persistent EAP-NOOB association", MUST include at
   least the data fields shown in Table 2.  They are used for
   identifying and authenticating the peer in the Reconnect Exchange.
   When a persistent EAP-NOOB association exists, the EAP server and
   peer are in the Registered (4) state or Reconnecting (3) state, as
   shown in Figure 1.

    | Data Field       | Value                    | Type              |
    | PeerId           | Peer identifier          | UTF-8 string      |
    |                  | allocated by server      | (typically 22     |
    |                  |                          | ASCII characters) |
    | Verp             | Negotiated protocol      | integer           |
    |                  | version                  |                   |
    | Cryptosuitep     | Negotiated cryptosuite   | integer           |
    | CryptosuitepPrev | Previous cryptosuite     | integer           |
    | (at peer only)   |                          |                   |
    | NAI              | NAI assigned by the      | UTF-8 string      |
    |                  | server or configured by  |                   |
    |                  | the user or the default  |                   |
    |                  | NAI "noob@eap-noob.arpa" |                   |
    | Kz               | Persistent key material  | 32 bytes          |
    | KzPrev (at peer  | Previous Kz value        | 32 bytes          |
    | only)            |                          |                   |

                  Table 2: Persistent EAP-NOOB Association

3.4.2.  Reconnect Exchange

   The server chooses the Reconnect Exchange when both the peer and the
   server are in a persistent state and fast reconnection is needed (see
   Section 3.2.1 for details).

   The Reconnect Exchange comprises the common handshake and three
   further EAP-NOOB request-response pairs: one for cryptosuite and
   parameter negotiation, another for the nonce and ECDHE key exchange,
   and the last one for exchanging message authentication codes.  In the
   first request and response (Type=7), the server and peer negotiate a
   protocol version and cryptosuite in the same way as in the Initial
   Exchange.  The server SHOULD NOT offer and the peer MUST NOT accept
   protocol versions or cryptosuites that it knows to be weaker than the
   one currently in the Cryptosuitep field of the persistent EAP-NOOB
   association.  The server SHOULD NOT needlessly change the
   cryptosuites it offers to the same peer because peer devices may have
   limited ability to update their persistent storage.  However, if the
   peer has different values in the Cryptosuitep and CryptosuitepPrev
   fields, it SHOULD also accept offers that are not weaker than
   CryptosuitepPrev.  Note that Cryptosuitep and CryptosuitePrev from
   the persistent EAP-NOOB association are only used to support the
   negotiation as described above; all actual cryptographic operations
   use the newly negotiated cryptosuite.  The request and response
   (Type=7) MAY additionally contain PeerInfo and ServerInfo objects.

   The server then determines the KeyingMode (defined in Section 3.5)
   based on changes in the negotiated cryptosuite and whether it desires
   to achieve forward secrecy or not.  The server SHOULD only select
   KeyingMode 3 when the negotiated cryptosuite differs from the
   Cryptosuitep in the server's persistent EAP-NOOB association,
   although it is technically possible to select this value without
   changing the cryptosuite.  In the second request and response
   (Type=8), the server informs the peer about the KeyingMode and the
   server and peer exchange nonces (Ns2, Np2).  When KeyingMode is 2 or
   3 (rekeying with ECDHE), they also exchange public components of
   ECDHE keys (PKs2, PKp2).  The server ECDHE key MUST be fresh, i.e.,
   not previously used with the same peer, and the peer ECDHE key SHOULD
   be fresh, i.e., not previously used.

   In the third and final request and response (Type=9), the server and
   peer exchange message authentication codes.  Both sides MUST compute
   the keys Kms2 and Kmp2, as defined in Section 3.5, and the message
   authentication codes MACs2 and MACp2, as defined in Section 3.3.2.
   Both sides MUST compare the received message authentication code with
   a locally computed value.

   The rules by which the peer compares the received MACs2 are
   nontrivial because, in addition to authenticating the current
   exchange, MACs2 may confirm the success or failure of a recent
   cryptosuite upgrade.  The peer processes the final request (Type=9)
   as follows:

   1.  The peer first compares the received MACs2 value with one it
       computed using the Kz stored in the persistent EAP-NOOB
       association.  If the received and computed values match, the peer
       deletes any data stored in the CryptosuitepPrev and KzPrev fields
       of the persistent EAP-NOOB association.  It does this because the
       received MACs2 confirms that the peer and server share the same
       Cryptosuitep and Kz, and any previous values must no longer be

   2.  On the other hand, if the peer finds that the received MACs2
       value does not match the one it computed locally with Kz, the
       peer checks whether the KzPrev field in the persistent EAP-NOOB
       association stores a key.  If it does, the peer repeats the key
       derivation (Section 3.5) and local MACs2 computation
       (Section 3.3.2) using KzPrev in place of Kz.  If this second
       computed MACs2 matches the received value, the match indicates
       synchronization failure caused by the loss of the last response
       (Type=9) in a previously attempted cryptosuite upgrade.  In this
       case, the peer rolls back that upgrade by overwriting
       Cryptosuitep with CryptosuitepPrev and Kz with KzPrev in the
       persistent EAP-NOOB association.  It also clears the
       CryptosuitepPrev and KzPrev fields.

   3.  If the received MACs2 matched one of the locally computed values,
       the peer proceeds to send the final response (Type=9).  The peer
       also moves to the Registered (4) state.  When KeyingMode is 1 or
       2, the peer stops here.  When KeyingMode is 3, the peer also
       updates the persistent EAP-NOOB association with the negotiated
       Cryptosuitep and the newly derived Kz value.  To prepare for
       possible synchronization failure caused by the loss of the final
       response (Type=9) during cryptosuite upgrade, the peer copies the
       old Cryptosuitep and Kz values in the persistent EAP-NOOB
       association to the CryptosuitepPrev and KzPrev fields.

   4.  Finally, if the peer finds that the received MACs2 does not match
       either of the two values that it computed locally (or one value
       if no KzPrev was stored), the peer sends an error message (error
       code 4001, see Section 3.6.5), which causes the Reconnect
       Exchange to end in EAP-Failure.

   The server rules for processing the final message are simpler than
   the peer rules because the server does not store previous keys and it
   never rolls back a cryptosuite upgrade.  Upon receiving the final
   response (Type=9), the server compares the received value of MACp2
   with one it computes locally.  If the values match, the Reconnect
   Exchange ends in EAP-Success.  When KeyingMode is 3, the server also
   updates Cryptosuitep and Kz in the persistent EAP-NOOB association.
   On the other hand, if the server finds that the values do not match,
   it sends an error message (error code 4001), and the Reconnect
   Exchange ends in EAP-Failure.

   The endpoints MAY send updated NewNAI, ServerInfo, and PeerInfo
   objects in the Reconnect Exchange.  When there is no update to the
   values, they SHOULD omit this information from the messages.  If the
   NewNAI was sent, each side updates NAI in the persistent EAP-NOOB
   association when moving to the Registered (4) state.

        EAP Peer                                        EAP Server
          |       ...continuing from common handshake        |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=7,Vers,PeerId,Cryptosuites,           |
          |       [NewNAI],[ServerInfo])                     |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=7,Verp,PeerId,Cryptosuitep,[PeerInfo])|
          |                                                  |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=8,PeerId,KeyingMode,[PKs2],Ns2)       |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=8,PeerId,[PKp2],Np2)                  |
          |                                                  |
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |      (Type=9,PeerId,MACs2)                       |
          |                                                  |
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |      (Type=9,PeerId,MACp2)                       |
          |                                                  |
          |                                                  |
          |<----------- EAP-Success -------------------------|
          |                                                  |

                        Figure 8: Reconnect Exchange

3.4.3.  User Reset

   As shown in the association state machine in Figure 1, the only
   specified way for the association to return from the Registered (4)
   state to the Unregistered (0) state is through user-initiated reset.
   After the reset, a new OOB message will be needed to establish a new
   association between the EAP server and peer.  Typical situations in
   which the user reset is required are when the other side has
   accidentally lost the persistent EAP-NOOB association data or when
   the peer device is decommissioned.

   The server could detect that the peer is in the Registered or
   Reconnecting state, but the server itself is in one of the ephemeral
   states 0..2 (including situations where the server does not recognize
   the PeerId).  In this case, effort should be made to recover the
   persistent server state, for example, from a backup storage --
   especially if many peer devices are similarly affected.  If that is
   not possible, the EAP server SHOULD log the error or notify an
   administrator.  The only way to continue from such a situation is by
   having the user reset the peer device.

   On the other hand, if the peer is in any of the ephemeral states
   0..2, including the Unregistered state, the server will treat the
   peer as a new peer device and allocate a new PeerId to it.  The
   PeerInfo can be used by the user as a clue to which physical device
   has lost its state.  However, there is no secure way of matching the
   "new" peer with the old PeerId without repeating the OOB Step.  This
   situation will be resolved when the user performs the OOB Step and
   thus identifies the physical peer device.  The server user interface
   MAY support situations where the "new" peer is actually a previously
   registered peer that has been reset by a user or otherwise lost its
   persistent data.  In those cases, the user could choose to merge the
   new peer identity with the old one in the server.  The alternative is
   to treat the device just like a new peer.

3.5.  Key Derivation

   EAP-NOOB derives the EAP output values MSK and EMSK and other secret
   keying material from the output of an Ephemeral Elliptic Curve
   Diffie-Hellman (ECDHE) algorithm following the NIST specification
   [NIST-DH].  In NIST terminology, we use a C(2e, 0s, ECC CDH) scheme,
   i.e., two ephemeral keys and no static keys.  In the Initial Exchange
   and Reconnect Exchange, the server and peer compute the ECDHE shared
   secret Z, as defined in Section 6.1.2 of the NIST specification
   [NIST-DH].  In the Completion Exchange and Reconnect Exchange, the
   server and peer compute the secret keying material from Z with the
   one-step key derivation function (KDF) defined in Section of
   the NIST specification.  The auxiliary function H is a hash function,
   and it is taken from the negotiated cryptosuite.

        | KeyingMode | Description                                |
        | 0          | Completion Exchange (always with ECDHE)    |
        | 1          | Reconnect Exchange, rekeying without ECDHE |
        | 2          | Reconnect Exchange, rekeying with ECHDE,   |
        |            | no change in cryptosuite                   |
        | 3          | Reconnect Exchange, rekeying with ECDHE,   |
        |            | new cryptosuite negotiated                 |

                           Table 3: Keying Modes

   The key derivation has four different modes (KeyingMode), which are
   specified in Table 3.  Table 4 defines the inputs to KDF in each

   In the Completion Exchange (KeyingMode=0), the input Z comes from the
   preceding Initial exchange.  The KDF takes some additional inputs
   (FixedInfo), for which we use the concatenation format defined in
   Section of the NIST specification [NIST-DH].  FixedInfo
   consists of the AlgorithmId, PartyUInfo, PartyVInfo, and SuppPrivInfo
   fields.  The first three fields are fixed-length bit strings, and
   SuppPrivInfo is a variable-length string with a one-byte Datalength
   counter.  AlgorithmId is the fixed-length, 8-byte ASCII string "EAP-
   NOOB".  The other input values are the server and peer nonces.  In
   the Completion Exchange, the inputs also include the secret nonce
   Noob from the OOB message.

   In the simplest form of the Reconnect Exchange (KeyingMode=1), fresh
   nonces are exchanged, but no ECDHE keys are sent.  In this case,
   input Z to the KDF is replaced with the shared key Kz from the
   persistent EAP-NOOB association.  The result is rekeying without the
   computational cost of the ECDHE exchange but also without forward

   When forward secrecy is desired in the Reconnect Exchange
   (KeyingMode=2 or KeyingMode=3), both nonces and ECDHE keys are
   exchanged.  Input Z is the fresh shared secret from the ECDHE
   exchange with PKs2 and PKp2.  The inputs also include the shared
   secret Kz from the persistent EAP-NOOB association.  This binds the
   rekeying output to the previously authenticated keys.

   | KeyingMode              | KDF input    | Value         | Length   |
   |                         | field        |               | (bytes)  |
   | 0 Completion            | Z            | ECDHE shared  | variable |
   |                         |              | secret from   |          |
   |                         |              | PKs and PKp   |          |
   |                         +--------------+---------------+----------+
   |                         | AlgorithmId  | "EAP-NOOB"    | 8        |
   |                         +--------------+---------------+----------+
   |                         | PartyUInfo   | Np            | 32       |
   |                         +--------------+---------------+----------+
   |                         | PartyVInfo   | Ns            | 32       |
   |                         +--------------+---------------+----------+
   |                         | SuppPubInfo  | (not          |          |
   |                         |              | allowed)      |          |
   |                         +--------------+---------------+----------+
   |                         | SuppPrivInfo | Noob          | 16       |
   | 1 Reconnect, rekeying   | Z            | Kz            | 32       |
   | without ECDHE           +--------------+---------------+----------+
   |                         | AlgorithmId  | "EAP-NOOB"    | 8        |
   |                         +--------------+---------------+----------+
   |                         | PartyUInfo   | Np2           | 32       |
   |                         +--------------+---------------+----------+
   |                         | PartyVInfo   | Ns2           | 32       |
   |                         +--------------+---------------+----------+
   |                         | SuppPubInfo  | (not          |          |
   |                         |              | allowed)      |          |
   |                         +--------------+---------------+----------+
   |                         | SuppPrivInfo | (null)        | 0        |
   | 2 or 3 Reconnect,       | Z            | ECDHE shared  | variable |
   | rekeying, with ECDHE,   |              | secret from   |          |
   | same or new cryptosuite |              | PKs2 and      |          |
   |                         |              | PKp2          |          |
   |                         +--------------+---------------+----------+
   |                         | AlgorithmId  | "EAP-NOOB"    | 8        |
   |                         +--------------+---------------+----------+
   |                         | PartyUInfo   | Np2           | 32       |
   |                         +--------------+---------------+----------+
   |                         | PartyVInfo   | Ns2           | 32       |
   |                         +--------------+---------------+----------+
   |                         | SuppPubInfo  | (not          |          |
   |                         |              | allowed)      |          |
   |                         +--------------+---------------+----------+
   |                         | SuppPrivInfo | Kz            | 32       |

                       Table 4: Key Derivation Input

   Table 5 defines how the output bytes of the KDF are used.  In
   addition to the EAP output values MSK and EMSK, the server and peer
   derive another shared secret key AMSK (Application Main Session Key),
   which MAY be used for application-layer security.  Further output
   bytes are used internally by EAP-NOOB for the message authentication
   keys (Kms, Kmp, Kms2, and Kmp2).

   The Completion Exchange (KeyingMode=0) produces the shared secret Kz,
   which the server and peer store in the persistent EAP-NOOB
   association.  When a new cryptosuite is negotiated in the Reconnect
   Exchange (KeyingMode=3), it similarly produces a new Kz.  In that
   case, the server and peer update both the cryptosuite and Kz in the
   persistent EAP-NOOB association.  Additionally, the peer stores the
   previous Cryptosuitep and Kz values in the CryptosuitepPrev and
   KzPrev fields of the persistent EAP-NOOB association.

    | KeyingMode                   | KDF output | Used as  | Length  |
    |                              | bytes      |          | (bytes) |
    | 0 Completion                 | 0..63      | MSK      | 64      |
    |                              +------------+----------+---------+
    |                              | 64..127    | EMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 128..191   | AMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 192..223   | MethodId | 32      |
    |                              +------------+----------+---------+
    |                              | 224..255   | Kms      | 32      |
    |                              +------------+----------+---------+
    |                              | 256..287   | Kmp      | 32      |
    |                              +------------+----------+---------+
    |                              | 288..319   | Kz       | 32      |
    | 1 or 2 Reconnect, rekeying   | 0..63      | MSK      | 64      |
    | without ECDHE, or with ECDHE +------------+----------+---------+
    | and unchanged cryptosuite    | 64..127    | EMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 128..191   | AMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 192..223   | MethodId | 32      |
    |                              +------------+----------+---------+
    |                              | 224..255   | Kms2     | 32      |
    |                              +------------+----------+---------+
    |                              | 256..287   | Kmp2     | 32      |
    | 3 Reconnect, rekeying with   | 0..63      | MSK      | 64      |
    | ECDHE, new cryptosuite       +------------+----------+---------+
    |                              | 64..127    | EMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 128..191   | AMSK     | 64      |
    |                              +------------+----------+---------+
    |                              | 192..223   | MethodId | 32      |
    |                              +------------+----------+---------+
    |                              | 224..255   | Kms2     | 32      |
    |                              +------------+----------+---------+
    |                              | 256..287   | Kmp2     | 32      |
    |                              +------------+----------+---------+
    |                              | 288..319   | Kz       | 32      |

                      Table 5: Key Derivation Output

   Finally, every EAP method must export a Server-Id, Peer-Id, and
   Session-Id [RFC5247].  In EAP-NOOB, the exported Peer-Id is the
   PeerId that the server has assigned to the peer.  The exported
   Server-Id is a zero-length string (i.e., null string) because EAP-
   NOOB neither knows nor assigns any server identifier.  The exported
   Session-Id is created by concatenating the one-byte Type-Code 0x38
   (decimal value 56) with the MethodId, which is obtained from the KDF
   output, as shown in Table 5.

3.6.  Error Handling

   Various error conditions in EAP-NOOB are handled by sending an error
   notification message (Type=0) instead of a next EAP request or
   response message.  Both the EAP server and the peer may send the
   error notification, as shown in Figures 9 and 10.  After sending or
   receiving an error notification, the server MUST send an EAP-Failure
   (as required by [RFC3748], Section 4.2).  The notification MAY
   contain an ErrorInfo field, which is a UTF-8-encoded text string with
   a maximum length of 500 bytes.  It is used for sending descriptive
   information about the error for logging and debugging purposes.

        EAP Peer                                        EAP Server
          ...                                                ...
          |                                                  |
          |<----------- EAP-Request/EAP-NOOB ----------------|
          |        (Type=0,[PeerId],ErrorCode,[ErrorInfo])   |
          |                                                  |
          |                                                  |
          |<----------- EAP-Failure -------------------------|
          |                                                  |

              Figure 9: Error Notification from Server to Peer

        EAP Peer                                        EAP Server
          ...                                                ...
          |                                                  |
          |------------ EAP-Response/EAP-NOOB -------------->|
          |        (Type=0,[PeerId],ErrorCode,[ErrorInfo])   |
          |                                                  |
          |                                                  |
          |<----------- EAP-Failure -------------------------|
          |                                                  |

             Figure 10: Error Notification from Peer to Server

   After the exchange fails due to an error notification, the server and
   peer set the association state as follows.  In the Initial Exchange,
   both the sender and recipient of the error notification MUST set the
   association state to the Unregistered (0) state.  In the Waiting
   Exchange and Completion Exchange, each side MUST remain in its old
   state as if the failed exchange had not taken place, with the
   exception that the recipient of error code 2003 processes it as
   specified in Section 3.2.4.  In the Reconnect Exchange, both sides
   MUST set the association state to the Reconnecting (3) state.

   Errors that occur in the OOB channel are not explicitly notified in-

3.6.1.  Invalid Messages

   If the NAI structure is invalid, the server SHOULD send the error
   code 1001 to the peer.  The recipient of an EAP-NOOB request or
   response SHOULD send the following error codes back to the sender:
   1002 if it cannot parse the message as a JSON object or the top-level
   JSON object has missing or unrecognized members; 1003 if a data field
   has an invalid value, such as an integer out of range, and there is
   no more specific error code available; 1004 if the received message
   type was unexpected in the current state; 2004 if the PeerId has an
   unexpected value; 2003 if the NoobId is not recognized; and 1005 if
   the ECDHE key is invalid.

3.6.2.  Unwanted Peer

   The preferred way for the EAP server to rate limit EAP-NOOB
   connections from a peer is to use the SleepTime parameter in the
   Waiting Exchange.  However, if the EAP server receives repeated EAP-
   NOOB connections from a peer that apparently should not connect to
   this server, the server MAY indicate that the connections are
   unwanted by sending the error code 2001.  After receiving this error
   message, the peer MAY refrain from reconnecting to the same EAP
   server, and, if possible, both the EAP server and peer SHOULD
   indicate this error condition to the user or server administrator.
   However, in order to avoid persistent denial of service, peer devices
   that are unable to alert a user SHOULD continue to try to reconnect
   infrequently (e.g., approximately every 3600 seconds).

3.6.3.  State Mismatch

   In the states indicated by "-" in Table 14 in Appendix A, user action
   is required to reset the association state or to recover it, for
   example, from backup storage.  In those cases, the server sends the
   error code 2002 to the peer.  If possible, both the EAP server and
   peer SHOULD indicate this error condition to the user or server

3.6.4.  Negotiation Failure

   If there is no matching protocol version, the peer sends the error
   code 3001 to the server.  If there is no matching cryptosuite, the
   peer sends the error code 3002 to the server.  If there is no
   matching OOB direction, the peer sends the error code 3003 to the

   In practice, there is no way of recovering from these errors without
   software or hardware changes.  If possible, both the EAP server and
   peer SHOULD indicate these error conditions to the user.

3.6.5.  Cryptographic Verification Failure

   If the receiver of the OOB message detects an unrecognized PeerId or
   incorrect fingerprint (Hoob) in the OOB message, the receiver MUST
   remain in the Waiting for OOB (1) state as if no OOB message was
   received.  The receiver SHOULD indicate the failure to accept the OOB
   message to the user.  No in-band error message is sent.

   Note that if the OOB message was delivered from the server to the
   peer and the peer does not recognize the PeerId, the likely cause is
   that the user has unintentionally delivered the OOB message to the
   wrong peer device.  If possible, the peer SHOULD indicate this to the
   user; however, the peer device may not have the capability for many
   different error indications to the user, and it MAY use the same
   indication as in the case of an incorrect fingerprint.

   The rationale for the above is that the invalid OOB message could
   have been presented to the receiver by mistake or intentionally by a
   malicious party; thus, it should be ignored in the hope that the
   honest user will soon deliver a correct OOB message.

   If the EAP server or peer detects an incorrect message authentication
   code (MACs, MACp, MACs2, or MACp2), it sends the error code 4001 to
   the other side.  As specified in the beginning of Section 3.6, the
   failed Completion Exchange will not result in server or peer state
   changes, while an error in the Reconnect Exchange will put both sides
   to the Reconnecting (3) state and thus lead to another reconnect

   The rationale for this is that the invalid cryptographic message may
   have been spoofed by a malicious party; thus, it should be ignored.
   In particular, a spoofed message on the in-band channel should not
   force the honest user to perform the OOB Step again.  In practice,
   however, the error may be caused by other failures, such as a
   software bug.  For this reason, the EAP server MAY limit the rate of
   peer connections with SleepTime after the above error.  Also, there
   SHOULD be a way for the user to reset the peer to the Unregistered
   (0) state so that the OOB Step can be repeated as the last resort.

3.6.6.  Application-Specific Failure

   Applications MAY define new error messages for failures that are
   specific to the application or to one type of OOB channel.  They MAY
   also use the generic application-specific error code 5001 or the
   error codes 5002 and 5004, which have been reserved for indicating
   invalid data in the ServerInfo and PeerInfo fields, respectively.
   Additionally, anticipating OOB channels that make use of a URL, the
   error code 5003 has been reserved for indicating an invalid server

4.  ServerInfo and PeerInfo Contents

   The ServerInfo and PeerInfo fields in the Initial Exchange and
   Reconnect Exchange enable the server and peer, respectively, to send
   information about themselves to the other endpoint.  They contain
   JSON objects whose structure may be specified separately for each
   application and each type of OOB channel.  ServerInfo and PeerInfo
   MAY contain auxiliary data needed for the OOB channel messaging and
   for EAP channel binding (see Section 6.7).  This section describes
   the optional initial data fields for ServerInfo and PeerInfo
   registered by this specification.  Further specifications may request
   new application-specific ServerInfo and PeerInfo data fields from
   IANA (see Sections 5.4 and 5.5).

   | Data Field     | Description                                     |
   | Type           | Type-tag string that can be used by the peer as |
   |                | a hint for how to interpret the ServerInfo      |
   |                | contents.                                       |
   | ServerName     | String that may be used to aid human            |
   |                | identification of the server.                   |
   | ServerURL      | Prefix string when the OOB message is formatted |
   |                | as a URL, as suggested in Appendix D.           |
   | SSIDList       | List of IEEE 802.11 wireless network service    |
   |                | set identifier (SSID) strings used for roaming  |
   |                | support, as suggested in Appendix C.  JSON      |
   |                | array of ASCII-encoded SSID strings.            |
   | Base64SSIDList | List of IEEE 802.11 wireless network identifier |
   |                | (SSID) strings used for roaming support, as     |
   |                | suggested in Appendix C.  JSON array of SSIDs,  |
   |                | each of which is base64url-encoded without      |
   |                | padding.  Peers SHOULD send at most one of the  |
   |                | fields SSIDList and Base64SSIDList in PeerInfo, |
   |                | and the server SHOULD ignore SSIDList if        |
   |                | Base64SSIDList is included.                     |

                     Table 6: ServerInfo Data Fields

   | Data Field   | Description                                       |
   | Type         | Type-tag string that can be used by the server as |
   |              | a hint for how to interpret the PeerInfo          |
   |              | contents.                                         |
   | PeerName     | String that may be used to aid human              |
   |              | identification of the peer.                       |
   | Manufacturer | Manufacturer or brand string.                     |
   | Model        | Manufacturer-specified model string.              |
   | SerialNumber | Manufacturer-assigned serial number.              |
   | MACAddress   | Peer link-layer 48-bit extended unique identifier |
   |              | (EUI-48) in the 12-digit base-16 form [EUI-48].   |
   |              | The string MAY be in upper or lower case and MAY  |
   |              | include additional colon ':' or dash '-'          |
   |              | characters that MUST be ignored by the server.    |
   | SSID         | IEEE 802.11 network SSID for channel binding.     |
   |              | The SSID is an ASCII string.                      |
   | Base64SSID   | IEEE 802.11 network SSID for channel binding.     |
   |              | The SSID is base64url encoded.  Peer SHOULD send  |
   |              | at most one of the fields SSID and Base64SSID in  |
   |              | PeerInfo, and the server SHOULD ignore SSID if    |
   |              | Base64SSID is included.                           |
   | BSSID        | Wireless network basic service set identifier     |
   |              | (BSSID) (EUI-48) in the 12-digit base-16 form     |
   |              | [EUI-48] for channel binding.  The string MAY be  |
   |              | in upper or lower case and MAY include additional |
   |              | colon ':' or dash '-' characters that MUST be     |
   |              | ignored by the server.                            |

                      Table 7: PeerInfo Data Fields

5.  IANA Considerations

   This section provides information regarding registration of values
   related to the EAP-NOOB method, in accordance with [RFC8126].

   The EAP Method Type for EAP-NOOB (value 56) has been assigned in the
   "Method Types" subregistry of the "Extensible Authentication Protocol
   (EAP) Registry".

   Per this memo, IANA has created and will maintain a new registry
   entitled "Nimble Out-of-Band Authentication for EAP Parameters (EAP-
   NOOB)" in the Extensible Authentication Protocol (EAP) category.
   Also, IANA has created and will maintain the subregistries defined in
   the following subsections.

5.1.  Cryptosuites

   IANA has created and will maintain a new subregistry entitled "EAP-
   NOOB Cryptosuites" in the "Nimble Out-of-Band Authentication for EAP
   Parameters (EAP-NOOB)" registry.  Cryptosuites are identified by an
   integer.  Each cryptosuite MUST specify an ECDHE curve for the key
   exchange, encoding of the ECDHE public key as a JWK object, and a
   cryptographic hash function for the fingerprint and HMAC computation
   and key derivation.  The hash value output by the cryptographic hash
   function MUST be at least 32 bytes in length.  The initial values for
   this registry are:

      | Cryptosuite | Algorithms                                    |
      | 0           | Reserved                                      |
      | 1           | ECDHE curve Curve25519 [RFC7748], public-key  |
      |             | format [RFC7517], hash function SHA-256       |
      |             | [RFC6234].  The JWK encoding of Curve25519    |
      |             | public key is defined in [RFC8037].  For      |
      |             | clarity, the "crv" parameter is "X25519", the |
      |             | "kty" parameter is "OKP", and the public-key  |
      |             | encoding contains only an x-coordinate.       |
      | 2           | ECDHE curve NIST P-256 [FIPS186-4], public-   |
      |             | key format [RFC7517], hash function SHA-256   |
      |             | [RFC6234].  The JWK encoding of NIST P-256    |
      |             | public key is defined in [RFC7518].  For      |
      |             | clarity, the "crv" parameter is "P-256", the  |
      |             | "kty" parameter is "EC", and the public-key   |
      |             | encoding has both an x and y coordinate, as   |
      |             | defined in Section 6.2.1 of [RFC7518].        |

                       Table 8: EAP-NOOB Cryptosuites

   EAP-NOOB implementations MUST support Cryptosuite 1.  Support for
   Cryptosuite 2 is RECOMMENDED.  An example of a Cryptosuite 1 public-
   key encoded as a JWK object is given below.  (Line breaks are for
   readability only.)


   Assignment of new values for new cryptosuites MUST be done through
   IANA with "Specification Required", as defined in [RFC8126].

5.2.  Message Types

   IANA has created and will maintain a new subregistry entitled "EAP-
   NOOB Message Types" in the "Nimble Out-of-Band Authentication for EAP
   Parameters (EAP-NOOB)" registry.  EAP-NOOB request and response pairs
   are identified by an integer Message Type.  The initial values for
   this registry are:

     | Message | Used in    | Purpose                                |
     | Type    | Exchange   |                                        |
     | 0       | Error      | Error notification                     |
     | 1       | All        | PeerId and PeerState discovery         |
     |         | exchanges  |                                        |
     | 2       | Initial    | Version, cryptosuite, and parameter    |
     |         |            | negotiation                            |
     | 3       | Initial    | Exchange of ECDHE keys and nonces      |
     | 4       | Waiting    | Indication to the peer that the server |
     |         |            | has not yet received an OOB message    |
     | 5       | Completion | NoobId discovery                       |
     | 6       | Completion | Authentication and key confirmation    |
     |         |            | with HMAC                              |
     | 7       | Reconnect  | Version, cryptosuite, and parameter    |
     |         |            | negotiation                            |
     | 8       | Reconnect  | Exchange of ECDHE keys and nonces      |
     | 9       | Reconnect  | Authentication and key confirmation    |
     |         |            | with HMAC                              |

                      Table 9: EAP-NOOB Message Types

   Assignment of new values for new Message Types MUST be done through
   IANA with "Specification Required", as defined in [RFC8126].

5.3.  Error Codes

   IANA has created and will maintain a new subregistry entitled "EAP-
   NOOB Error codes" in the "Nimble Out-of-Band Authentication for EAP
   Parameters (EAP-NOOB)" registry.  Cryptosuites are identified by an
   integer.  The initial values for this registry are:

        | Error code | Purpose                                   |
        | 1001       | Invalid NAI                               |
        | 1002       | Invalid message structure                 |
        | 1003       | Invalid data                              |
        | 1004       | Unexpected message type                   |
        | 1005       | Invalid ECDHE key                         |
        | 2001       | Unwanted peer                             |
        | 2002       | State mismatch, user action required      |
        | 2003       | Unrecognized OOB message identifier       |
        | 2004       | Unexpected peer identifier                |
        | 3001       | No mutually supported protocol version    |
        | 3002       | No mutually supported cryptosuite         |
        | 3003       | No mutually supported OOB direction       |
        | 4001       | HMAC verification failure                 |
        | 5001       | Application-specific error                |
        | 5002       | Invalid server info                       |
        | 5003       | Invalid server URL                        |
        | 5004       | Invalid peer info                         |
        | 6001-6999  | Reserved for Private and Experimental Use |

                      Table 10: EAP-NOOB Error Codes

   Assignment of new error codes MUST be done through IANA with
   "Specification Required", as defined in [RFC8126], except for the
   range 6001-6999.  This range is reserved for "Private Use" and
   "Experimental Use", both locally and on the open Internet.

5.4.  ServerInfo Data Fields

   IANA has created and will maintain a new subregistry entitled "EAP-
   NOOB ServerInfo Data Fields" in the "Nimble Out-of-Band
   Authentication for EAP Parameters (EAP-NOOB)" registry.  The initial
   values for this registry are:

                 | Data Field     | Specification       |
                 | Type           | RFC 9140, Section 4 |
                 | ServerName     | RFC 9140, Section 4 |
                 | ServerURL      | RFC 9140, Section 4 |
                 | SSIDList       | RFC 9140, Section 4 |
                 | Base64SSIDList | RFC 9140, Section 4 |

                     Table 11: ServerInfo Data Fields

   Assignment of new values for new ServerInfo data fields MUST be done
   through IANA with "Specification Required", as defined in [RFC8126].

5.5.  PeerInfo Data Fields

   IANA is requested to create and maintain a new subregistry entitled
   "EAP-NOOB PeerInfo Data Fields" in the "Nimble Out-of-Band
   Authentication for EAP Parameters (EAP-NOOB)" registry.  The initial
   values for this registry are:

                  | Data Field   | Specification       |
                  | Type         | RFC 9140, Section 4 |
                  | PeerName     | RFC 9140, Section 4 |
                  | Manufacturer | RFC 9140, Section 4 |
                  | Model        | RFC 9140, Section 4 |
                  | SerialNumber | RFC 9140, Section 4 |
                  | MACAddress   | RFC 9140, Section 4 |
                  | SSID         | RFC 9140, Section 4 |
                  | Base64SSID   | RFC 9140, Section 4 |
                  | BSSID        | RFC 9140, Section 4 |

                      Table 12: PeerInfo Data Fields

   Assignment of new values for new PeerInfo data fields MUST be done
   through IANA with "Specification Required", as defined in [RFC8126].

5.6.  Domain Name Reservation

   The special-use domain "eap-noob.arpa" has been registered in the
   .arpa registry (https://www.iana.org/domains/arpa) and the "Special-
   Use Domain Names" registry (https://www.iana.org/assignments/special-

5.7.  Guidance for Designated Experts

   Experts SHOULD be conservative in the allocation of new Cryptosuites.
   Experts MUST ascertain that the requested values match the current
   Crypto Forum Research Group (CFRG) guidance on cryptographic
   algorithm security.  Experts MUST ensure that any new Cryptosuites
   fully specify the encoding of the ECDHE public key and should include
   details, such as the value of the "kty" (key type) parameter when JWK
   [RFC7517] encoding is used.

   Experts SHOULD be conservative in the allocation of new Message
   Types.  Experts SHOULD ascertain that a well-defined specification
   for the new Message Type is permanently and publicly available.

   Experts SHOULD be conservative in the allocation of new Error codes,
   since the 6001-6999 range is already reserved for private and
   experimental use.

   Experts MAY be liberal in the allocation of new ServerInfo and
   PeerInfo data fields.  Experts MUST ensure that the data field
   requested has a unique name that is not easily confused with existing
   registrations.  For example, requests for a new PeerInfo data field
   "ssid" should be rejected even though it is unique because it can be
   confused with the existing registration of "SSID".  Experts MUST
   ensure that a suitable Description for the data field is available.

6.  Security Considerations

   EAP-NOOB is an authentication and key derivation protocol; thus,
   security considerations can be found in most sections of this
   specification.  In the following, we explain the protocol design and
   highlight some other special considerations.

6.1.  Authentication Principle

   EAP-NOOB establishes a shared secret with an authenticated ECDHE key
   exchange.  The mutual authentication in EAP-NOOB is based on two
   separate features, both conveyed in the OOB message.  The first
   authentication feature is the secret nonce Noob.  The peer and server
   use this secret in the Completion Exchange to mutually authenticate
   the session key previously created with ECDHE.  The message
   authentication codes computed with the secret nonce Noob are alone
   sufficient for authenticating the key exchange.  The second
   authentication feature is the integrity-protecting fingerprint Hoob.
   Its purpose is to prevent impersonation attacks even in situations
   where the attacker is able to eavesdrop on the OOB channel and the
   nonce Noob is compromised.  In some human-assisted OOB channels, such
   as human-perceptible audio or a user-typed URL, it may be easier to
   detect tampering than disclosure of the OOB message, and such
   applications benefit from the second authentication feature.

   The additional security provided by the cryptographic fingerprint
   Hoob is somewhat intricate to understand.  The endpoint that receives
   the OOB message uses Hoob to verify the integrity of the ECDHE
   exchange.  Thus, the OOB receiver can detect impersonation attacks
   that may have happened on the in-band channel.  The other endpoint,
   however, is not equally protected because the OOB message and
   fingerprint are sent only in one direction.  Some protection to the
   OOB sender is afforded by the fact that the user may notice the
   failure of the association at the OOB receiver and therefore reset
   the OOB sender.  Other device-pairing protocols have solved similar
   situations by requiring the user to confirm to the OOB sender that
   the association was accepted by the OOB receiver, e.g., with a button
   press on the sender side.  Applications MAY implement EAP-NOOB in
   this way.  Nevertheless, since EAP-NOOB was designed to work with
   strictly one-directional OOB communication and the fingerprint is
   only the second authentication feature, the EAP-NOOB specification
   does not mandate such explicit confirmation to the OOB sender.

   To summarize, EAP-NOOB uses the combined protection of the secret
   nonce Noob and the cryptographic fingerprint Hoob, both conveyed in
   the OOB message.  The secret nonce Noob alone is sufficient for
   mutual authentication unless the attacker can eavesdrop on it from
   the OOB channel.  Even if an attacker is able to eavesdrop on the
   secret nonce Noob, it nevertheless cannot perform a full
   impersonation attack on the in-band channel because a mismatching
   fingerprint would alert the OOB receiver, which would reject the OOB
   message.  The attacker that eavesdropped on the secret nonce can
   impersonate the OOB receiver to the OOB sender.  If it does, the
   association will appear to be complete only on the OOB sender side,
   and such situations have to be resolved by the user by resetting the
   OOB sender to the initial state.

   The expected use cases for EAP-NOOB are ones where it replaces a
   user-entered access credential in IoT appliances.  In wireless
   network access without EAP, the user-entered credential is often a
   passphrase that is shared by all the network stations.  The advantage
   of an EAP-based solution, including EAP-NOOB, is that it establishes
   a different shared secret for each peer device, which makes the
   system more resilient against device compromise.  Another advantage
   is that it is possible to revoke the security association for an
   individual device on the server side.

   Forward secrecy during fast reconnect in EAP-NOOB is optional.  The
   Reconnect Exchange in EAP-NOOB provides forward secrecy only if both
   the server and peer send their fresh ECDHE keys.  This allows both
   the server and peer to limit the frequency of the costly computation
   that is required for forward secrecy.  The server MAY adjust the
   frequency of its attempts at ECDHE rekeying based on what it knows
   about the peer's computational capabilities.

   Another way in which some servers may control their computational
   load is to reuse the same ECDHE key for all peers over a short
   server-specific time window.  In that case, forward secrecy will be
   achieved only after the server updates its ECDHE key, which may be a
   reasonable trade-off between security and performance.  However, the
   server MUST NOT reuse the same ECDHE key with the same peer when
   rekeying with ECDHE (KeyingMode=2 or KeyingMode=3).  Instead, it can
   simply not send an ECDHE key (KeyingMode=1).

   The users delivering the OOB messages will often authenticate
   themselves to the EAP server, e.g., by logging into a secure web page
   or API.  In this case, the server can associate the peer device with
   the user account.  Applications that make use of EAP-NOOB can use
   this information for configuring the initial owner of the freshly
   registered device.

6.2.  Identifying Correct Endpoints

   Potential weaknesses in EAP-NOOB arise from the fact that the user
   must physically identify the correct peer device.  If the user
   mistakenly delivers the OOB message from the wrong peer device to the
   server, the server may create an association with the wrong peer.
   The reliance on the user in identifying the correct endpoints is an
   inherent property of user-assisted, out-of-band authentication.  To
   understand the potential consequences of the user mistake, we need to
   consider a few different scenarios.  In the first scenario, there is
   no malicious party, and the user makes an accidental mistake between
   two out-of-the-box devices that are both ready to be registered to a
   server.  If the user delivers the OOB message from the wrong device
   to the server, confusion may arise but usually no security issues.
   In the second scenario, an attacker intentionally tricks the user,
   for example, by substituting the original peer device with a
   compromised one.  This is essentially a supply chain attack where the
   user accepts a compromised physical device.

   There is also a third scenario, in which an opportunistic attacker
   tries to take advantage of the user's accidental mistake.  For
   example, the user could play an audio or a blinking LED message to a
   device that is not expecting to receive it.  In simple security
   bootstrapping solutions that transfer a primary key to the device via
   the OOB channel, the device could misuse or leak the accidentally
   received primary key.  EAP-NOOB is not vulnerable to such
   opportunistic attackers because the OOB message has no value to
   anyone who did not take part in the corresponding Initial Exchange.

   One mechanism that can mitigate user mistakes is certification of
   peer devices.  A certificate or an attestation token (e.g., [TLS-CWT]
   and [RATS-EAT]) can convey to the server authentic identifiers and
   attributes, such as model and serial number, of the peer device.
   Compared to a fully certificate-based authentication, however, EAP-
   NOOB can be used without trusted third parties and does not require
   the user to know any identifier of the peer device; physical access
   to the device is sufficient for bootstrapping with EAP-NOOB.

   Similarly, the attacker can try to trick the user into delivering the
   OOB message to the wrong server so that the peer device becomes
   associated with the wrong server.  If the EAP server is accessed
   through a web user interface, the attack is akin to phishing attacks
   where the user is tricked into accessing the wrong URL and wrong web
   page.  OOB implementation with a dedicated app on a mobile device,
   which communicates with a server API at a preconfigured URL, can
   protect against such attacks.

   After the device registration, an attacker could clone the device
   identity by copying the keys from the persistent EAP-NOOB association
   into another device.  The attacker can be an outsider who gains
   access to the keys or the device owner who wants to have two devices
   matching the same registration.  The cloning threats can be mitigated
   by creating the cryptographic keys and storing the persistent EAP-
   NOOB association on the peer device in a secure hardware component
   such as a trusted execution environment (TEE).  Furthermore, remote
   attestation on the application level could provide assurance to the
   server that the device has not been cloned.  Reconnect Exchange with
   a new cryptosuite (KeyingMode=3) will also disconnect all but the
   first clone that performs the update.

6.3.  Trusted Path Issues and Misbinding Attacks

   Another potential threat is spoofed user input or output on the peer
   device.  When the user is delivering the OOB message to or from the
   correct peer device, a trusted path between the user and the peer
   device is needed.  That is, the user must communicate directly with
   an authentic operating system and EAP-NOOB implementation in the peer
   device and not with a spoofed user interface.  Otherwise, a
   registered device that is under the control of the attacker could
   emulate the behavior of an unregistered device.  The secure path can
   be implemented, for example, by having the user press a reset button
   to return the device to the Unregistered (0) state and to invoke a
   trusted UI.  The problem with such trusted paths is that they are not
   standardized across devices.

   Another potential consequence of a spoofed UI is the misbinding
   attack where the user tries to register a correct but compromised
   device, which tricks the user into registering another
   (uncompromised) device instead.  For example, the compromised device
   might have a malicious, full-screen app running, which presents to
   the user QR codes copied, in real time, from another device's screen.
   If the unwitting user scans the QR code and delivers the OOB message
   in it to the server, the wrong device may become registered in the
   server.  Such misbinding vulnerabilities arise because the user does
   not have any secure way of verifying that the in-band cryptographic
   handshake and the out-of-band physical access are terminated at the
   same physical device.  Sethi et al.  [Sethi19] analyze the misbinding
   threat against device-pairing protocols and also EAP-NOOB.
   Essentially, all protocols where the authentication relies on the
   user's physical access to the device are vulnerable to misbinding,
   including EAP-NOOB.

   A standardized trusted path for communicating directly with the
   trusted computing base in a physical device would mitigate the
   misbinding threat, but such paths rarely exist in practice.  Careful
   asset tracking on the server side can also prevent most misbinding
   attacks if the peer device sends its identifiers or attributes in the
   PeerInfo field and the server compares them with the expected values.
   The wrong but uncompromised device's PeerInfo will not match the
   expected values.  Device certification by the manufacturer can
   further strengthen the asset tracking.

6.4.  Peer Identifiers and Attributes

   The PeerId value in the protocol is a server-allocated identifier for
   its association with the peer and SHOULD NOT be shown to the user
   because its value is initially ephemeral.  Since the PeerId is
   allocated by the server and the scope of the identifier is the single
   server, the so-called identifier squatting attacks, where a malicious
   peer could reserve another peer's identifier, are not possible in
   EAP-NOOB.  The server SHOULD assign a random or pseudorandom PeerId
   to each new peer.  It SHOULD NOT select the PeerId based on any peer
   characteristics that it may know, such as the peer's link-layer
   network address.

   User reset or failure in the OOB Step can cause the peer to perform
   many Initial Exchanges with the server, which allocates many PeerId
   values and stores the ephemeral protocol state for them.  The peer
   will typically only remember the latest ones.  EAP-NOOB leaves it to
   the implementation to decide when to delete these ephemeral
   associations.  There is no security reason to delete them early, and
   the server does not have any way to verify that the peers are
   actually the same one.  Thus, it is safest to store the ephemeral
   states on the server for at least one day.  If the OOB messages are
   sent only in the server-to-peer direction, the server SHOULD NOT
   delete the ephemeral state before all the related Noob values have

   After completion of EAP-NOOB, the server may store the PeerInfo data,
   and the user may use it to identify the peer and its attributes, such
   as the make and model or serial number.  A compromised peer could lie
   in the PeerInfo that it sends to the server.  If the server stores
   any information about the peer, it is important that this information
   is approved by the user during or after the OOB Step.  Without
   verification by the user or authentication on the application level,
   the PeerInfo is not authenticated information and should not be
   relied on.  One possible use for the PeerInfo field is EAP channel
   binding (see Section 6.7).

6.5.  Downgrading Threats

   The fingerprint Hoob protects all the information exchanged in the
   Initial Exchange, including the cryptosuite negotiation.  The message
   authentication codes MACs and MACp also protect the same information.
   The message authentication codes MACs2 and MACp2 protect information
   exchanged during key renegotiation in the Reconnect Exchange.  This
   prevents downgrading attacks to weaker cryptosuites, as long as the
   possible attacks take more time than the maximum time allowed for the
   EAP-NOOB completion.  This is typically the case for recently
   discovered cryptanalytic attacks.

   As an additional precaution, the EAP server and peer MUST check for
   downgrading attacks in the Reconnect Exchange as follows.  As long as
   the server or peer saves any information about the other endpoint, it
   MUST also remember the previously negotiated cryptosuite and MUST NOT
   accept renegotiation of any cryptosuite that is known to be weaker
   than the previous one, such as a deprecated cryptosuite.  Determining
   the relative strength of the cryptosuites is out of scope of this
   specification and may be managed by implementations or by local
   policies at the peer and server.

   Integrity of the direction negotiation cannot be verified in the same
   way as the integrity of the cryptosuite negotiation.  That is, if the
   OOB channel used in an application is critically insecure in one
   direction, an on-path attacker could modify the negotiation messages
   and thereby cause that direction to be used.  Applications that
   support OOB messages in both directions SHOULD, therefore, ensure
   that the OOB channel has sufficiently strong security in both
   directions.  While this is a theoretical vulnerability, it could
   arise in practice if EAP-NOOB is deployed in new applications.
   Currently, we expect most peer devices to support only one OOB
   direction; in which case, interfering with the direction negotiation
   can only prevent the completion of the protocol.

   The long-term shared key material Kz in the persistent EAP-NOOB
   association is established with an ECDHE key exchange when the peer
   and server are first associated.  It is a weaker secret than a
   manually configured random shared key because advances in
   cryptanalysis against the used ECDHE curve could eventually enable
   the attacker to recover Kz.  EAP-NOOB protects against such attacks
   by allowing cryptosuite upgrades in the Reconnect Exchange and by
   updating the shared key material Kz whenever the cryptosuite is
   upgraded.  We do not expect the cryptosuite upgrades to be frequent,
   but, if an upgrade becomes necessary, it can be done without manual
   reset and reassociation of the peer devices.

6.6.  Protected Success and Failure Indications

   Section 7.16 of [RFC3748] allows EAP methods to specify protected
   result indications because EAP-Success and EAP-Failure packets are
   neither acknowledged nor integrity protected.  [RFC3748] notes that
   these indications may be explicit or implicit.

   EAP-NOOB relies on implicit, protected success indicators in the
   Completion Exchange and Reconnect Exchange.  Successful verification
   of MACs and MACs2 in the EAP-Request message from the server (message
   type 6 and message type 9, respectively) acts as an implicit,
   protected success indication to the peer.  Similarly, successful
   verification of MACp and MACp2 in the EAP-Response message from the
   peer (message type 6 and message type 9, respectively) act as an
   implicit, protected success indication to the server.

   EAP-NOOB failure messages are not protected.  Protected failure
   result indications would not significantly improve availability since
   EAP-NOOB reacts to most malformed data by ending the current EAP
   conversation in EAP-Failure.  However, since EAP-NOOB spans multiple
   conversations, failure in one conversation usually means no state
   change on the level of the EAP-NOOB state machine.

6.7.  Channel Binding

   EAP channel binding, defined in [RFC6677], means that the endpoints
   compare their perceptions of network properties, such as lower-layer
   identifiers, over the secure channel established by EAP
   authentication.  Section 4.1 of [RFC6677] defines two approaches to
   channel binding.  EAP-NOOB follows the first approach, in which the
   peer and server exchange plaintext information about the network over
   a channel that is integrity protected with keys derived during the
   EAP execution.  More specifically, channel information is exchanged
   in the plaintext PeerInfo and ServerInfo objects and is later
   verified with message authentication codes (MACp, MACs, MACp2, and
   MACs2).  This allows policy-based comparison with locally perceived
   network properties on either side, as well as logging for debugging
   purposes.  The peer MAY include in PeerInfo any data items that it
   wants to bind to the EAP-NOOB association and to the exported keys.
   These can be properties of the authenticator or the access link, such
   as the SSID and BSSID of the wireless network (see Table 6).  As
   noted in Section 4.3 of [RFC6677], the scope of the channel binding
   varies between deployments.  For example, the server may have less
   link-layer information available from roaming networks than from a
   local enterprise network, and it may be unable to verify all the
   network properties received in PeerInfo.  There are also privacy
   considerations related to exchanging the ServerInfo and PeerInfo
   while roaming (see Section 6.10).

   Channel binding to secure channels, defined in [RFC5056], binds
   authentication at a higher protocol layer to a secure channel at a
   lower layer.  Like most EAP methods, EAP-NOOB exports the session
   keys MSK and EMSK, and an outer tunnel or a higher-layer protocol can
   bind its authentication to these keys.  Additionally, EAP-NOOB
   exports the key AMSK, which may be used to bind application-layer
   authentication to the secure channel created by EAP-NOOB and to the
   session keys MSK and EMSK.

6.8.  Denial of Service

   While denial-of-service (DoS) attacks by on-link attackers cannot be
   fully prevented, the design goal in EAP-NOOB is to void long-lasting
   failure caused by an attacker who is present only temporarily or
   intermittently.  The main defense mechanism is the persistent EAP-
   NOOB association, which is never deleted automatically due to in-band
   messages or error indications.  Thus, the endpoints can always use
   the persistent association for reconnecting after the DoS attacker
   leaves the network.  In this sense, the persistent association serves
   the same function in EAP-NOOB as a permanent primary key or
   certificate in other authentication protocols.  We discuss logical
   attacks against the updates of the persistent association in
   Section 6.9.

   In addition to logical DoS attacks, it is necessary to consider
   resource exhaustion attacks against the EAP server.  The number of
   persistent EAP-NOOB associations created in the server is limited by
   the need for a user to assist in delivering the OOB message.  The
   users can be authenticated for the input or output of the OOB message
   at the EAP server, and any users who create excessive numbers of
   persistent associations can be held accountable and their
   associations can be deleted by the server administrator.  What the
   attacker can do without user authentication is to perform the Initial
   Exchange repeatedly and create a large number of ephemeral
   associations (server in Waiting for OOB (1) state) without ever
   delivering the OOB message.  In Section 6.4, it was suggested that
   the server should store the ephemeral states for at least a day.
   This may require off-loading the state storage from memory to disk
   during a DoS attack.  However, if the server implementation is unable
   to keep up with a rate of Initial Exchanges performed by a DoS
   attacker and needs to drop some ephemeral states, no damage is caused
   to already-created persistent associations, and the honest users can
   resume registering new peers when the DoS attacker leaves the

   There are some trade-offs in the protocol design between politely
   backing off and giving way to DoS attackers.  An on-link DoS attacker
   could spoof the SleepTime value in the Initial Exchange or Waiting
   Exchange to cause denial of service against a specific peer device.
   There is an upper limit on the SleepTime (3600 seconds) to mitigate
   the spoofing threat.  This means that, in the presence of an on-link
   DoS attacker who spoofs the SleepTime, it could take up to one hour
   after the delivery of the OOB message before the device performs the
   Completion Exchange and becomes functional.  Similarly, the Unwanted
   peer error (error code 2001) could cause the peer to stop connecting
   to the network.  If the peer device is able to alert the user about
   the error condition, it can safely stop connecting to the server and
   wait for the user to trigger a reconnection attempt, e.g., by
   resetting the device.  As mentioned in Section 3.6.2, peer devices
   that are unable to alert the user should continue to retry the
   Initial Exchange infrequently to avoid a permanent DoS condition.  We
   believe a maximum backoff time of 3600 seconds is reasonable for a
   new protocol because malfunctioning or misconfigured peer
   implementations are at least as great a concern as DoS attacks, and
   politely backing off within some reasonable limits will increase the
   acceptance of the protocol.  The maximum backoff times could be
   updated to be shorter as the protocol implementations mature.

6.9.  Recovery from Loss of Last Message

   The EAP-NOOB Completion Exchange, as well as the Reconnect Exchange
   with cryptosuite update, results in a persistent state change that
   should take place either on both endpoints or on neither; otherwise,
   the result is a state mismatch that requires user action to resolve.
   The state mismatch can occur if the final EAP response of the
   exchanges is lost.  In the Completion Exchange, the loss of the final
   response (Type=6) results in the peer moving to the Registered (4)
   state and creating a persistent EAP-NOOB association while the server
   stays in an ephemeral state (1 or 2).  In the Reconnect Exchange, the
   loss of the final response (Type=9) results in the peer moving to the
   Registered (4) state and updating its persistent key material Kz
   while the server stays in the Reconnecting (3) state and keeps the
   old key material.

   The state mismatch is an example of an unavoidable problem in
   distributed systems: it is theoretically impossible to guarantee
   synchronous state changes in endpoints that communicate
   asynchronously.  The protocol will always have one critical message
   that may get lost, so that one side commits to the state change and
   the other side does not.  In EAP, the critical message is the final
   response from the peer to the server.  While the final response is
   normally followed by EAP-Success, [RFC3748], Section 4.2 states that
   the peer MAY assume that the EAP-Success was lost and the
   authentication was successful.  Furthermore, EAP method
   implementations in the peer do not receive notification of the EAP-
   Success message from the parent EAP state machine [RFC4137].  For
   these reasons, EAP-NOOB on the peer side commits to a state change
   already when it sends the final response.

   The best available solution to the loss of the critical message is to
   keep trying.  EAP retransmission behavior defined in Section 4.3 of
   [RFC3748] suggests 3-5 retransmissions.  In the absence of an
   attacker, this would be sufficient to reduce the probability of
   failure to an acceptable level.  However, a determined attacker on
   the in-band channel can drop the final EAP-Response message and all
   subsequent retransmissions.  In the Completion Exchange
   (KeyingMode=0) and Reconnect Exchange with cryptosuite upgrade
   (KeyingMode=3), this could result in a state mismatch and persistent
   denial of service until the user resets the peer state.

   EAP-NOOB implements its own recovery mechanism that allows unlimited
   retries of the Reconnect Exchange.  When the DoS attacker eventually
   stops dropping packets on the in-band channel, the protocol will
   recover.  The logic for this recovery mechanism is specified in
   Section 3.4.2.

   EAP-NOOB does not implement the same kind of retry mechanism in the
   Completion Exchange.  The reason is that there is always a user
   involved in the initial association process, and the user can repeat
   the OOB Step to complete the association after the DoS attacker has
   left.  On the other hand, Reconnect Exchange needs to work without
   user involvement.

6.10.  Privacy Considerations

   There are privacy considerations related to performing the Reconnect
   Exchange while roaming.  The peer and server may send updated
   PeerInfo and ServerInfo fields in the Reconnect Exchange.  This data
   is sent unencrypted between the peer and the EAP authenticator, such
   as a wireless access point.  Thus, it can be observed by both
   outsiders and the access network.  The PeerInfo field contains
   identifiers and other information about the peer device (see
   Table 6).  While the information refers to the peer device and not
   directly to the user, it can leak information about the user to the
   access network and to outside observers.  The ServerInfo, on the
   other hand, can leak information about the peer's affiliation with
   the home network.  For this reason, the optional PeerInfo and
   ServerInfo in the Reconnect Exchange SHOULD be omitted unless some
   critical data has changed and it cannot be updated on the application

   Peer devices that randomize their Layer 2 address to prevent tracking
   can do this whenever the user resets the EAP-NOOB association.
   During the lifetime of the association, the PeerId is a unique
   identifier that can be used to track the peer in the access network.
   Later versions of this specification may consider updating the PeerId
   at each Reconnect Exchange.  In that case, it is necessary to
   consider how the authenticator and access-network administrators can
   recognize and add misbehaving peer devices to a deny list, as well as
   how to avoid loss of synchronization between the server and the peer
   if messages are lost during the identifier update.

   To enable stronger identity protection in later versions of EAP-NOOB,
   the optional server-assigned NAI (NewNAI) SHOULD have a constant
   username part.  The RECOMMENDED username is "noob".  The server MAY,
   however, send a different username in NewNAI to avoid username
   collisions within its realm or to conform to a local policy on

6.11.  EAP Security Claims

   EAP security claims are defined in Section 7.2.1 of [RFC3748].  The
   security claims for EAP-NOOB are listed in Table 13.

   | Security        | EAP-NOOB Claim                                  |
   | Property        |                                                 |
   | Authentication  | ECDHE key exchange with out-of-band             |
   | mechanism       | authentication                                  |
   | Protected       | yes                                             |
   | cryptosuite     |                                                 |
   | negotiation     |                                                 |
   | Mutual          | yes                                             |
   | authentication  |                                                 |
   | Integrity       | yes                                             |
   | protection      |                                                 |
   | Replay          | yes                                             |
   | protection      |                                                 |
   | Confidentiality | no                                              |
   | Key derivation  | yes                                             |
   | Key strength    | The specified cryptosuites provide              |
   |                 | key strength of at least 128 bits.              |
   | Dictionary      | yes                                             |
   | attack          |                                                 |
   | protection      |                                                 |
   | Fast reconnect  | yes                                             |
   | Cryptographic   | not applicable                                  |
   | binding         |                                                 |
   | Session         | yes                                             |
   | independence    |                                                 |
   | Fragmentation   | no                                              |
   | Channel binding | yes (The ServerInfo and PeerInfo can            |
   |                 | be used to convey integrity-protected           |
   |                 | channel properties, such as network             |
   |                 | SSID or peer MAC address.)                      |

                       Table 13: EAP Security Claims

7.  References

7.1.  Normative References

   [EUI-48]   IEEE, "IEEE Standard for Local and Metropolitan Area
              Networks: Overview and Architecture",
              DOI 10.1109/IEEESTD.2014.6847097, IEEE Standard 802-2014,
              June 2014, <https://doi.org/10.1109/IEEESTD.2014.6847097>.

              National Institute of Standards and Technology (NIST),
              "Digital Signature Standard (DSS)",
              DOI 10.6028/NIST.FIPS.186-4, FIPS 186-4, July 2013,

   [NIST-DH]  Barker, E., Chen, L., Roginsky, A., Vassilev, A., and R.
              Davis, "Recommendation for Pair-Wise Key-Establishment
              Schemes Using Discrete Logarithm Cryptography",
              DOI 10.6028/NIST.SP.800-56Ar3, NIST Special
              Publication 800-56A Revision 3, April 2018,

   [RFC2104]  Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed-
              Hashing for Message Authentication", RFC 2104,
              DOI 10.17487/RFC2104, February 1997,

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,

   [RFC3748]  Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H.
              Levkowetz, Ed., "Extensible Authentication Protocol
              (EAP)", RFC 3748, DOI 10.17487/RFC3748, June 2004,

   [RFC4648]  Josefsson, S., "The Base16, Base32, and Base64 Data
              Encodings", RFC 4648, DOI 10.17487/RFC4648, October 2006,

   [RFC5247]  Aboba, B., Simon, D., and P. Eronen, "Extensible
              Authentication Protocol (EAP) Key Management Framework",
              RFC 5247, DOI 10.17487/RFC5247, August 2008,

   [RFC6234]  Eastlake 3rd, D. and T. Hansen, "US Secure Hash Algorithms
              (SHA and SHA-based HMAC and HKDF)", RFC 6234,
              DOI 10.17487/RFC6234, May 2011,

   [RFC7517]  Jones, M., "JSON Web Key (JWK)", RFC 7517,
              DOI 10.17487/RFC7517, May 2015,

   [RFC7518]  Jones, M., "JSON Web Algorithms (JWA)", RFC 7518,
              DOI 10.17487/RFC7518, May 2015,

   [RFC7542]  DeKok, A., "The Network Access Identifier", RFC 7542,
              DOI 10.17487/RFC7542, May 2015,

   [RFC7748]  Langley, A., Hamburg, M., and S. Turner, "Elliptic Curves
              for Security", RFC 7748, DOI 10.17487/RFC7748, January
              2016, <https://www.rfc-editor.org/info/rfc7748>.

   [RFC8037]  Liusvaara, I., "CFRG Elliptic Curve Diffie-Hellman (ECDH)
              and Signatures in JSON Object Signing and Encryption
              (JOSE)", RFC 8037, DOI 10.17487/RFC8037, January 2017,

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC8259]  Bray, T., Ed., "The JavaScript Object Notation (JSON) Data
              Interchange Format", STD 90, RFC 8259,
              DOI 10.17487/RFC8259, December 2017,

7.2.  Informative References

              Bluetooth Special Interest Group, "Bluetooth Core
              Specification Version 5.3", July 2021,

              IEEE, "IEEE Standard for Local and Metropolitan Area
              Networks--Port-Based Network Access Control", IEEE
              Standard 802.1X-2020, February 2020.

   [RATS-EAT] Lundblade, L., Mandyam, G., and J. O'Donoghue, "The Entity
              Attestation Token (EAT)", Work in Progress, Internet-
              Draft, draft-ietf-rats-eat-11, 24 October 2021,

   [RFC2904]  Vollbrecht, J., Calhoun, P., Farrell, S., Gommans, L.,
              Gross, G., de Bruijn, B., de Laat, C., Holdrege, M., and
              D. Spence, "AAA Authorization Framework", RFC 2904,
              DOI 10.17487/RFC2904, August 2000,

   [RFC3986]  Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
              Resource Identifier (URI): Generic Syntax", STD 66,
              RFC 3986, DOI 10.17487/RFC3986, January 2005,

   [RFC4137]  Vollbrecht, J., Eronen, P., Petroni, N., and Y. Ohba,
              "State Machines for Extensible Authentication Protocol
              (EAP) Peer and Authenticator", RFC 4137,
              DOI 10.17487/RFC4137, August 2005,

   [RFC5056]  Williams, N., "On the Use of Channel Bindings to Secure
              Channels", RFC 5056, DOI 10.17487/RFC5056, November 2007,

   [RFC5216]  Simon, D., Aboba, B., and R. Hurst, "The EAP-TLS
              Authentication Protocol", RFC 5216, DOI 10.17487/RFC5216,
              March 2008, <https://www.rfc-editor.org/info/rfc5216>.

   [RFC6677]  Hartman, S., Ed., Clancy, T., and K. Hoeper, "Channel-
              Binding Support for Extensible Authentication Protocol
              (EAP) Methods", RFC 6677, DOI 10.17487/RFC6677, July 2012,

   [Sethi14]  Sethi, M., Oat, E., Di Francesco, M., and T. Aura, "Secure
              bootstrapping of cloud-managed ubiquitous displays",
              Proceedings of ACM International Joint Conference on
              Pervasive and Ubiquitous Computing (UbiComp 2014), pp.
              739-750, Seattle, USA, DOI 10.1145/2632048.2632049,
              September 2014,

   [Sethi19]  Sethi, M., Peltonen, A., and T. Aura, "Misbinding Attacks
              on Secure Device Pairing and Bootstrapping",
              DOI 10.1145/3321705.3329813, February 2019,

   [TLS-CWT]  Tschofenig, H. and M. Brossard, "Using CBOR Web Tokens
              (CWTs) in Transport Layer Security (TLS) and Datagram
              Transport Layer Security (DTLS)", Work in Progress,
              Internet-Draft, draft-tschofenig-tls-cwt-02, 13 July 2020,

Appendix A.  Exchanges and Events per State

   Table 14 shows how the EAP server chooses the exchange type depending
   on the server and peer states.  In the state combinations marked with
   hyphen "-", there is no possible exchange and user action is required
   to make progress.  Note that peer state 4 is omitted from the table
   because the peer never connects to the server when the peer is in
   that state.  The table also shows the handling of errors in each
   exchange.  A notable detail is that the recipient of error code 2003
   moves to state 1.

    | Peer States | Exchange Chosen by the Server | Next Peer and    |
    |             |                               | Server States    |
    |                 Server State: Unregistered (0)                 |
    | 0..2        | Initial Exchange              | both 1 (0 on     |
    |             |                               | error)           |
    | 3           | -                             | no change,       |
    |             |                               | notify user      |
    |               Server State: Waiting for OOB (1)                |
    | 0           | Initial Exchange              | both 1 (0 on     |
    |             |                               | error)           |
    | 1           | Waiting Exchange              | both 1 (no       |
    |             |                               | change on error) |
    | 2           | Completion Exchange           | both 4 (A)       |
    | 3           | -                             | no change,       |
    |             |                               | notify user      |
    |                 Server State: OOB Received (2)                 |
    | 0           | Initial Exchange              | both 1 (0 on     |
    |             |                               | error)           |
    | 1           | Completion Exchange           | both 4 (B)       |
    | 2           | Completion Exchange           | both 4 (A)       |
    | 3           | -                             | no change,       |
    |             |                               | notify user      |
    |        Server State: Reconnecting (3) or Registered (4)        |
    | 0..2        | -                             | no change,       |
    |             |                               | notify user      |
    | 3           | Reconnect Exchange            | both 4 (3 on     |
    |             |                               | error)           |

            Table 14: How the Server Chooses the Exchange Type

   (A)  peer to 1 on error 2003; no other changes on error

   (B)  server to 1 on error 2003; no other changes on error

   Table 15 lists the local events that can take place in the server or
   peer.  Both the server and peer output and accept OOB messages in
   association state 1, leading the receiver to state 2.  Communication
   errors and timeouts in states 0..2 lead back to state 0, while
   similar errors in states 3..4 lead to state 3.  An application
   request for rekeying (e.g., to refresh session keys or to upgrade
   cryptosuite) also takes the association from state 3..4 to state 3.
   The user can always reset the association state to 0.  Recovering
   association data, e.g., from a backup, leads to state 3.

         | Server/Peer State | Possible Local Events  | Next    |
         |                   | in the Server and Peer | State   |
         | 1                 | OOB Output             | 1       |
         | 1                 | OOB Input              | 2 (1 on |
         |                   |                        | error)  |
         | 0..2              | Mobility/timeout/      | 0       |
         |                   | network failure        |         |
         | 3..4              | Mobility/timeout/      | 3       |
         |                   | network failure        |         |
         | 3..4              | Rekeying request       | 3       |
         | 0..4              | User resets            | 0       |
         |                   | association            |         |
         | 0..4              | Association state      | 3       |
         |                   | recovery               |         |

              Table 15: Local Events in the Server and Peer

Appendix B.  Application-Specific Parameters

   Table 16 lists OOB channel parameters that need to be specified in
   each application that makes use of EAP-NOOB.  The list is not
   exhaustive and is included for the convenience of implementers only.

     | Parameter          | Description                             |
     | OobDirs            | Allowed directions of the OOB channel.  |
     | OobMessageEncoding | How the OOB message data fields are     |
     |                    | encoded for the OOB channel.            |
     | SleepTimeDefault   | Default minimum time in seconds that    |
     |                    | the peer should sleep before the next   |
     |                    | Waiting Exchange.                       |
     | OobRetries         | Number of received OOB messages with    |
     |                    | invalid Hoob, after which the receiver  |
     |                    | moves to Unregistered (0) state.  When  |
     |                    | the OOB channel has error detection or  |
     |                    | correction, the RECOMMENDED value is 5. |
     | NoobTimeout        | How many seconds the sender of the OOB  |
     |                    | message remembers the sent Noob value.  |
     |                    | The RECOMMENDED value is 3600 seconds.  |
     | ServerInfoType     | The value of the Type field and the     |
     |                    | other required fields in ServerInfo.    |
     | PeerInfoType       | The value of the Type field and the     |
     |                    | other required fields in PeerInfo.      |

                  Table 16: OOB Channel Characteristics

Appendix C.  EAP-NOOB Roaming

   AAA architectures [RFC2904] allow for roaming of network-connected
   appliances that are authenticated over EAP.  While the peer is
   roaming in a visited network, authentication still takes place
   between the peer and an authentication server at its home network.
   EAP-NOOB supports such roaming by allowing the server to assign a NAI
   to the peer.  After the NAI has been assigned, it enables the visited
   network to route the EAP session to the peer's home AAA server.

   A peer device that is new or has gone through a hard reset should be
   connected first to the home network and establish an EAP-NOOB
   association with its home AAA server before it is able to roam.
   After that, it can perform the Reconnect Exchange from the visited

   Alternatively, the device may provide some method for the user to
   configure the NAI of the home network.  This is the user or
   application-configured NAI mentioned in Section 3.3.1.  In that case,
   the EAP-NOOB association can be created while roaming.  The
   configured NAI enables the EAP messages to be routed correctly to the
   home AAA server.

   While roaming, the device needs to identify the networks where the
   EAP-NOOB association can be used to gain network access.  For 802.11
   access networks, the server MAY send a list of SSID strings in the
   ServerInfo field, called either SSIDList or Base64SSIDList.  The list
   is formatted as explained in Table 6.  If present, the peer MAY use
   this list as a hint to determine the networks where the EAP-NOOB
   association can be used for access authorization, in addition to the
   access network where the Initial Exchange took place.

Appendix D.  OOB Message as a URL

   While EAP-NOOB does not mandate any particular OOB communication
   channel, typical OOB channels include graphical displays and emulated
   NFC tags.  In the peer-to-server direction, it may be convenient to
   encode the OOB message as a URL, which is then encoded as a QR code
   for displays and printers or as an NFC Data Exchange Format (NDEF)
   record for dynamic NFC tags.  A user can then simply scan the QR code
   or NFC tag and open the URL, which causes the OOB message to be
   delivered to the authentication server.  The URL MUST specify https
   or another server-authenticated scheme so that there is a secure
   connection to the server and the on-path attacker cannot read or
   modify the OOB message.

   The ServerInfo in this case includes a field called ServerURL of the
   following format with a RECOMMENDED length of at most 60 characters:


   To this, the peer appends the OOB message fields (PeerId, Noob, and
   Hoob) as a query string.  PeerId is provided to the peer by the
   server and might be a 22-character ASCII string.  The peer base64url
   encodes, without padding, the 16-byte values Noob and Hoob into
   22-character ASCII strings.  The query parameters MAY be in any
   order.  The resulting URL is of the following format:


   The following is an example of a well-formed URL encoding the OOB
   message (without line breaks):



   Max Crone, Shiva Prasad TP, and Raghavendra MS implemented parts of
   this protocol with wpa_supplicant and hostapd.  Eduardo Inglés and
   Dan Garcia-Carrillo were involved in the implementation of this
   protocol on Contiki.  Their inputs helped us in improving the

   The authors would like to thank Rhys Smith and Josh Howlett for
   providing valuable feedback, as well as new use cases and
   requirements for the protocol.  Thanks to Eric Rescorla, Alan Dekok,
   Darshak Thakore, Stefan Winter, Hannes Tschofenig, Daniel Migault,
   Roman Danyliw, Benjamin Kaduk, Francesca Palombini, Steve Hanna, Lars
   Eggert, and Éric Vyncke for their comments and reviews.

   We would also like to express our sincere gratitude to Dave Thaler
   for his thorough review of the document.

Authors' Addresses

   Tuomas Aura
   Aalto University
   FI-00076 Aalto

   Email: tuomas.aura@aalto.fi

   Mohit Sethi
   FI-02420 Jorvas

   Email: mohit@iki.fi

   Aleksi Peltonen
   Aalto University
   FI-00076 Aalto

   Email: aleksi.peltonen@aalto.fi