3.1. AT_KDF_INPUT

The format of the AT_KDF_INPUT attribute is shown below.¶

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | AT_KDF_INPUT  | Length        | Actual Network Name Length    |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                                                               |
   .                        Network Name                           .
   .                                                               .
   |                                                               |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

The fields are as follows:¶

AT_KDF_INPUT

This is set to 23.¶

Length

The length of the attribute, calculated as defined in [RFC4187], Section 8.1.¶

Actual Network Name Length

This is a 2-byte actual length field, needed due to the requirement that the previous field is expressed in multiples of 4 bytes per the usual EAP-AKA rules. The Actual Network Name Length field provides the length of the network name in bytes.¶

Network Name

This field contains the network name of the access network for which the authentication is being performed. The name does not include any terminating null characters. Because the length of the entire attribute must be a multiple of 4 bytes, the sender pads the name with 1, 2, or 3 bytes of all zero bits when necessary.¶

Only the server sends the AT_KDF_INPUT attribute. The value is sent as specified in [TS-3GPP.24.302] for both non-3GPP access networks and for 5G access networks. Per [TS-3GPP.33.402], the server always verifies the authorization of a given access network to use a particular name before sending it to the peer over EAP-AKA'. The value of the AT_KDF_INPUT attribute from the server MUST be non-empty, with a greater than zero length in the Actual Network Name Length field. If the AT_KDF_INPUT attribute is empty, the peer behaves as if AUTN had been incorrect and authentication fails. See Section 3 and Figure 3 of [RFC4187] for an overview of how authentication failures are handled.¶

In addition, the peer MAY check the received value against its own understanding of the network name. Upon detecting a discrepancy, the peer either warns the user and continues, or fails the authentication process. More specifically, the peer SHOULD have a configurable policy that it can follow under these circumstances. If the policy indicates that it can continue, the peer SHOULD log a warning message or display it to the user. If the peer chooses to proceed, it MUST use the network name as received in the AT_KDF_INPUT attribute. If the policy indicates that the authentication should fail, the peer behaves as if AUTN had been incorrect and authentication fails.¶

The Network Name field contains a UTF-8 string. This string MUST be constructed as specified in [TS-3GPP.24.302] for "Access Network Identity". The string is structured as fields separated by colons (:). The algorithms and mechanisms to construct the identity string depend on the used access technology.¶

On the network side, the network name construction is a configuration issue in an access network and an authorization check in the authentication server. On the peer, the network name is constructed based on the local observations. For instance, the peer knows which access technology it is using on the link, it can see information in a link-layer beacon, and so on. The construction rules specify how this information maps to an access network name. Typically, the network name consists of the name of the access technology or the name of the access technology followed by some operator identifier that was advertised in a link-layer beacon. In all cases, [TS-3GPP.24.302] is the normative specification for the construction in both the network and peer side. If the peer policy allows running EAP-AKA' over an access technology for which that specification does not provide network name construction rules, the peer SHOULD rely only on the information from the AT_KDF_INPUT attribute and not perform a comparison.¶

If a comparison of the locally determined network name and the one received over EAP-AKA' is performed on the peer, it MUST be done as follows. First, each name is broken down to the fields separated by colons. If one of the names has more colons and fields than the other one, the additional fields are ignored. The remaining sequences of fields are compared, and they match only if they are equal character by character. This algorithm allows a prefix match where the peer would be able to match "", "FOO", and "FOO:BAR" against the value "FOO:BAR" received from the server. This capability is important in order to allow possible updates to the specifications that dictate how the network names are constructed. For instance, if a peer knows that it is running on access technology "FOO", it can use the string "FOO" even if the server uses an additional, more accurate description, e.g., "FOO:BAR", that contains more information.¶

The allocation procedures in [TS-3GPP.24.302] ensure that conflicts potentially arising from using the same name in different types of networks are avoided. The specification also has detailed rules about how a client can determine these based on information available to the client, such as the type of protocol used to attach to the network, beacons sent out by the network, and so on. Information that the client cannot directly observe (such as the type or version of the home network) is not used by this algorithm.¶

The AT_KDF_INPUT attribute MUST be sent and processed as explained above when AT_KDF attribute has the value 1. Future definitions of new AT_KDF values MUST define how this attribute is sent and processed.¶

3.2. AT_KDF

AT_KDF is an attribute that the server uses to reference a specific key derivation function. It offers a negotiation capability that can be useful for future evolution of the key derivation functions.¶

The format of the AT_KDF attribute is shown below.¶

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   | AT_KDF        | Length        |    Key Derivation Function    |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

The fields are as follows:¶

AT_KDF

This is set to 24.¶

Length

The length of the attribute, calculated as defined in [RFC4187], Section 8.1. For AT_KDF, the Length field MUST be set to 1.¶

Key Derivation Function

An enumerated value representing the key derivation function that the server (or peer) wishes to use. Value 1 represents the default key derivation function for EAP-AKA', i.e., employing CK' and IK' as defined in Section 3.3.¶

Servers MUST send one or more AT_KDF attributes in the EAP-Request/AKA'-Challenge message. These attributes represent the desired functions ordered by preference, the most preferred function being the first attribute.¶

Upon receiving a set of these attributes, if the peer supports and is willing to use the key derivation function indicated by the first attribute, the function is taken into use without any further negotiation. However, if the peer does not support this function or is unwilling to use it, it does not process the received EAP-Request/AKA'-Challenge in any way except by responding with the EAP-Response/AKA'-Challenge message that contains only one attribute, AT_KDF with the value set to the selected alternative. If there is no suitable alternative, the peer behaves as if AUTN had been incorrect and authentication fails (see Figure 3 of [RFC4187]). The peer fails the authentication also if there are any duplicate values within the list of AT_KDF attributes (except where the duplication is due to a request to change the key derivation function; see below for further information).¶

Upon receiving an EAP-Response/AKA'-Challenge with AT_KDF from the peer, the server checks that the suggested AT_KDF value was one of the alternatives in its offer. The first AT_KDF value in the message from the server is not a valid alternative since the peer should have accepted it without further negotiation. If the peer has replied with the first AT_KDF value, the server behaves as if AT_MAC of the response had been incorrect and fails the authentication. For an overview of the failed authentication process in the server side, see Section 3 and Figure 2 of [RFC4187]. Otherwise, the server re-sends the EAP-Response/AKA'-Challenge message, but adds the selected alternative to the beginning of the list of AT_KDF attributes and retains the entire list following it. Note that this means that the selected alternative appears twice in the set of AT_KDF values. Responding to the peer's request to change the key derivation function is the only legal situation where such duplication may occur.¶

When the peer receives the new EAP-Request/AKA'-Challenge message, it MUST check that the requested change, and only the requested change, occurred in the list of AT_KDF attributes. If so, it continues with processing the received EAP-Request/AKA'-Challenge as specified in [RFC4187] and Section 3.1 of this document. If not, it behaves as if AT_MAC had been incorrect and fails the authentication. If the peer receives multiple EAP-Request/AKA'-Challenge messages with differing AT_KDF attributes without having requested negotiation, the peer MUST behave as if AT_MAC had been incorrect and fail the authentication.¶

Note that the peer may also request sequence number resynchronization [RFC4187]. This happens after AT_KDF negotiation has already completed. That is, the EAP-Request/AKA'-Challenge and, possibly, the EAP-Response/AKA'-Challenge messages are exchanged first to determine a mutually acceptable key derivation function, and only then is the possible AKA'-Synchronization-Failure message sent. The AKA'-Synchronization-Failure message is sent as a response to the newly received EAP-Request/AKA'-Challenge, which is the last message of the AT_KDF negotiation. Note that if the first proposed KDF is acceptable, then the first EAP-Request/AKA'-Challenge message is also the last message. The AKA'-Synchronization-Failure message MUST contain the AUTS parameter as specified in [RFC4187] and a copy the AT_KDF attributes as they appeared in the last message of the AT_KDF negotiation. If the AT_KDF attributes are found to differ from their earlier values, the peer and server MUST behave as if AT_MAC had been incorrect and fail the authentication.¶

3.3. Key Derivation

Both the peer and server MUST derive the keys as follows.¶

AT_KDF parameter has the value 1

In this case, MK is derived and used as follows:¶

    MK = PRF'(IK'|CK',"EAP-AKA'"|Identity)
    K_encr = MK[0..127]
    K_aut  = MK[128..383]
    K_re   = MK[384..639]
    MSK    = MK[640..1151]
    EMSK   = MK[1152..1663]

¶

Here [n..m] denotes the substring from bit n to m, including bits n and m. PRF' is a new pseudorandom function specified in Section 3.4. The first 1664 bits from its output are used for K_encr (encryption key, 128 bits), K_aut (authentication key, 256 bits), K_re (re-authentication key, 256 bits), MSK (Master Session Key, 512 bits), and EMSK (Extended Master Session Key, 512 bits). These keys are used by the subsequent EAP-AKA' process. K_encr is used by the AT_ENCR_DATA attribute, and K_aut by the AT_MAC attribute. K_re is used later in this section. MSK and EMSK are outputs from a successful EAP method run [RFC3748].¶

IK' and CK' are derived as specified in [TS-3GPP.33.402]. The functions that derive IK' and CK' take the following parameters: CK and IK produced by the AKA algorithm, and value of the Network Name field comes from the AT_KDF_INPUT attribute (without length or padding).¶

The value "EAP-AKA'" is an eight-characters-long ASCII string. It is used as is, without any trailing NUL characters.¶

Identity is the peer identity as specified in Section 7 of [RFC4187] and in Section 5.3.2 of in this document for the 5G cases.¶

When the server creates an AKA challenge and corresponding AUTN, CK, CK', IK, and IK' values, it MUST set the Authentication Management Field (AMF) separation bit to 1 in the AKA algorithm [TS-3GPP.33.102]. Similarly, the peer MUST check that the AMF separation bit is set to 1. If the bit is not set to 1, the peer behaves as if the AUTN had been incorrect and fails the authentication.¶

On fast re-authentication, the following keys are calculated:¶

    MK = PRF'(K_re,"EAP-AKA' re-auth"|Identity|counter|NONCE_S)
    MSK  = MK[0..511]
    EMSK = MK[512..1023]

¶

MSK and EMSK are the resulting 512-bit keys, taking the first 1024 bits from the result of PRF'. Note that K_encr and K_aut are not re-derived on fast re-authentication. K_re is the re-authentication key from the preceding full authentication and stays unchanged over any fast re-authentication(s) that may happen based on it. The value "EAP-AKA' re-auth" is a sixteen-characters-long ASCII string, again represented without any trailing NUL characters. Identity is the fast re-authentication identity, counter is the value from the AT_COUNTER attribute, NONCE_S is the nonce value from the AT_NONCE_S attribute, all as specified in Section 7 of [RFC4187]. To prevent the use of compromised keys in other places, it is forbidden to change the network name when going from the full to the fast re-authentication process. The peer SHOULD NOT attempt fast re-authentication when it knows that the network name in the current access network is different from the one in the initial, full authentication. Upon seeing a re-authentication request with a changed network name, the server SHOULD behave as if the re-authentication identifier had been unrecognized, and fall back to full authentication. The server observes the change in the name by comparing where the fast re-authentication and full authentication EAP transactions were received at the Authentication, Authorization, and Accounting (AAA) protocol level.¶

AT_KDF has any other value

Future variations of key derivation functions may be defined, and they will be represented by new values of AT_KDF. If the peer does not recognize the value, it cannot calculate the keys and behaves as explained in Section 3.2.¶

AT_KDF is missing

The peer behaves as if the AUTN had been incorrect and MUST fail the authentication.¶

If the peer supports a given key derivation function but is unwilling to perform it for policy reasons, it refuses to calculate the keys and behaves as explained in Section 3.2.¶

3.4. Hash Functions

EAP-AKA' uses SHA-256 / HMAC-SHA-256, not SHA-1 / HMAC-SHA-1 (see [FIPS.180-4] and [RFC2104]) as in EAP-AKA. This requires a change to the pseudorandom function (PRF) as well as the AT_MAC and AT_CHECKCODE attributes.¶

PRF'(K,S) = T1 | T2 | T3 | T4 | ...

   where:
   T1 = HMAC-SHA-256 (K, S | 0x01)
   T2 = HMAC-SHA-256 (K, T1 | S | 0x02)
   T3 = HMAC-SHA-256 (K, T2 | S | 0x03)
   T4 = HMAC-SHA-256 (K, T3 | S | 0x04)
   ...

 0                   1                   2                   3
 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| AT_CHECKCODE  | Length        |           Reserved            |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|                                                               |
|                     Checkcode (0 or 32 bytes)                 |
|                                                               |
|                                                               |
|                                                               |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

3.5. Summary of Attributes for EAP-AKA'

Table 1 identifies which attributes may be found in which kinds of messages, and in what quantity.¶

Messages are denoted with numbers as follows:¶

1

EAP-Request/AKA-Identity¶

2

EAP-Response/AKA-Identity¶

3

EAP-Request/AKA-Challenge¶

4

EAP-Response/AKA-Challenge¶

5

EAP-Request/AKA-Notification¶

6

EAP-Response/AKA-Notification¶

7

EAP-Response/AKA-Client-Error¶

8

EAP-Request/AKA-Reauthentication¶

9

EAP-Response/AKA-Reauthentication¶

10

EAP-Response/AKA-Authentication-Reject¶

11

EAP-Response/AKA-Synchronization-Failure¶

The column denoted with "E" indicates whether the attribute is a nested attribute that MUST be included within AT_ENCR_DATA.¶

In addition, the numbered columns indicate the quantity of the attribute within the message as follows:¶

"0"

Indicates that the attribute MUST NOT be included in the message.¶

"1"

Indicates that the attribute MUST be included in the message.¶

"0-1"

Indicates that the attribute is sometimes included in the message¶

"0+"

Indicates that zero or more copies of the attribute MAY be included in the message.¶

"1+"

Indicates that there MUST be at least one attribute in the message but more than one MAY be included in the message.¶

"0*"

Indicates that the attribute is not included in the message in cases specified in this document, but MAY be included in the future versions of the protocol.¶

The attribute table is shown below. The table is largely the same as in the EAP-AKA attribute table ([RFC4187], Section 10.1), but changes how many times AT_MAC may appear in an EAP-Response/AKA'-Challenge message as it does not appear there when AT_KDF has to be sent from the peer to the server. The table also adds the AT_KDF and AT_KDF_INPUT attributes.¶

4.1. Summary of Attributes for EAP-AKA

The appearance of the AT_BIDDING attribute in EAP-AKA exchanges is shown below, using the notation from Section 3.5:¶

5.1. Username Types in EAP-AKA' Identities

Section 4.1.1.3 of [RFC4187] specifies that there are three types of usernames: permanent, pseudonym, and fast re-authentication usernames. This specification extends this definition as follows. There are four types of usernames:¶

(1)

Regular usernames. These are external names given to EAP-AKA' peers. The regular usernames are further subdivided into to categories:¶

(a)

Permanent usernames, for instance, IMSI-based usernames.¶

(b)

Privacy-friendly temporary usernames, for instance, 5G GUTI (5G Globally Unique Temporary Identifier) or 5G privacy identifiers (see Section 5.3.2) such as SUCI (Subscription Concealed Identifier).¶

(2)

EAP-AKA' pseudonym usernames. For example, 2s7ah6n9q@example.com might be a valid pseudonym identity. In this example, 2s7ah6n9q is the pseudonym username.¶

(3)

EAP-AKA' fast re-authentication usernames. For example, 43953754@example.com might be a valid fast re-authentication identity and 43953754 the fast re-authentication username.¶

The permanent, privacy-friendly temporary, and pseudonym usernames are only used with full authentication, and fast re-authentication usernames only with fast re-authentication. Unlike permanent usernames and pseudonym usernames, privacy-friendly temporary usernames and fast re-authentication usernames are one-time identifiers, which are not reused across EAP exchanges.¶

5.2. Generating Pseudonyms and Fast Re-Authentication Identities

This section provides some additional guidance to implementations for producing secure pseudonyms and fast re-authentication identities. It does not impact backwards compatibility because each server consumes only the identities that it generates itself. However, adherence to the guidance will provide better security.¶

As specified by [RFC4187], Section 4.1.1.7, pseudonym usernames and fast re-authentication identities are generated by the EAP server in an implementation-dependent manner. RFC 4187 provides some general requirements on how these identities are transported, how they map to the NAI syntax, how they are distinguished from each other, and so on.¶

However, to enhance privacy, some additional requirements need to be applied.¶

The pseudonym usernames and fast re-authentication identities MUST be generated in a cryptographically secure way so that it is computationally infeasible for an attacker to differentiate two identities belonging to the same user from two identities belonging to different users. This can be achieved, for instance, by using random or pseudorandom identifiers such as random byte strings or ciphertexts. See also [RFC4086] for guidance on random number generation.¶

Note that the pseudonym and fast re-authentication usernames also MUST NOT include substrings that can be used to relate the username to a particular entity or a particular permanent identity. For instance, the usernames cannot include any subscriber-identifying part of an IMSI or other permanent identifier. Similarly, no part of the username can be formed by a fixed mapping that stays the same across multiple different pseudonyms or fast re-authentication identities for the same subscriber.¶

When the identifier used to identify a subscriber in an EAP-AKA' authentication exchange is a privacy-friendly identifier that is used only once, the EAP-AKA' peer MUST NOT use a pseudonym provided in that authentication exchange in subsequent exchanges more than once. To ensure that this does not happen, the EAP-AKA' server MAY decline to provide a pseudonym in such authentication exchanges. An important case where such privacy-friendly identifiers are used is in 5G networks (see Section 5.3).¶

5.3. Identifier Usage in 5G

In EAP-AKA', the peer identity may be communicated to the server in one of three ways:¶

• As a part of link-layer establishment procedures, externally to EAP.¶
• With the EAP-Response/Identity message in the beginning of the EAP exchange, but before the selection of EAP-AKA'.¶
• Transmitted from the peer to the server using EAP-AKA' messages instead of EAP-Response/Identity. In this case, the server includes an identity-requesting attribute (AT_ANY_ID_REQ, AT_FULLAUTH_ID_REQ, or AT_PERMANENT_ID_REQ) in the EAP-Request/AKA-Identity message, and the peer includes the AT_IDENTITY attribute, which contains the peer's identity, in the EAP-Response/AKA-Identity message.¶

The identity carried above may be a permanent identity, privacy-friendly identity, pseudonym identity, or fast re-authentication identity as defined in Section 5.1.¶

5G supports the concept of privacy identifiers, and it is important for interoperability that the right type of identifier is used.¶

5G defines the SUbscription Permanent Identifier (SUPI) and SUbscription Concealed Identifier (SUCI) [TS-3GPP.23.501] [TS-3GPP.33.501] [TS-3GPP.23.003]. SUPI is globally unique and allocated to each subscriber. However, it is only used internally in the 5G network and is privacy sensitive. The SUCI is a privacy-preserving identifier containing the concealed SUPI, using public key cryptography to encrypt the SUPI.¶

Given the choice between these two types of identifiers, EAP-AKA' ensures interoperability as follows:¶

• Where identifiers are used within EAP-AKA' (such as key derivation) determine the exact values of the identity to be used, to avoid ambiguity (see Section 5.3.1).¶
• Where identifiers are carried within EAP-AKA' packets (such as in the AT_IDENTITY attribute) determine which identifiers should be filled in (see Section 5.3.2).¶

In 5G, the normal mode of operation is that identifiers are only transmitted outside EAP. However, in a system involving terminals from many generations and several connectivity options via 5G and other mechanisms, implementations and the EAP-AKA' specification need to prepare for many different situations, including sometimes having to communicate identities within EAP.¶

The following sections clarify which identifiers are used and how.¶

7.1. Privacy

[RFC6973] suggests that the privacy considerations of IETF protocols be documented.¶

The confidentiality properties of EAP-AKA' itself have been discussed above under "Confidentiality" (Section 7).¶

EAP-AKA' uses several different types of identifiers to identify the authenticating peer. It is strongly RECOMMENDED to use the privacy-friendly temporary or hidden identifiers, i.e., the 5G GUTI or SUCI, pseudonym usernames, and fast re-authentication usernames. The use of permanent identifiers such as the IMSI or SUPI may lead to an ability to track the peer and/or user associated with the peer. The use of permanent identifiers such as the IMSI or SUPI is strongly NOT RECOMMENDED.¶

As discussed in Section 5.3, when authenticating to a 5G network, only the SUCI identifier is normally used. The use of EAP-AKA' pseudonyms in this situation is at best limited because the SUCI already provides a stronger mechanism. In fact, reusing the same pseudonym multiple times will result in a tracking opportunity for observers that see the pseudonym pass by. To avoid this, the peer and server need to follow the guidelines given in Section 5.2.¶

When authenticating to a 5G network, per Section 5.3.1, both the EAP-AKA' peer and server need to employ the permanent identifier SUPI as an input to key derivation. However, this use of the SUPI is only internal. As such, the SUPI need not be communicated in EAP messages. Therefore, SUPI MUST NOT be communicated in EAP-AKA' when authenticating to a 5G network.¶

While the use of SUCI in 5G networks generally provides identity privacy, this is not true if the null-scheme encryption is used to construct the SUCI (see [TS-3GPP.33.501], Annex C). The use of this scheme makes the use of SUCI equivalent to the use of SUPI or IMSI. The use of the null scheme is NOT RECOMMENDED where identity privacy is important.¶

The use of fast re-authentication identities when authenticating to a 5G network does not have the same problems as the use of pseudonyms, as long as the 5G authentication server generates the fast re-authentication identifiers in a proper manner specified in Section 5.2.¶

Outside 5G, the peer can freely choose between the use of permanent, pseudonym, or fast re-authentication identifiers:¶

• A peer that has not yet performed any EAP-AKA' exchanges does not typically have a pseudonym available. If the peer does not have a pseudonym available, then the privacy mechanism cannot be used, and the permanent identity will have to be sent in the clear.¶

  The terminal SHOULD store the pseudonym in nonvolatile memory so that it can be maintained across reboots. An active attacker that impersonates the network may use the AT_PERMANENT_ID_REQ attribute ([RFC4187], Section 4.1.2) to learn the subscriber's IMSI. However, as discussed in [RFC4187], Section 4.1.2, the terminal can refuse to send the cleartext permanent identity if it believes that the network should be able to recognize the pseudonym.¶

• When pseudonyms and fast re-authentication identities are used, the peer relies on the properly created identifiers by the server.¶

  It is essential that an attacker cannot link a privacy-friendly identifier to the user in any way or determine that two identifiers belong to the same user as outlined in Section 5.2. The pseudonym usernames and fast re-authentication identities MUST NOT be used for other purposes (e.g., in other protocols).¶

If the peer and server cannot guarantee that SUCI can be used or that pseudonyms will be available, generated properly, and maintained reliably, and identity privacy is required, then additional protection from an external security mechanism such as tunneled EAP methods like Tunneled Transport Layer Security (TTLS) [RFC5281] or Tunnel Extensible Authentication Protocol (TEAP) [RFC7170] may be used. The benefits and the security considerations of using an external security mechanism with EAP-AKA are beyond the scope of this document.¶

Finally, as with other EAP methods, even when privacy-friendly identifiers or EAP tunneling is used, typically the domain part of an identifier (e.g., the home operator) is visible to external parties.¶

7.2. Discovered Vulnerabilities

There have been no published attacks that violate the primary secrecy or authentication properties defined for Authentication and Key Agreement (AKA) under the originally assumed trust model. The same is true of EAP-AKA'.¶

However, there have been attacks when a different trust model is in use, with characteristics not originally provided by the design, or when participants in the protocol leak information to outsiders on purpose, and there have been some privacy-related attacks.¶

For instance, the original AKA protocol does not prevent an insider supplying keys to a third party, e.g., as described by Mjølsnes and Tsay in [MT2012] where a serving network lets an authentication run succeed, but then it misuses the session keys to send traffic on the authenticated user's behalf. This particular attack is not different from any on-path entity (such as a router) pretending to send traffic, but the general issue of insider attacks can be a problem, particularly in a large group of collaborating operators.¶

Another class of attacks is the use of tunneling of traffic from one place to another, e.g., as done by Zhang and Fang in [ZF2005] to leverage security policy differences between different operator networks, for instance. To gain something in such an attack, the attacker needs to trick the user into believing it is in another location. If policies between locations differ, for instance, if payload traffic is not required to be encrypted in some location, the attacker may trick the user into opening a vulnerability. As an authentication mechanism, EAP-AKA' is not directly affected by most of these attacks. EAP-AKA' network name binding can also help alleviate some of the attacks. In any case, it is recommended that EAP-AKA' configuration not be dependent on the location of request origin, unless the location information can be cryptographically confirmed, e.g., with the network name binding.¶

Zhang and Fang also looked at denial-of-service attacks [ZF2005]. A serving network may request large numbers of authentication runs for a particular subscriber from a home network. While the resynchronization process can help recover from this, eventually it is possible to exhaust the sequence number space and render the subscriber's card unusable. This attack is possible for both original AKA and EAP-AKA'. However, it requires the collaboration of a serving network in an attack. It is recommended that EAP-AKA' implementations provide the means to track, detect, and limit excessive authentication attempts to combat this problem.¶

There have also been attacks related to the use of AKA without the generated session keys (e.g., [BT2013]). Some of those attacks relate to the use of HTTP Digest AKAv1 [RFC3310], which was originally vulnerable to man-in-the-middle attacks. This has since been corrected in [RFC4169]. The EAP-AKA' protocol uses session keys and provides channel binding, and as such, it is resistant to the above attacks except where the protocol participants leak information to outsiders.¶

Basin, et al. [Basin2018] have performed formal analysis and concluded that the AKA protocol would have benefited from additional security requirements such as key confirmation.¶

In the context of pervasive monitoring revelations, there were also reports of compromised long-term pre-shared keys used in SIM and AKA [Heist2015]. While no protocol can survive the theft of key material associated with its credentials, there are some things that alleviate the impacts in such situations. These are discussed further in Section 7.3.¶

Arapinis, et al. [Arapinis2012] describe an attack that uses the AKA resynchronization protocol to attempt to detect whether a particular subscriber is in a given area. This attack depends on the attacker setting up a false base station in the given area and on the subscriber performing at least one authentication between the time the attack is set up and run.¶

Borgaonkar, et al. discovered that the AKA resynchronization protocol may also be used to predict the authentication frequency of a subscriber if a non-time-based sequence number (SQN) generation scheme is used [Borgaonkar2018]. The attacker can force the reuse of the keystream that is used to protect the SQN in the AKA resynchronization protocol. The attacker then guesses the authentication frequency based on the lowest bits of two XORed SQNs. The researchers' concern was that the authentication frequency would reveal some information about the phone usage behavior, e.g., number of phone calls made or number of SMS messages sent. There are a number of possible triggers for authentication, so such an information leak is not direct, but it can be a concern. The impact of the attack differs depending on whether the SQN generation scheme that is used is time-based or not.¶

Similar attacks are possible outside AKA in the cellular paging protocols where the attacker can simply send application-layer data, send short messages, or make phone calls to the intended victim and observe the air interface (e.g., [Kune2012] and [Shaik2016]). Hussain, et al. demonstrated a slightly more sophisticated version of the attack that exploits the fact that the 4G paging protocol uses the IMSI to calculate the paging timeslot [Hussain2019]. As this attack is outside AKA, it does not impact EAP-AKA'.¶

Finally, bad implementations of EAP-AKA' may not produce pseudonym usernames or fast re-authentication identities in a manner that is sufficiently secure. While it is not a problem with the protocol itself, following the recommendations in Section 5.2 can mitigate this concern.¶

7.3. Pervasive Monitoring

As required by [RFC7258], work on IETF protocols needs to consider the effects of pervasive monitoring and mitigate them when possible.¶

As described in Section 7.2, after the publication of RFC 5448, new information has come to light regarding the use of pervasive monitoring techniques against many security technologies, including AKA-based authentication.¶

For AKA, these attacks relate to theft of the long-term, shared-secret key material stored on the cards. Such attacks are conceivable, for instance, during the manufacturing process of cards, through coercion of the card manufacturers, or during the transfer of cards and associated information to an operator. Since the publication of reports about such attacks, manufacturing and provisioning processes have gained much scrutiny and have improved.¶

In particular, it is crucial that manufacturers limit access to the secret information and the cards only to necessary systems and personnel. It is also crucial that secure mechanisms be used to store and communicate the secrets between the manufacturer and the operator that adopts those cards for their customers.¶

Beyond these operational considerations, there are also technical means to improve resistance to these attacks. One approach is to provide Perfect Forward Secrecy (PFS). This would prevent any passive attacks merely based on the long-term secrets and observation of traffic. Such a mechanism can be defined as a backwards-compatible extension of EAP-AKA' and is pursued separately from this specification [EMU-AKA-PFS]. Alternatively, EAP-AKA' authentication can be run inside a PFS-capable, tunneled authentication method. In any case, the use of some PFS-capable mechanism is recommended.¶

7.4. Security Properties of Binding Network Names

The ability of EAP-AKA' to bind the network name into the used keys provides some additional protection against key leakage to inappropriate parties. The keys used in the protocol are specific to a particular network name. If key leakage occurs due to an accident, access node compromise, or another attack, the leaked keys are only useful when providing access with that name. For instance, a malicious access point cannot claim to be network Y if it has stolen keys from network X. Obviously, if an access point is compromised, the malicious node can still represent the compromised node. As a result, neither EAP-AKA' nor any other extension can prevent such attacks; however, the binding to a particular name limits the attacker's choices, allows better tracking of attacks, makes it possible to identify compromised networks, and applies good cryptographic hygiene.¶

The server receives the EAP transaction from a given access network, and verifies that the claim from the access network corresponds to the name that this access network should be using. It becomes impossible for an access network to claim over AAA that it is another access network. In addition, if the peer checks that the information it has received locally over the network-access link-layer matches with the information the server has given it via EAP-AKA', it becomes impossible for the access network to tell one story to the AAA network and another one to the peer. These checks prevent some "lying NAS" (Network Access Server) attacks. For instance, a roaming partner, R, might claim that it is the home network H in an effort to lure peers to connect to itself. Such an attack would be beneficial for the roaming partner if it can attract more users, and damaging for the users if their access costs in R are higher than those in other alternative networks, such as H.¶

Any attacker who gets hold of the keys CK and IK, produced by the AKA algorithm, can compute the keys CK' and IK' and, hence, the Master Key (MK) according to the rules in Section 3.3. The attacker could then act as a lying NAS. In 3GPP systems in general, the keys CK and IK have been distributed to, for instance, nodes in a visited access network where they may be vulnerable. In order to reduce this risk, the AKA algorithm MUST be computed with the AMF separation bit set to 1, and the peer MUST check that this is indeed the case whenever it runs EAP-AKA'. Furthermore, [TS-3GPP.33.402] requires that no CK or IK keys computed in this way ever leave the home subscriber system.¶

The additional security benefits obtained from the binding depend obviously on the way names are assigned to different access networks. This is specified in [TS-3GPP.24.302]. See also [TS-3GPP.23.003]. Ideally, the names allow separating each different access technology, each different access network, and each different NAS within a domain. If this is not possible, the full benefits may not be achieved. For instance, if the names identify just an access technology, use of compromised keys in a different technology can be prevented, but it is not possible to prevent their use by other domains or devices using the same technology.¶

8.1. Type Value

IANA has updated the reference for EAP-AKA' (0x32) in the "Method Types" subregistry under the "Extensible Authentication Protocol (EAP) Registry" to point to this document. Per Section 6.2 of [RFC3748], this allocation can be made with Specification Required [RFC8126].¶