Protocol for Carrying Authentication for Network Access (PANA)

Providing secure network access service requires access control based on the authentication and authorization of the clients and the access networks. Client-to-network authentication provides parameters that are needed to police the traffic flow through the enforcement points. A protocol is needed to carry authentication methods between the client and the access network. Currently there is no standard network-layer solution for authenticating clients for network access. Appendix A of describes the problem statement that led to the development of PANA. Scope of this work is identified as designing a link-layer agnostic transport for network access authentication methods. The Extensible Authentication Protocol (EAP) provides such authentication methods. In other words, PANA will carry EAP which can carry various authentication methods. By the virtue of enabling transport of EAP above IP, any authentication method that can be carried as an EAP method is made available to PANA and hence to any link-layer technology. There is a clear division of labor between PANA (an EAP lower layer), EAP and EAP methods as described in . Various environments and usage models for PANA are identified in Appendix A of . Potential security threats for network-layer access authentication protocol are discussed in . These have been essential in defining the requirements on the PANA protocol. Note that some of these requirements are imposed by the chosen payload, EAP . There are components that are part of a complete secure network solution but are outside of the PANA protocol specification, including IP address configuration, authentication method choice, filter rule installation, data traffic protection and PAA-EP protocol. These components are described in separate documents (see and ). The readers are recommended to go through the PANA Framework document prior to reading this protocol specification document.

In this document, several words are used to signify the requirements of the specification. These words are often capitalized. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in .

The client side of the protocol that resides in the access device (e.g., laptop, PDA, etc.). It is responsible for providing the credentials in order to prove its identity for network access authorization. The protocol entity in the access network whose responsibility is to verify the credentials provided by a PANA client (PaC) and authorize network access to the device associated with the client and identified by a Device Identifier (DI). Note the authentication and authorization procedure can, according to the EAP model, be also offloaded to the backend AAA infrastructure. A PANA session begins with the handshake between the PANA Client (PaC) and the PANA Authentication Agent (PAA), and terminates as a result of an authentication failure, a timeout, or an explicit termination message. A fixed session identifier is maintained throughout a session. A session cannot be shared across multiple network interfaces. This identifier is used to uniquely identify a PANA session on the PAA and PaC. It includes an identifier of the PAA, therefore it cannot be shared across multiple PAAs. It is included in PANA messages to bind the message to a specific PANA session. This bidirectional identifier is allocated by the PAA following the handshake and freed when the session terminates. A PANA security association is a relationship between the PaC and PAA, formed by the sharing of cryptographic keying material and associated context. Security associations are duplex. That is, one security association is needed to protect the bidirectional traffic between the PaC and the PAA. The identifier used by the network as a handle to control and police the network access of a client. Depending on the access technology, this identifier may contain an address that is carried in protocol headers (e.g., IP or link-layer address), or a locally significant identifier that is made available by the local protocol stack (e.g., circuit id, PPP interface id) of a connected device. A node on the access network where per-packet enforcement policies (i.e., filters) are applied on the inbound and outbound traffic of client devices. Information such as the DI and (optionally) cryptographic keys are provided by the PAA per client for generating filters on the EP. EP and PAA may be co-located. A service provider that provides physical and link-layer connectivity to an access network it manages. A key derived by the EAP peer and EAP server and transported to the authenticator .

The PANA protocol is run between a client (PaC) and a server (PAA) in order to perform authentication and authorization for the network access service. The protocol messaging consists of a series of request and responses, some of which may be initiated by either ends. Each message can carry zero or more AVPs as payload. The main payload of PANA is EAP which performs authentication. PANA helps PaC and PAA establish an EAP session. PANA is a UDP-based protocol. It has its own retransmission mechanism to reliably deliver messages. PANA messages are sent between a PaC and PAA as part of a PANA session. A PANA session consists of distinct phases: Discovery and handshake phase: This is the phase that initiates a new PANA session. The PaC discovers the PAA(s) by either explicitly soliciting advertisements for them or receiving unsolicited advertisements. The PaC's answer sent in response to an advertisement starts a new session. Authentication phase: Immediately following the handshake phase is the EAP execution between the PAA and PaC. The EAP payloads (which carry an EAP method inside) is what is used for authentication. Authentication phase may involve execution of two EAP sessions back-to-back, one for the NAP and one for the ISP. Authorized phase: Following a successful PANA authentication phase, the PaC gains access to the network and can send and receive IP data traffic through EP. During this phase, the PaC and PAA may optionally ping each other to test liveness of the PANA session on each end. Re-authentication phase: Following an authorized phase, the PAA must initiate re-authentication before the PANA session lifetime expires. Again EAP is carried by PANA to perform authentication. This phase may be optionally triggered by both the PaC and the PAA without any respect to the session lifetime. The session moves to this phase from authorized phase, and returns back there upon successful re-authentication. Termination phase: The PaC or PAA may choose to discontinue the access service at any time. An explicit disconnect message can be sent by either end. If either the PaC or the PAA disconnects without engaging in termination messaging, it is expected that either the expiration of a finite session lifetime or failed liveness tests would do the job.

PaC PAA Message ----------------------------------------------------- // Discovery and handshake phase -----> PANA-PAA-Discover <----- PANA-Start-Request -----> PANA-Start-Answer // Authentication phase <----- PANA-Auth-Request /* EAP Request */ -----> PANA-Auth-Answer -----> PANA-Auth-Request /* EAP Response */ <----- PANA-Auth-Answer <----- PANA-Bind-Request /* EAP Success */ -----> PANA-Bind-Answer // Authorized phase (IP data traffic allowed) <----- PANA-Ping-Request -----> PANA-Ping-Answer // Termination phase -----> PANA-Termination-Request <----- PANA-Termination-Answer Note that depending on the environment and deployment the protocol flow depicted in can be abbreviated. Cryptographic protection of messages between the PaC and PAA is possible as soon as EAP in conjunction with the EAP method exports a shared key. That shared key is used to create a PANA SA. The PANA SA helps generating per-message authentication codes that provide integrity protection and authentication. PANA also allows creation of a new PANA session with a new PAA out of an existing session with another PAA. This optimization allows PaC achieve quicker authorization without having to run EAP upon movement (changing PAAs). Throughout the lifetime of a session, various problems found with the incoming messages can generate a PANA error message sent in response.

The following sections explain in detail the various phases of a PANA session.

The payload of any PANA message consists of zero or more AVPs (Attribute Value Pairs). The subsequent sections refer to these AVPs, therefore the list of AVPs are provided with a brief description before more extensive descriptions are included later in the document. Cookie AVP: contains a random value that is used for making PAA discovery robust against blind resource consumption DoS attacks. Protection-Capability AVP: contains the type of per-packet protection (link-layer vs. network-layer) when a cryptographic mechanism should be enabled after PANA authentication. Device-Id AVP: contains a device identifier (link-layer address or an IP address) of the PaC or an EP. EAP AVP: contains an EAP PDU. MAC AVP: contains a Message Authentication Code that integrity protects the PANA message. Termination-Cause AVP: contains the reason of session termination. Result-Code AVP: contains information about the protocol execution results. Session-Id AVP: contains the PANA session identifier value. Session-Lifetime AVP: contains the duration of authorized access. Failed-AVP: contains an offending AVP that caused a failure. NAP-Information AVP, ISP-Information AVP: contains the identifier of a NAP and an ISP, respectively. Key-Id AVP: contains a AAA-Key identifier. PPAC AVP: Post-PANA-Address-Configuration AVP. Used to indicate the available/chosen IP address configuration methods that can be used by the PaC after successful PANA authentication. Nonce AVP: contains a randomly chosen value that is used in cyrptographic key computations. IP-Address AVP: contains an IP Address of the PaC.

When a PaC attaches to a network, and knows that it has to discover a PAA, it SHOULD send a PANA-PAA-Discover message to a well-known link local multicast address (TBD) and UDP port (TBD). The PAA discovery assumes that the PaC and the PAA are one hop away from each other. If the PaC knows the IP address of the PAA (based on pre-configuration), it MAY unicast the PANA-PAA-Discover message to that address. When the PAA receives a PANA-PAA-Discover message from a PaC, the PAA SHOULD unicast a PANA-Start-Request message to the PaC. The PaC MAY also choose to start sending data packets before getting authenticated. The EP in an access network that implements PANA SHOULD drop unauthorized packets upon receipt. Additionally, the EP MAY also take this traffic as an indication of unauthorized PaC and notify the PAA. The EP-to-PAA notification SHOULD be sent via . In response, the PAA SHOULD send an unsolicited PANA-Start-Request message to the PaC. This is called "traffic-driven PAA discovery" (an alternative to PaC explicitly soliciting for PAA). Note that this optional feature MAY NOT be present in all deployments, therefore PaCs MUST NOT assume its availability. When a PaC receives a PANA-Start-Request message from a PAA, it responds with a PANA-Start-Answer message if it wishes to enter an authentication phase. There can be multiple PAAs on the link and a PaC may receive multiple PANA-Start-Request messages from those PAAs. The authentication and authorization result does not depend on which PAA is chosen by the PaC. By default the PaC MAY choose the PAA that sent the first response. A PANA-Start-Request message MAY carry a Cookie AVP that contains a cookie. The cookie is used for preventing the PAA from resource consumption DoS attacks by blind attackers which bombard the PAA with PANA-PAA-Discover messages. By relying on a cookie mechanism the PAA can avoid per-PaC state creation until after the PaC can produce the same cookie in its PANA-Start-Answer message. In order to do that, the cookie MUST be computed in such a way that it does not require any per-session state maintenance on the PAA in order to verify the cookie returned in a PANA-Start-Answer message. The PAA discovery that takes advantage of cookies is called "stateless PAA discovery". The exact algorithms and syntax used for generating cookies does not affect interoperability and hence is not specified here. An example algorithm is described below. Cookie = <secret-version> | HMAC_SHA1( <Device-Id of PaC> , <secret> ) where <secret> is a randomly generated secret known only to the PAA, <secret-version> is an index used for choosing the secret for generating the cookie and '|' indicates concatenation. The secret-version should be changed frequently enough to prevent replay attacks. The secret key is valid for a certain time frame. When the PaC sends a PANA-Start-Answer message in response to a PANA-Start-Request containing a Cookie AVP, the answer MUST contain a Cookie AVP with the cookie value copied from the request. When the PAA receives the PANA-Start-Answer message from the PaC, it verifies the cookie. The cookie is considered as valid if the received cookie has the expected value. If the computed cookie is valid, the protocol enters the authentication phase. Otherwise, it MUST silently discard the received message. Initial EAP Request MAY be optionally carried by the PANA-Start-Request (as opposed to by a later PANA-Auth-Request) message in order to reduce the number of round-trips. This optimization SHOULD NOT be used if the PAA discovery is desired to be stateless. A Protection-Capability AVP and a Post-PANA-Address-Configuration (PPAC) AVP MAY be included in the PANA-Start-Request in order to indicate required and available capabilities for the network access. These AVPs MAY be used by the PaC for assessing the capability match even before the authentication takes place. Since these AVPs are provided during the insecure discovery and handshake phase, there are certain security risks involved in using the provided information. See for further discussion on this. If the initial EAP Request message is carried in the PANA-Start-Request message, an EAP Response message MUST be carried in the PANA-Start-Answer message returned to the PAA. The PANA-Start-Request/Answer exchange is needed before entering an authentication phase even when the PaC is pre-configured with PAAs IP address and the PANA-PAA-Discover message is unicast. A Nonce AVP MUST be included in PANA-Start-Request and PANA-Start-Answer messages. The nonces are used to establish a PANA SA. A PANA-Start-Request message of a stateless PAA discovery MUST NOT be retransmitted as this voids the statelessness on the PAA. Instead, the PaC MUST retransmit the PANA-PAA-Discover until it receives a PANA-Start-Request message, and retransmit the PANA-Start-Answer message until it receives a PANA-Auth-Request message. The PaC can determine whether the PAA is using stateless discovery by the presence of Cookie AVP. The PANA-Start-Request message MUST be retransmitted instead of the PANA-Start-Answer message when stateful PAA discovery is used. It is possible that both the PAA and the PaC initiate the discovery and handshake procedure at the same time, i.e., the PAA sends a PANA-Start-Request message while the PaC sends a PANA-PAA-Discover message. To resolve the race condition, the PAA SHOULD silently discard the PANA-PAA-Discover message received from the PaC after it has sent a PANA-Start-Request message with creating a state (i.e., no Cookie AVP is included in the message) for the PaC. In this case the PAA will retransmit PANA-Start-Request based on a timer, if the PaC doesn't respond in time (message was lost for example). If the PAA had sent a PANA-Start-Request message without creating a state for the PaC (i.e., a Cookie AVP was included in the message), then it SHOULD answer to the PANA-PAA-Discover message. shows an example sequence for the discovery and handshake phase when a PANA-PAA-Discover message is sent by the PaC. shows an example sequence for the discovery and handshake phase that is triggered by data traffic.

PaC PAA Message(seqno)[AVPs] ------------------------------------------------------ -----> PANA-PAA-Discover(0) <----- PANA-Start-Request(x)[Nonce, Cookie] -----> PANA-Start-Answer(x)[Nonce, Cookie] (continued to authentication phase)

PaC EP PAA Message(seqno)[AVPs] ------------------------------------------------------ ---->o (Data packet arrival or L2 trigger) ------> PAA-to-EP protocol, or another mechanism <------------ PANA-Start-Request(x)[Nonce, Cookie] ------------> PANA-Start-Answer(x)[Nonce, Cookie] (continued to authentication phase)

The main task in authentication phase is to carry EAP messages between the PaC and the PAA. EAP Request and Response messages are carried in PANA-Auth-Request messages. PANA-Auth-Answer messages are simply used to acknowledge receipt of the requests. As an optimization, a PANA-Auth-Answer message MAY include the EAP Response. Another optimization allows optionally carrying the first EAP Request/Response in PANA-Start-Request/Answer message as described in PANA allows execution of two separate authentication methods, one with NAP and one with ISP under the same PANA session. This optional feature may be offered by the PAA and accepted by the PaC. When performed separately, the result of first EAP authentication is signaled via PANA-FirstAuth-End-Request and PANA-FirstAuth-End-Answer message exchange which delineates the first method execution from the next. See for a detailed discussion on NAP/ISP separate authentication. The result of PANA authentication is carried in a PANA-Bind-Request message sent from PAA to PaC. This message carries the final EAP authentication method message (whether it is the second method of NAP and ISP separate authentication, or the sole authentication method) and the result of PANA authentication. The PANA-Bind-Request message MUST be acknowledged with a PANA-Bind-Answer (PBA) message. shows an example sequence in an authentication phase (no separate authentication).

PaC PAA Message(seqno)[AVPs] -------------------------------------------------------------------- (continued from discovery and handshake phase) <----- PANA-Auth-Request(x+1) [Session-Id, EAP{Request}] -----> PANA-Auth-Answer(x+1) // No piggybacking EAP-Response [Session-Id] -----> PANA-Auth-Request(y) [Session-Id, EAP{Response}] <----- PANA-Auth-Answer(y) [Session-Id] <----- PANA-Auth-Request(x+2) [Session-Id, EAP{Request}] -----> PANA-Auth-Answer(x+2) // Piggybacking EAP-Response [Session-Id, EAP{Response}] <----- PANA-Bind-Request(x+3) [Session-Id, EAP{Success}, Device-Id, Lifetime, Protection-Cap., PPAC, MAC] -----> PANA-Bind-Answer(x+3) [Session-Id, Device-Id, PPAC, MAC] When an EAP method that is capable of deriving keys is used during the authentication phase and the keys are successfully derived, the PANA message that carries the EAP Success (PANA-FirstAuth-End-Request, PANA-Bind-Request) and any subsequent message MUST contain a MAC AVP. The PANA-Bind-Request and the PANA-Bind-Answer message exchange is also used for binding device identifiers of the PaC and EP(s), and the IP address of the PAA to the PANA SA. To achieve this, the PANA-Bind-Request SHOULD contain the device identifier(s) of the EP(s) in Device-Id AVP(s) when they are either MAC or IP address(es), and the IP address of the PAA in an IP-Address AVP. PANA-Bind-Answer SHOULD contain PaC's device identifier in a Device-Id AVP when it is already presented with that of EP(s). The PaC MUST use the same type of device identifier as contained in the PANA-Bind-Request message. This exchange when protected by a MAC AVP prevents man-in-the-middle attacks. The PANA-Bind-Request message MAY also contain a Protection-Capability AVP to indicate if link-layer or network-layer ciphering should be initiated after PANA. No link-layer or network-layer specific information is included in the Protection-Capability AVP. It is assumed that the PAA is aware of the security capabilities of the access network. The PANA protocol does not specify how the PANA SA and the Protection-Capability AVP will be used to provide per-packet protection for data traffic. Additionally, the PANA-Bind-Request message MUST include a Post-PANA-Address-Configuration (PPAC) AVP, which helps the PAA to inform the PaC about whether a new IP address MUST be configured and the available methods to do so. The PaC MUST include a PPAC AVP in order to indicate its choice of method when there is a match between the methods offered by the PAA and the methods available on the PaC. When there is no match, a PPAC AVP MUST NOT be included and the Result-Code AVP MUST be set to PANA_PPAC_CAPABILITY_UNSUPPORTED in the PANA-Bind-Answer message. PANA-Bind-Request and PANA-Bind-Answer messages MUST be retransmitted based on the retransmission rule described in . EAP authentication can fail at a pass-through authenticator without sending an EAP-Failure message . When this occurs, the PAA SHOULD send a PANA-Error-Request message to the PaC with using PANA_UNABLE_TO_COMPLY result code. The PaC SHOULD not change its state unless the error message is secured by PANA or lower layer. In any case, a more appropriate way is to rely on a timeout on the PaC. There is a case where EAP authentication succeeds with producing an EAP-Success message but network access authorization fails due to, e.g., authorization rejected by a AAA proxy or authorization locally rejected by the PAA. When this occurs, the PAA MUST send PANA-Bind-Request with a result code PANA_AUTHORIZATION_REJECTED. If a AAA-Key is established between the PaC and the PAA by the time when the EAP-Success message is generated by the EAP server (this is the case when the EAP method provides protected success indication), this PANA-Bind message exchange MUST be protected with a MAC AVP and carry a Key-Id AVP. The AAA-Key and the PANA session MUST be deleted immediately after the PANA-Bind message exchange.

Once an authentication phase or a re-authentication phase successfully completes, the PaC gains access to the network and can send and receive IP data traffic through EP and the PANA session enters an authorized phase. In this phase, PANA-Ping-Request and PANA-Ping-Answer messages can be used for testing the liveness of the PANA session on the PANA peer. Both the PaC and the PAA are allowed to send a PANA-Ping-Request message to the communicating peer whenever they need to make sure the availability of the session on the peer and expect the peer to return a PANA-Ping-Answer message. Both PANA-Ping-Request and PANA-Ping-Answer messages MUST be protected with a MAC AVP when a PANA SA is available. Implementations MUST limit the rate of performing this test. The PaC and the PAA can handle rate limitation on their own, they do not have to perform any coordination with each other. There is no negotiation of timers for this purpose. and show liveness tests as they are initiated by the PaC and the PAA respectively.

PaC PAA Message(seqno)[AVPs] ------------------------------------------------------ -----> PANA-Ping-Request(q)[Session-Id, MAC] <----- PANA-Ping-Answer(q)[Session-Id, MAC]

PaC PAA Message(seqno)[AVPs] ------------------------------------------------------ <----- PANA-Ping-Request(p)[Session-Id, MAC] -----> PANA-Ping-Answer(p)[Session-Id, MAC]

A PANA session in an authorized phase can enter a re-authentication phase to extend the current session lifetime by re-executing EAP. Once the re-authentication phase successfully completes, the session re-enters the authorized phase. Otherwise, the session is deleted. When a PaC wants to initiate re-authentication, it sends a PANA-Reauth-Request message to the PAA. This message MUST contain a Session-Id AVP which is used for identifying the PANA session on the PAA. If the PAA already has an established PANA session for the PaC with the matching identifier, it MUST first respond with a PANA-Reauth-Answer, followed by a PANA-Auth-Request that starts a new EAP authentication. If PAA cannot identify the session, it MUST respond with a PANA-Error-Request with the error code PANA_UNKNOWN_SESSION_ID. PANA-Reauth-Request/Answer messages MUST contain a MAC AVP when PANA SA is available. PaC may receive a PANA-Auth-Request before receiving the answer to its outstanding PANA-Reauth-Request. This condition can arise due to packet re-ordering or a race condition between the PaC and PAA when they both attempt to engage in re-authentication. PaC MUST keep discarding the received PANA-Auth-Requests until it receives the answer to its request. When the PAA initiates re-authentication, it sends a PANA-Auth-Request message containing the session identifier for the PaC to enter an authentication phase. PAA SHOULD initiate EAP authentication before the current session lifetime expires. Re-authentication of an on-going PANA session MUST maintain the existing sequence numbers. For any re-authentication, if there is an established PANA SA, PANA-Auth-Request and PANA-Auth-Answer messages MUST be protected by adding a MAC AVP to each message. Any subsequent EAP-based authentication MUST be performed with the same ISP and NAP that was selected during the initial authentication. An example sequence for a re-authentication initiated by a PaC is shown in .

PaC PAA Message(seqno)[AVPs] ------------------------------------------------------ -----> PANA-Reauth-Request(q) [Session-Id, MAC] <----- PANA-Reauth-Answer(q) [Session-Id, MAC] <----- PANA-Auth-Request(p) [Session-Id, EAP{Request}, MAC] -----> PANA-Auth-Answer(p) // No piggybacking EAP-Response [Session-Id, MAC] -----> PANA-Auth-Request(q+1) [Session-Id, EAP{Response}, MAC] <----- PANA-Auth-Answer(q+1) // No piggybacking EAP-Response [Session-Id, MAC] <----- PANA-Auth-Request(p+1) [Session-Id, EAP{Request}, MAC] -----> PANA-Auth-Answer(p+1) // Piggybacking EAP-Response [Session-Id, EAP{Response}, MAC] <----- PANA-Bind-Request(p+2) [Session-Id, EAP{Success}, Device-Id, Key-Id, IP-Address, Lifetime, Protection-Cap., PPAC, MAC] -----> PANA-Bind-Answer(p+2) [Session-Id, Device-Id, Key-Id, PPAC, MAC]

A procedure for explicitly terminating a PANA session can be initiated either from the PaC (i.e., disconnect indication) or from the PAA (i.e., session revocation). The PANA-Termination-Request and the PANA-Termination-Answer message exchanges are used for disconnect indication and session revocation procedures. The reason for termination is indicated in the Termination-Cause AVP. When there is an established PANA SA established between the PaC and the PAA, all messages exchanged during the termination phase MUST be protected with a MAC AVP. When the sender of the PANA-Termination-Request receives a valid acknowledgment, all states maintained for the PANA session MUST be deleted immediately.

PaC PAA Message(seqno)[AVPs] ------------------------------------------------------ -----> PANA-Termination-Request(q)[Session-Id, MAC] <----- PANA-Termination-Answer(q)[Session-Id, MAC]

PANA allows running at most two EAP sessions in sequence in an authentication phase to support separate NAP and ISP authentication as described in this section. A typical network access authentication includes execution of one EAP method with the ISP. This separation allows PaC to perform an additional authentication method for receiving differentiated services from the NAP. Currently, running multiple EAP sessions in sequence in an authentication phase is designed only for separate NAP and ISP authentication. It is not for running arbitrary number of EAP sessions in sequence, or giving the PaC another chance to try another EAP authentication method within an integrated NAP and ISP authentication when an EAP authentication method fails. Within separate NAP and ISP authentication, the NAP authentication and the ISP authentication are considered completely independent. Presence or success of one should not effect the other. Making a network access authorization decision based on the success or failure of each authentication is a network policy issue.

When the PaC and PAA negotiates in the discovery and handshake phase to perform separate NAP and ISP authentication, the PaC and the PAA operate in the following way in addition to the behavior defined in In the discovery and handshake phase, the PAA MAY advertise availability of separate NAP and ISP authentication () by setting the S-flag on the message header of the PANA-Start-Request. If the S-flag of the received PANA-Start-Request message is set, the PaC can indicate its desire to perform separate NAP and ISP authentication by setting the S-flag in the PANA-Start-Answer message. If the S-flag of the received PANA-Start-Request message is not set, the PaC MUST NOT set the S-flag in the PANA-Start-Answer message sent back to the PAA. If the S-flag in the PANA-Start-Answer message is not set, only one authentication is performed (ISP-only) and the processing occurs as described in . When the S-flag is set in a PANA-Start-Request message, the initial EAP Request MUST NOT be carried in the PANA-Start-Request message. (If the initial EAP Request were contained in the PANA-Start-Request message during the S-flag negotiation, the PaC cannot tell whether the EAP Request is for NAP authentication or ISP authentication.)

When the PaC and PAA have negotiated in the discovery and handshake phase to perform separate NAP and ISP authentication, the PaC and the PAA operate in the following way in addition to the behavior defined in The S-flag of PANA-Auth-Request and PANA-Auth-Answer messages MUST be set. An EAP Success/Failure message is carried in a PANA-FirstAuth-End-Request (PFER) message as well as a PANA-Bind-Request (PBR) message. The PANA-FirstAuth-End-Request message MUST be used at the end of the first EAP authentication and the PANA-Bind-Request MUST be used for the second EAP authentication. The PANA-FirstAuth-End-Request messages MUST be acknowledged with a PANA-FirstAuth-End-Answer (PFEA) message. If the first EAP authentication has failed, the PAA can choose not to perform the second EAP authentication by clearing the S-flag of the PANA-FirstAuth-End-Request message. In this case, the S-flag of the PANA-FirstAuth-End-Answer message sent by the PaC MUST be cleared. If the S-flag of the PANA-FirstAuth-End-Request message is set when the first EAP authentication has failed, the PaC can choose not to perform the second EAP authentication by clearing the S-flag of the PANA-FirstAuth-End-Answer message. If the first EAP authentication failed and the S-flag is not set in the PANA-FirstAuth-End-Answer message as a result of those operations, the PANA session MUST be immediately deleted. Otherwise, the second EAP authentication MUST be performed. The PAA determines the execution order of NAP authentication and ISP authentication. In this case, the PAA can indicate which authentication (NAP authentication or ISP authentication) is currently occurring by using N-flag in the PANA message header. When NAP authentication is being performed, the N-flag MUST be set. When ISP authentication is being performed, the N-flag MUST NOT be set. The N-flag MUST NOT be set when S-flag is not set. When the PaC and PAA have negotiated in the discovery and handshake phase to perform separate NAP and ISP authentication, and the lower-layer is insecure, the two EAP authentication methods used in the separate authentication MUST be capable of deriving keys (AAA-Key).

When the PaC and PAA have negotiated in the discovery and handshake phase to perform separate NAP and ISP authentication, if the lower-layer is insecure, the two EAP authentication methods used in the separate authentication MUST be capable of deriving keys. In this case, if the first EAP authentication is successful, the PANA-FirstAuth-End-Request and PANA-FirstAuth-End-Answer messages as well as PANA-Auth-Request and PANA-Auth-Answer messages in the second EAP authentication MUST be protected with the key derived from the AAA-Key for the first EAP authentication. The PANA-Bind-Request and PANA-Bind-Answer messages and all subsequent PANA messages exchanged in authorized phase, re-authentication phase and termination phase MUST be protected either with the AAA-Key for the first EAP authentication if the first EAP authentication succeeds and the second EAP authentication fails, or with the AAA-Key for the second EAP authentication if the first EAP authentication fails and the second EAP authentication succeeds, or with the compound AAA-Key derived from the two AAA-Keys, one for the first EAP authentication and the other from the second EAP authentication, if both the first and second EAP authentications succeed.

PANA uses UDP as its transport layer protocol. The UDP port number is TBD. All messages except for PANA-PAA-Discover are always unicast. PANA-PAA-Discover MAY be unicast when the PaC knows the IP address of the PAA.

PANA does not provide fragmentation of PANA messages. Instead, it relies on fragmentation provided by EAP methods and IP layer when needed.

PANA uses sequence numbers to provide ordered and reliable delivery of messages. PaC and PAA maintain two sequence numbers: the next one to be used for a request it initiates and the next one it expects to see in a request from the other end. These sequence numbers are 32-bit unsigned numbers. They are monotonically incremented by 1 as new requests are generated and received, and wrapped to zero on the next message after 2^32-1. Answers always contain the same sequence number as the corresponding request. Retransmissions maintain the same sequence number. The initial sequence numbers (ISN) are randomly picked by PaC and PAA as they send their very first request messages. PANA-PAA-Discover message carries sequence number 0. When a request message is received, it is considered valid in terms of sequence numbers if and only if its sequence number matches the expected value. This check does not apply to PANA-PAA-Discover, and the very first request messages. When an answer message is received, it is considered valid in terms of sequence numbers if and only if its sequence number matches that of the currently outstanding request. A peer can only have one outstanding request at a time. PANA messages are retransmitted based on a timer until a response is received (in which case the retransmission timer is stopped) or the number of retransmission reaches the maximum value (in which case the PANA session MUST be deleted immediately). The initial discovery and handshake phase requires special handling. PaC MUST retransmit PANA-PAA-Discover if a subsequent PANA-Start-Request is not received in time. Even though a PANA-Start-Request is received, PANA-PAA-Discover may still have to be retransmitted. This is because a stateless PAA discovery requires one time transmission of a solicited PANA-Start-Request. PAA MUST NOT start a timer and retransmit the request in order to avoid state creation. If the received PANA-Start-Request included a Cookie AVP (an indication of stateless discovery), PaC MUST retransmit PANA-PAA-Discover until the first PANA-Auth-Request is received. Otherwise, PaC can rely on PAA to retransmit the PANA-Start-Requests as soon as PaC receives the first one (i.e., PaC can stop sending PANA-PAA-Discover). The retransmission timers SHOULD be calculated as described in to provide congestion control. See for default timer and maximum retransmission count parameters. PaC and PAA MUST respond to duplicate requests. Last transmitted PANA answer MAY be cached in case it is not received by the peer and that generates a retransmission of the last request. When available, a cached answer can be used instead of fully processing the retransmitted request and forming a new answer from scratch. PANA MUST NOT generate EAP message duplication. EAP payload of a retransmitted PANA message MUST NOT be passed to the EAP layer.

A PANA SA is created as an attribute of a PANA session when EAP authentication succeeds with a creation of a AAA-Key. A PANA SA is not created when the PANA authentication fails or no AAA-Key is produced by any EAP authentication method. In the case where two EAP authentications are performed in sequence in a single PANA authentication phase, it is possible that two AAA-Keys are derived. If this happens, the PANA SA MUST be generated from both AAA-Keys. When a new AAA-Key is derived as a result of EAP-based re-authentication, any key derived from the old AAA-Key MUST be updated to a new one that is derived from the new AAA-Key. In order to distinguish the new AAA-Key from old ones, one Key-Id AVP MUST be carried in PANA-Bind-Request and PANA-Bind-Answer messages or PANA-FirstAuth-End-Request and PANA-FirstAuth-End-Answer messages at the end of the EAP authentication which resulted in deriving a new AAA-Key. The Key-Id AVP is of type Unsigned32 and MUST contain a value that uniquely identifies the AAA-Key within the PANA session. The PANA-Bind-Answer message (or the PANA-FirstAuth-End-Answer message) sent in response to a PANA-Bind-Request message (or a PANA-FirstAuth-End-Request message) with a Key-Id AVP MUST contain a Key-Id AVP with the same AAA-Key identifier carried in the request. PANA-Bind-Request, PANA-Bind-Answer, PANA-FirstAuth-End-Request and PANA-FirstAuth-End-Answer messages with a Key-Id AVP MUST also carry a MAC AVP whose value is computed by using the new PANA-MAC-KEY derived from the new AAA-Key (or the new pair of AAA-Keys when the PANA_MAC_KEY is derived from two AAA-Keys). Although the specification does not mandate a particular method for calculation of Key-Id AVP value, a simple method is to use monotonically increasing numbers. The created PANA SA is deleted when the corresponding PANA session is deleted. The lifetime of the PANA SA is the same as the lifetime of the PANA session for simplicity. PANA SA attributes as well as PANA session attributes are listed below: Session-Id Device-Id of PaC IP address of PaC (may be the same as the Device-Id of PaC) IP address of PAA List of device identifiers of EPs Sequence number of the last transmitted request Sequence number of the last received request Last transmitted message payload Retransmission interval Session lifetime Protection-Capability PANA SA attributes: Nonce generated by PaC (PaC_nonce) Nonce generated by PAA (PAA_nonce) AAA-Key AAA-Key Identifier PANA_MAC_KEY The PANA_MAC_KEY is derived from the available AAA-Key(s) and it is used to integrity protect PANA messages. If there is only one AAA-Key available, e.g., due to ISP-only authentication, or with one failed and one successful NAP and ISP separate authentication (see ), the PANA_MAC_KEY computation is based on that single key. Otherwise, two AAA-Keys available to PANA can be combined in following way ('|' indicates concatenation): AAA-Key = AAA-Key1 | AAA-Key2 The PANA_MAC_KEY is computed in the following way: PANA_MAC_KEY = The first N bits of HMAC_SHA1(AAA-Key, PaC_nonce | PAA_nonce | Session-ID) where the value of N depends on the integrity protection algorithm in use, i.e., N=160 for HMAC-SHA1. The length of AAA-Key MUST be N bits or longer. See Section for the detailed usage of the PANA_MAC_KEY.

A PANA message can contain a MAC (Message Authentication Code) AVP for cryptographically protecting the message. When a MAC AVP is included in a PANA message, the value field of the MAC AVP is calculated by using the PANA_MAC_KEY in the following way: MAC AVP value = PANA_MAC_PRF(PANA_MAC_KEY, PANA_PDU) where PANA_PDU is the PANA message including the PANA header, with the MAC AVP value field first initialized to 0. PANA_MAC_PRF represents the pseudo random function corresponding to the MAC algorithm specified in the MAC AVP. In this version of draft, PANA_MAC_PRF is HMAC-SHA1. The PaC and PAA MUST use the same algorithm to calculate a MAC AVP they originate and receive. The algorithm is determined by the PAA when a PANA-Bind-Request with a MAC AVP is sent. When the PaC does not support the MAC algorithm specified in the PANA-Bind-Request message, it MUST silently discard the message. The PAA MUST NOT change the MAC algorithm throughout the continuation of the PANA session.

When a PANA message is received, the message is considered to be invalid at least when one of the following conditions are not met: The IP Hop Limit (or TTL) field has a value of 255, i.e., the packet could not possibly have been forwarded by a router. Each field in the message header contains a valid value including sequence number, message length, message type, version number, flags, etc. When a device identifier of the PaC is bound to the PANA session, it matches the device identifier carried in MAC or or IP header, or other locally-significant identifier provided by the lower-layers (e.g., circuit ID) unless the message is a PANA-Update-Request with an IP-Address AVP. The message type is one of the expected types in the current state. Specifically the following messages are unexpected and invalid: In discovery and handshake phase: PANA-Termination-Request and PANA-Ping-Request. PANA-Bind-Request. PANA-Update-Request. In authentication phase: PANA-PAA-Discover. PANA-Update-Request. PANA-Start-Request after a PaC receives the first valid PANA-Auth-Request. PANA-Termination-Request before the PaC receives the first successful PANA-Bind-Request. Authorized phase: PANA-Start-Request as well as a non-duplicate PANA-Bind-Request. PANA-PAA-Discover. In termination phase: PANA-PAA-Discover. All requests but PANA-Termination-Request. The message payload contains a valid set of AVPs allowed for the message type and there is no missing AVP that needs to be included in the payload. Each AVP is decoded correctly. When a MAC AVP is included, the AVP value matches the MAC value computed against the received message. When a Device-Id AVP is included, the AVP is valid if the device identifier type contained in the AVP is supported (check performed by both PaC and PAA) and is the requested one (check performed by PAA only) and the device identifier value contained in the AVP matches the value extracted from the lower-layer encapsulation header corresponding to the device identifier type contained in the AVP (check performed by PAA only). Note that a Device-Id AVP carries the PaC's device identifier in messages from PaC to PAA and EP(s)' device identifier in messages from PAA to PaC. When an IP-Address AVP is received in a message, the AVP is valid if the IP address matches the source address in the IP header. Invalid messages MUST be discarded in order to provide robustness against DoS attacks. In addition, an error notification message MAY be returned to the sender. See for details.

The device identifier used in the context of PANA can be an IP address, a MAC address, or an identifier that is not carried in data packets but has local significance in identifying a connected device (e.g., circuit id, PPP interface id). The last type of identifiers are commonly used in point-to-point links where MAC addresses are not available and lower-layers are already physically or cryptographically secured. It is assumed that the PAA knows the link type and the security mechanisms being provided or required on the access network (e.g., based on physical security, link-layer ciphers enabled before or after PANA, or IPsec). Based on that information, the PAA can decide what type of EP device id will be used when running PANA with the client. When IPsec-based security is the choice of access control, the PAA SHOULD provide IP address(es) as EP(s)' device ID, and expect the PaC to provide its IP address in return. In case IPsec is not used, MAC addresses are used as device IDs when available. If non-IPsec access control is enabled, and a MAC address is not available, device ID exchange does not occur within PANA. Instead, peers rely on lower-layers to provide locally-significant identifiers along with received PANA packets.

A PaC's IP address can change in certain situations. For example, the PANA framework describes a case in which a PaC replaces a pre-PANA address (PRPA) with a post-PANA address (POPA), and the PaC and PAA create host routes to each other in order to maintain on-link communication based on the POPA. The PAA needs to be notified about the change of PaC address. After the PaC has changed its address, it MUST send a PANA-Update-Request message to the PAA. The message MUST carry the new PaC address in an IP-Address AVP. If the address contained in the request is invalid, the PAA MUST send a PANA-Error message with the result code PANA_INVALID_IP_ADDRESS. Otherwise, the PAA MUST update the PANA session with the new PaC address and return a PANA-Update-Answer message. If there is an established PANA SA, both PANA-Update-Request and PANA-Update-Answer messages MUST be protected with a MAC AVP.

The authentication phase determines the PANA session lifetime when the network access authorization succeeds. The Session-Lifetime AVP MAY be optionally included in the PANA-Bind-Request message to inform PaC about the valid lifetime of the PANA session. It MUST be ignored when included in other PANA messages. The lifetime is a non-negotiable parameter that can be used by PaC to manage PANA-related state. PaC does not have to perform any actions when the lifetime expires, other than optionally purging local state. PAA SHOULD initiate EAP authentication before the current session lifetime expires. PaC and PAA MAY optionally rely on lower-layer indications to expedite the detection of a disconnected peer. Availability and reliability of such indications depend on the specific access technologies. PANA peer can use PANA-Ping-Request message to verify the disconnection before taking an action. The session lifetime parameter is not related to the transmission of PANA-Ping-Request messages. These messages can be used for asynchronously verifying the liveness of the peer. The decision to send PANA-Ping-Request message is taken locally and does not require coordination between the peers. When separate ISP and NAP authentication is performed, it is possible that different authorization lifetime values are associated with the two authentications. In this case, the smaller authorization lifetime value MUST be used for calculating the PANA Session-Lifetime value. As a result, when entering a re-authentication phase, both NAP and ISP authentication will be performed in the same re-authentication phase.

In a discovery and handshake phase, a PANA-Start-Request message sent from the PAA MAY contain zero or one NAP-Information AVP and zero or more ISP-Information AVPs to advertise the information on the NAP and/or ISPs. The PaC MAY indicate its choice of ISP by including an ISP-Information AVP in the PANA-Start-Answer message. When a AAA backend is used, the identity of the destination AAA server or realm MUST be determined based on the explicitly chosen ISP. When the ISP-Information AVP is not present, the access network MAY rely on the client identifier carried in the EAP authentication method to make this determination. The PaC can choose an ISP and contain an ISP-Information AVP for the chosen ISP in a PANA-Start-Answer message even when there is no ISP-Information AVP contained in the PANA-Start-Request message.

A PANA-Error-Request message MAY be sent by either the PaC or the PAA when a badly formed PANA message is received or in case of other errors. The receiver of this request MUST respond with a PANA-Error-Answer message. If the cause of this error message was a request message (e.g., PANA-PAA-Discover or *-Request), then the request MAY be retransmitted immediately without waiting for its retransmission timer to go off. If the cause of the error was a response message, the receiver of the PANA-Error-Request message SHOULD NOT resend the same response until it receives the next request. Erroneous PANA messages may be exploited by adverseries to launch DoS attacks on the victims. Unless the PaC or PAA rate-limits the generated PANA-Error-Request messages it may be overburdened by having tp respond to bogus packets. Limiting the number of error notifications sent to a given peer during a (configurable) period of time may be useful. When an error message is sent unprotected (i.e., no MAC AVP) and the lower-layer is insecure, the error message is treated as an informational message. The receiver of such an error message MUST NOT change its state unless the error persists and the PANA session is not making any progress.

A mobile PaC's network access authentication performance can be enhanced by deploying a context-transfer-based mechanism, where some session attributes are transferred from the previous PAA to the new one in order to avoid performing a full EAP authentication (reactive approach). Additional mechanisms that are based on the proactive AAA state establishment at one or more candidate PAAs may be developed in the future . The details of a context-transfer-based mechanism is provided in this section. Upon changing its point of attachment, a PaC that wants to quickly resume its ongoing PANA session without running EAP MAY send its unexpired PANA session identifier in its PANA-Start-Answer message. Along with the Session-Id AVP, a MAC AVP MUST be included in this message. The MAC AVP is computed by using the PANA_MAC_KEY shared between the PaC and its previous PAA that has an unexpired PANA session with the PaC. This action signals PaC's desire to perform the mobility optimization. In the absence of a Session-Id AVP in this message, the PANA session takes its usual course (i.e., EAP-based authentication is performed). If a PAA receives a session identifier in the PANA-Start-Answer message, and it is configured to enable this optimization, it SHOULD retrieve the PANA session attributes from the previous PAA. Current PAA determines the identity of the previous PAA by looking at the DiameterIdentity part of the PANA session identifier. The MAC AVP can only be verified by the previous PAA, therefore a copy of the PANA message MUST be provided to the previous PAA. The mechanism required to send a copy of the PANA-Start-Answer message from current PAA to the previous PAA, and retrieve the session attributes is outside the scope of PANA protocol. The Context Transfer Protocol might be useful for this purpose. When the previous or current PAA is not configured to enable this optimization, the current PAA can not retrieve the PANA session attributes, or the PANA session has already expired (i.e., session lifetime is zero), the PAA MUST send the PANA-Auth-Request message with a new session identifier and let the PANA exchange take its usual course. This action will engage EAP-based authentication and create a fresh PANA session from scratch. In case the current PAA can retrieve the on-going PANA session attributes from the previous PAA, the PANA session continues with a PANA-Bind exchange. As part of the context transfer, an intermediate AAA-Key material is provided by the previous PAA to the current PAA. AAA-Key-int = The first N bits of HMAC-SHA1(AAA-Key, DiameterIdentity | Session-ID) The value of N depends on the integrity protection algorithm in use, i.e., N=160 for HMAC-SHA1. DiameterIdentity is the identifier of the current PAA. Session-ID is the identifier of the PaC's PANA session with the previous PAA. The current PAA and PaC compute the new AAA-Key by using the nonce values and the AAA-Key-int. AAA-Key-new = The first N bits of HMAC-SHA1(AAA-Key-int, PaC_nonce | PAA_nonce) New PANA_MAC_KEY is computed based on the algorithm described in , by using the new AAA-Key and the new Session-ID assigned by the current PAA. The MAC AVP contained in the PANA-Bind-Request and PANA-Bind-Answer messages MUST be generated and verified by using the new PANA_MAC_KEY. The Session-ID AVP MUST include a new session identifier assigned by the current PAA. A new PANA session is created upon successful completion of this exchange. Note that correct operation of this optimization relies on many factors, including applicability of authorization state from one network attachment to another. identifies this operation as "fast handoff" and provides deployment considerations. Operators are recommended to take those guidelines into account when using this optimization in their networks.

This section defines message formats for PANA protocol.

The Hop Limit (or TTL) field of the IP header MUST be set to 255. When a PANA-PAA-Discover message is multicast, IP destination address of the message is set to a well-known link-local multicast address (TBD). A PANA-PAA-Discover message MAY be unicast in some cases as specified in . Any other PANA packet is unicast between the PaC and the PAA. The source and destination addresses SHOULD be set to the addresses on the interfaces from which the message will be sent and received, respectively. When the PANA packet is sent in response to a request, the UDP source and destination ports of the response packet MUST be copied from the destination and source ports of the request packet, respectively. The destination port of an unsolicited PANA packet MUST be set to an assigned value (TBD), and the source port MUST be set to a value chosen by the sender.

A summary of the PANA header format is shown below. The fields are transmitted in network byte order. 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Version | Reserved | Message Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Flags | Message Type | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Sequence Number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | AVPs ... +-+-+-+-+-+-+-+-+-+-+-+-+- This Version field MUST be set to 1 to indicate PANA Version 1. This 8-bit field is reserved for future use, and MUST be set to zero, and ignored by the receiver. The Message Length field is three octets and indicates the length of the PANA message including the header fields. The Flags field is eight bits. The following bits are assigned: 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |R S N r r r r r r r r r r r r r| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ If set, the message is a request. If cleared, the message is an answer. When the S-flag is set in a PANA-Start-Request message it indicates that PAA is willing to offer separate NAP and ISP authentication. When the S-flag is set in a PANA-Start-Answer message it indicates that the PaC accepts on performing separate NAP and ISP authentication. When the S-flag is set in a PANA-Auth-Request/Answer, PANA-FirstAuth-End-Request/Answer and PANA-Bind-Request/Answer messages it indicates that separate NAP and ISP authentication is being performed in the authentication phase. For other cases, S-flag MUST NOT be set. When the N-flag is set in a PANA-Auth-Request message, it indicates that the current EAP authentication is for NAP authentication. When the N-flag is unset in a PANA-Auth-Request message, it indicates that the current EAP authentication is for ISP authentication. The PaC MUST copy the value of the flag in its requests from the last received request of the PAA. The value of the flag on an answer MUST be copied from the request. The N-flag MUST NOT be set when S-flag is not set. These flag bits are reserved for future use, and MUST be set to zero, and ignored by the receiver. The Message Type field is two octets, and is used in order to communicate the message type with the message. The 16-bit address space is managed by IANA . PANA uses its own address space for this field. The Sequence Number field contains a 32 bit value. AVPs are a method of encapsulating information relevant to the PANA message. See section for more information on AVPs.

Each AVP of type OctetString MUST be padded to align on a 32-bit boundary, while other AVP types align naturally. A number of zero-valued bytes are added to the end of the AVP Data field till a word boundary is reached. The length of the padding is not reflected in the AVP Length field . The fields in the AVP header MUST be sent in network byte order. The format of the header is: 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | AVP Code | AVP Flags | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | AVP Length | Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-Id (opt) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Data ... +-+-+-+-+-+-+-+-+ The AVP Code, combined with the Vendor-Id field, identifies the attribute uniquely. AVP numbers are allocated by IANA . PANA uses its own address space for this field although some of the AVP formats are borrowed from Diameter protocol . The AVP Flags field is two octets. The following bits are assigned: 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V M r r r r r r r r r r r r r r| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The 'M' Bit, known as the Mandatory bit, indicates whether support of the AVP is required. The 'V' bit, known as the Vendor-Specific bit, indicates whether the optional Vendor-Id field is present in the AVP header. These flag bits are reserved for future use, and MUST be set to zero, and ignored by the receiver. The AVP Length field is four octets, and indicates the number of octets in this AVP including the AVP Code, AVP Length, AVP Flags, and the AVP data. This two-octet field is reserved for future use, and MUST be set to zero, and ignored by the receiver. The Vendor-Id field is present if the 'V' bit is set in the AVP Flags field. The optional four-octet Vendor-Id field contains the IANA assigned "SMI Network Management Private Enterprise Codes" value, encoded in network byte order. Any vendor wishing to implement a vendor-specific PANA AVP MUST use their own Vendor-Id along with their privately managed AVP address space, guaranteeing that they will not collide with any other vendor's vendor-specific AVP(s), nor with future IETF applications. The Data field is zero or more octets and contains information specific to the Attribute. The format and length of the Data field is determined by the AVP Code and AVP Length fields.

lists all PANA messages defined in this document.

Message Direction: PaC---PAA ---------------------------------------- PANA-PAA-Discover --------> PANA-Start-Request <-------- PANA-Start-Answer --------> PANA-Auth-Request <-------> PANA-Auth-Answer <-------> PANA-Reauth-Request --------> PANA-Reauth-Answer <-------- PANA-FirstAuth-End-Request <-------- PANA-FirstAuth-End-Answer --------> PANA-Bind-Request <-------- PANA-Bind-Answer --------> PANA-Ping-Request <-------> PANA-Ping-Answer <-------> PANA-Termination-Request <-------> PANA-Termination-Answer <-------> PANA-Update-Request --------> PANA-Update-Answer <-------- PANA-Error-Request <-------> PANA-Error-Answer <------->

Every PANA message MUST include a corresponding ABNF specification found in . Example: message ::= < PANA-Header: <Message type>, [REQ] [SEP] > * [ AVP ]

The PANA-PAA-Discover (PDI) message is used to discover the address of PAA(s). The sequence number in this message is always set to zero (0). PANA-PAA-Discover ::= < PANA-Header: 1 > * [ AVP ]

PANA-Start-Request (PSR) is sent by the PAA to the PaC to advertise availability of the PAA and start PANA authentication. The PAA sets the sequence number to an initial random value. PANA-Start-Request ::= < PANA-Header: 2, REQ [SEP] > { Nonce } [ Cookie ] [ EAP-Payload ] [ NAP-Information ] * [ ISP-Information ] [ Protection-Capability] [ PPAC ] * [ AVP ]

PANA-Start-Answer (PSA) is sent by the PaC to the PAA in response to a PANA-Start-Request message. This message completes the handshake to start PANA authentication. PANA-Start-Answer ::= < PANA-Header: 2 [SEP] > { Nonce } [ Session-Id ] [ Cookie ] [ EAP-Payload ] [ ISP-Information ] * [ AVP ] 0*1 < MAC >

PANA-Auth-Request (PAR) is either sent by the PAA or the PaC. Its main task is to carry an EAP-Payload AVP. PANA-Auth-Request ::= < PANA-Header: 3, REQ [SEP] [NAP] > < Session-Id > < EAP-Payload > * [ AVP ] 0*1 < MAC >

PANA-Auth-Answer (PAN) is sent by either the PaC or the PAA in response to a PANA-Auth-Request message. It MAY carry an EAP-Payload AVP. PANA-Auth-Answer ::= < PANA-Header: 3 [SEP] [NAP] > < Session-Id > [ EAP-Payload ] * [ AVP ] 0*1 < MAC >

PANA-Reauth-Request (PRAR) is sent by the PaC to the PAA to re-initiate EAP authentication. PANA-Reauth-Request ::= < PANA-Header: 4, REQ > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-Reauth-Answer (PRAA) is sent by the PAA to the PaC in response to a PANA-Reauth-Request message. PANA-Reauth-Answer ::= < PANA-Header: 4 > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-Bind-Request (PBR) is sent by the PAA to the PaC to deliver the result of PANA authentication. PANA-Bind-Request ::= < PANA-Header: 5, REQ [SEP] [NAP] > < Session-Id > { Result-Code } { PPAC } { IP-Address } [ EAP-Payload ] [ Session-Lifetime ] [ Protection-Capability ] [ Key-Id ] * [ Device-Id ] * [ AVP ] 0*1 < MAC >

PANA-Bind-Answer (PBA) is sent by the PaC to the PAA in response to a PANA-Result-Request message. PANA-Bind-Answer ::= < PANA-Header: 5 [,SEP] [NAP] > < Session-Id > { Result-Code } [ PPAC ] [ Device-Id ] [ Key-Id ] * [ AVP ] 0*1 < MAC >

PANA-Ping-Request (PPR) is either sent by the PaC or the PAA for performing liveness test. PANA-Ping-Request ::= < PANA-Header: 6, REQ > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-Ping-Answer (PPA) is sent in response to a PANA-Ping-Request. PANA-Ping-Answer ::= < PANA-Header: 6 > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-Termination-Request (PTR) is sent either by the PaC or the PAA to terminate a PANA session. PANA-Termination-Request ::= < PANA-Header: 7, REQ > < Session-Id > < Termination-Cause > * [ AVP ] 0*1 < MAC >

PANA-Termination-Answer (PTA) is sent either by the PaC or the PAA in response to PANA-Termination-Request. PANA-Termination-Answer ::= < PANA-Header: 7 > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-Error is sent either by the PaC or the PAA to report an error with the last received PANA message. PANA-Error-Request ::= < PANA-Header: 8 REQ > < Session-Id > < Result-Code > { Failed-AVP } * [ AVP ] 0*1 < MAC >

PANA-Error-Answer is sent in response to a PANA-Error-Request. PANA-Error-Answer ::= < PANA-Header: 8 > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA-FirstAuth-End-Request (PFER) is sent by the PAA to the PaC to signal the result of the first EAP authentication method when separate NAP and ISP authentication is performed. PANA-FirstAuth-End-Request ::= < PANA-Header: 9, REQ [SEP] [NAP] > < Session-Id > { EAP-Payload } { Result-Code } [ Key-Id ] * [ AVP ] 0*1 < MAC >

PANA-FirstAuth-End-Answer (PFEA) is sent by the PaC to the PAA in response to a PANA-FirstAuth-End-Request message. PANA-FirstAuth-End-Answer ::= < PANA-Header: 9, REQ [SEP] [NAP] > < Session-Id > [ Key-Id ] * [ AVP ] 0*1 < MAC >

PANA-Update-Request (PUR) is sent by the PaC to the PAA to update the attributes of the PANA session. Currently only the PaC IP address attribute can be updated via this mechanism. PANA-Update-Request ::= < PANA-Header: 10, REQ > < Session-Id > < IP-Address > * [ AVP ] 0*1 < MAC >

PANA-Update-Answer (PUA) is sent by the PAA to the PaC in response to a PANA-Update-Request. PANA-Update-Answer ::= < PANA-Header: 10 > < Session-Id > * [ AVP ] 0*1 < MAC >

PANA defines several AVPs that are specific to the protocol. A number of others AVPs are reused. These are specified in other documents such as . The following tables lists the AVPs used in this document, and specifies in which PANA messages they MAY, or MAY NOT be present. The table uses the following symbols: The AVP MUST NOT be present in the message. Zero or more instances of the AVP MAY be present in the message. Zero or one instance of the AVP MAY be present in the message. It is considered an error if there are more than one instance of the AVP. One instance of the AVP MUST be present in the message. At least one instance of the AVP MUST be present in the message.

+-----------------------------------------+ | Message | | Type | +-----+-----+-----+-----+-----+-----+-----+ Attribute Name | PSR | PSA | PAR | PAN | PBR | PBA | PDI | --------------------+-----+-----+-----+-----+-----+-----+-----+ Result-Code | 0 | 0 | 0 | 0 | 1 | 1 | 0 | Session-Id | 0 | 0-1 | 1 | 1 | 1 | 1 | 0 | Termination-Cause | 0 | 0 | 0 | 0 | 0 | 0 | 0 | EAP-Payload | 0-1 | 0-1 | 1 | 0-1 | 0-1 | 0 | 0 | MAC | 0 | 0-1 | 0-1 | 0-1 | 0-1 | 0-1 | 0 | Nonce | 1 | 1 | 0 | 0 | 0 | 0 | 0 | Device-Id | 0 | 0 | 0 | 0 | 0+ | 0-1 | 0 | Cookie | 0-1 | 0-1 | 0 | 0 | 0 | 0 | 0 | Protection-Cap. | 0-1 | 0 | 0 | 0 | 0-1 | 0 | 0 | PPAC | 0-1 | 0 | 0 | 0 | 1 | 0-1 | 0 | Session-Lifetime | 0 | 0 | 0 | 0 | 0-1 | 0 | 0 | Failed-AVP | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ISP-Information | 0+ | 0-1 | 0 | 0 | 0 | 0 | 0 | NAP-Information | 0-1 | 0 | 0 | 0 | 0 | 0 | 0 | Key-Id | 0 | 0 | 0 | 0 | 0-1 | 0-1 | 0 | IP-Address | 0 | 0 | 0 | 0 | 0 | 0 | 0 | --------------------+-----+-----+-----+-----+-----+-----+-----+

+-------------------------------------+ | Message | | Type | +-----+-----+-----+-----+------+------+ Attribute Name | PPR | PPA | PTR | PTA | PFER | PFEA | --------------------+-----+-----+-----+-----+------+------+ Result-Code | 0 | 0 | 0 | 0 | 1 | 0 | Session-Id | 1 | 1 | 1 | 1 | 1 | 1 | Termination-Cause | 0 | 0 | 1 | 0 | 0 | 0 | EAP-Payload | 0 | 0 | 0 | 0 | 1 | 0 | MAC | 0-1 | 0-1 | 0-1 | 0-1 | 0-1 | 0-1 | Nonce | 0 | 0 | 0 | 0 | 0 | 0 | Device-Id | 0 | 0 | 0 | 0 | 0 | 0 | Cookie | 0 | 0 | 0 | 0 | 0 | 0 | Protection-Cap. | 0 | 0 | 0 | 0 | 0 | 0 | PPAC | 0 | 0 | 0 | 0 | 0 | 0 | Session-Lifetime | 0 | 0 | 0 | 0 | 0 | 0 | Failed-AVP | 0 | 0 | 0 | 0 | 0 | 0 | ISP-Information | 0 | 0 | 0 | 0 | 0 | 0 | NAP-Information | 0 | 0 | 0 | 0 | 0 | 0 | Key-Id | 0 | 0 | 0 | 0 | 0-1 | 0-1 | IP-Address | 0 | 0 | 0 | 0 | 0 | 0 | --------------------+-----+-----+-----+-----+------+------+

+-------------------------------------+ | Message | | Type | +-----+-----+-----+-----+------+------+ Attribute Name | PUR | PUA | PER | PEA | PRAR | PRAA | --------------------+-----+-----+-----+-----+------+------+ Result-Code | 0 | 0 | 1 | 0 | 0 | 0 | Session-Id | 1 | 1 | 1 | 1 | 1 | 1 | Termination-Cause | 0 | 0 | 0 | 0 | 0 | 0 | EAP-Payload | 0 | 0 | 0 | 0 | 0 | 0 | MAC | 0-1 | 0-1 | 0-1 | 0-1 | 0-1 | 0-1 | Nonce | 0 | 0 | 0 | 0 | 0 | 0 | Device-Id | 0 | 0 | 0 | 0 | 0 | 0 | Cookie | 0 | 0 | 0 | 0 | 0 | 0 | Protection-Cap. | 0 | 0 | 0 | 0 | 0 | 0 | PPAC | 0 | 0 | 0 | 0 | 0 | 0 | Session-Lifetime | 0 | 0 | 0 | 0 | 0 | 0 | Failed-AVP | 0 | 0 | 1 | 0 | 0 | 0 | ISP-Information | 0 | 0 | 0 | 0 | 0 | 0 | NAP-Information | 0 | 0 | 0 | 0 | 0 | 0 | Key-Id | 0 | 0 | 0 | 0 | 0 | 0 | IP-Address | 1 | 0 | 0 | 0 | 0 | 0 | --------------------+-----+-----+-----+-----+------+------+

The MAC (Message Authentication Code) AVP is used to integrity protect PANA messages. The first octet of the this AVP (AVP Code 1) data contains the MAC algorithm type. Rest of the AVP data payload contains the MAC encoded in network byte order. The 8-bit Algorithm name space is managed by IANA . The AVP length varies depending on the used algorithm. 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Algorithm | MAC... +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1 HMAC-SHA1 (20 bytes) The Message Authentication Code is encoded in network byte order.

The Device-Id AVP (AVP Code 2) is used for carrying device identifiers of PaC and EP(s). The AVP data is of Address type . IPv4 and IPv6 addresses are encoded as specified in . The content and format of data (including byte and bit ordering) for link-layer addresses is expected to be specified in specific documents that describe how IP operates over different link-layers. For instance, . Address families other than that are defined for link-layer or IP addresses MUST NOT be used for this AVP.

All messages pertaining to a specific PANA session MUST include a Session-Id AVP (AVP Code 3) which carries a PAA-assigned fixed session identifier value throughout the lifetime of a session. When present, the Session-Id SHOULD appear immediately following the PANA header. The Session-Id MUST be globally and eternally unique, as it is meant to identify a PANA Session without reference to any other information, and may be needed to correlate historical authentication information with accounting information. The PANA Session-Id AVP has the same format as the Diameter Session-Id AVP .

The Cookie AVP (AVP Code 4) is used for carrying a cookie value generated by the PAA. The AVP data is of type OctetString. It is opaque and the exact content is outside the scope of this protocol.

The Protection-Capability AVP (AVP Code 5) indicates the cryptographic data protection capability supported and required by the EPs. The AVP data is of type Unsigned32. Below is a list of valid data values and associated protection capabilities: 0 L2_PROTECTION 1 IPSEC_PROTECTION

The Termination-Cause AVP (AVP Code 6) is used for indicating the reason why a session is terminated by the requester. The AVP data is of type Enumerated. The following Termination-Cause data values are used with PANA. The client initiated a disconnect The client was not granted access, or was disconnected, due to administrative reasons, such as the receipt of a Abort-Session-Request message. The session has timed out, and service has been terminated.

The Result-Code AVP (AVP Code 7) is of type Unsigned32 and indicates whether an EAP authentication was completed successfully or whether an error occurred. Here are Result-Code AVP values taken from and adapted for PANA.

These result code values inform the PaC about the authentication and authorization result. The authentication result and authorization result can be different as described below, but only one result is returned to the PaC. These codes are used with PANA-Bind-Request and PANA-FirstAuth-End-Request messages. Both authentication and authorization processes are successful. Authentication has failed. When this error is returned, it is assumed that authorization is automatically failed. Authorization has failed. This error could occur when authorization is rejected by a AAA proxy or rejected locally by a PAA, even if the authentication has succeeded.

These codes are used with PANA-Error-Request messages. Unless stated otherwise, they can be generated by both the PaC and the PAA. Message type not recognized or supported. PAA was unable to deliver the EAP payload to the authentication server. Only PAA can generate this code. A message was received whose bits in the PANA header were either set to an invalid combination, or to a value that is inconsistent with the message type definition. A message was received that included an AVP whose flag bits are set to an unrecognized value, or that is inconsistent with the AVP's definition. The received message contained an AVP that is not recognized or supported and was marked with the Mandatory bit. A PANA message with this error MUST contain one or more Failed-AVP AVP containing the AVPs that caused the failure. The message contained an unknown Session-Id. PAA MUST NOT send this error result code value to PaC if PaC sent an unknown Session-Id in the PANA-Start-Answer message (session resumption). The message contained an AVP with an invalid value in its data portion. A PANA message indicating this error MUST include the offending AVPs within a Failed-AVP AVP. The message did not contain an AVP that is required by the message type definition. If this value is sent in the Result-Code AVP, a Failed-AVP AVP SHOULD be included in the message. The Failed-AVP AVP MUST contain an example of the missing AVP complete with the Vendor-Id if applicable. The value field of the missing AVP should be of correct minimum length and contain zeroes. A message was received that cannot be authorized because the client has already expended allowed resources. An example of this error condition is a client that is restricted to one PANA session and attempts to establish a second session. Only PAA can generate this code. The PAA has detected AVPs in the message that contradicted each other, and is not willing to provide service to the client. One or more Failed-AVP AVPs MUST be present, containing the AVPs that contradicted each other. Only PAA can generate this code. A message was received with an AVP that MUST NOT be present. The Failed-AVP AVP MUST be included and contain a copy of the offending AVP. A message was received that included an AVP that appeared more often than permitted in the message definition. The Failed-AVP AVP MUST be included and contain a copy of the first instance of the offending AVP that exceeded the maximum number of occurrences. This error is returned when a message was received, whose version number is unsupported. This error is returned when a request is rejected for unspecified reasons. For example, when an EAP authentication fails at an EAP pass-through authenticator without passing an EAP-Failure message to the PAA, a Result-Code AVP with this error code is carried in PANA-Error-Request message. The message contained an AVP with an invalid length. The PANA-Error-Request message indicating this error MUST include the offending AVPs within a Failed-AVP AVP. This error is returned when a message is received with an invalid message length. This error is returned when the PaC receives a PANA-Bind-Request with a Protection-Capability AVP and a valid MAC AVP but does not support the protection capability specified in the Protection-Capability AVP. Only PaC can generate this code. This error is returned in a PANA-Bind-Answer message when there is no match between the list of PPAC methods offered by the PAA and the ones available on the PaC. Only PaC can generate this code. This error is returned in a PANA-Error-Request message when the IP-Address AVP in the received PANA-Update-Request message is invalid (e.g., a non-unicast address). Only PAA can generate this code.

The EAP-Payload AVP (AVP Code 8) is used for encapsulating the actual EAP packet that is being exchanged between the EAP peer and the EAP authenticator. The AVP data is of type OctetString.

The Session-Lifetime AVP (AVP Code 9) contains the number of seconds remaining before the current session is considered expired. The AVP data is of type Unsigned32.

The Failed-AVP AVP (AVP Code 10) provides debugging information in cases where a request is rejected or not fully processed due to erroneous information in a specific AVP. The AVP data is of type Grouped. The format of the Failed-AVP AVP is defined in .

The NAP-Information AVP (AVP Code 11) contains zero or one Provider-Identifier AVP which carries the identifier of the NAP and one Provider-Name AVP which carries the name of the NAP. The AVP data is of type Grouped, and it has the following ABNF grammar: NAP-Information ::= < AVP Header: 11 > 0*1 { Provider-Identifier } { Provider-Name } * [ AVP ]

The ISP-Information AVP (AVP Code 12) contains zero or one Provider-Identifier AVP which carries the identifier of the ISP and one Provider-Name AVP which carries the name of the ISP. The AVP data is of type Grouped, and it has the following ABNF grammar: ISP-Information ::= < AVP Header: 12 > 0*1 { Provider-Identifier } { Provider-Name } * [ AVP ]

The Provider-Identifier AVP (AVP Code 13) is of type Unsigned32, and contains an IANA assigned "SMI Network Management Private Enterprise Codes" value, encoded in network byte order.

The Provider-Name AVP (AVP Code 14) is of type UTF8String, and contains the UTF8-encoded name of the provider.

The Key-Id AVP (AVP Code 15) is of type Integer32, and contains an AAA-Key identifier. The AAA-Key identifier is assigned by PAA and MUST be unique within the PANA session.

The PPAC AVP (AVP Code 16) is used for conveying the available types of post-PANA IP address configuration mechanisms when sent by the PAA, and the chosen one when sent by the PaC. Each possible mechanisms is represented by a flag. At least one or more of the flags MUST be set when sent by the PAA, and exactly one flag MUST be set when sent by the PaC. The AVP data is of type Unsigned32. The format of the AVP data is as follows: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |N|D|A|T|I| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The PaC does not have to (if sent by PAA) or will not (if sent by PaC) configure a new IP address after PANA. The PaC can (if sent by PAA) or will (if sent by PaC) use DHCP to configure a new IP address after PANA. The PaC can/will use stateless IPv6 address autoconfiguration to configure a new IP address after PANA. The PaC can/will use to configure a new IP address after PANA. The PaC can/will use to configure a new IP address after PANA. These flag bits are reserved for future use, and MUST be set to zero, and ignored by the receiver. Unless the N-flag is set, the PaC MUST configure a new IP address using one of the methods indicated by the other flags. Refer to for a detailed discussion on when these methods can be used.

The Nonce AVP (AVP Code 17) carries a randomly chosen value that is used in cyrptographic key computations. The AVP data is of type OctetString and it contains a randomly generated value in opaque format. The data length MUST be between 8 and 256 bytes inclusive.

The IP-Address AVP (AVP Code 18) contains an IP address of the PaC or PAA. When it is sent by the PaC, it is used to convey the new IP address of the PaC to the PAA when the PaC reconfigures its IP address after the successful PANA authentication. This AVP is not used if the PaC's IP address used during the PANA authentication phase is still valid. It is sent by the PAA in PANA-Bind-Request to bind the IP address of the PAA to the PANA session. The payload format of the IP-Address AVP is the same as that of the Device-Id AVP (see See ). Address families for IPv4 or IPv6 MUST be used for this AVP.

The PANA protocol provides retransmissions for the PANA-PAA-Discover and request messages. The rule is that the sender of the request message retransmits the request if the corresponding answer is not received in time. Answer messages are sent as answers to the request messages, not based on a timer. PANA retransmission timers are based on the model used in DHCPv6 . Variables used here are also borrowed from this specification. PANA is a request response like protocol. The message exchange terminates when either the request sender successfully receives the appropriate answer, or when the message exchange is considered to have failed according to the retransmission mechanism described below. The retransmission behavior is controlled and described by the following variables: RT Retransmission timeout IRT Initial retransmission time MRC Maximum retransmission count MRT Maximum retransmission time MRD Maximum retransmission duration RAND Randomization factor With each message transmission or retransmission, the sender sets RT according to the rules given below. If RT expires before the message exchange terminates, the sender recomputes RT and retransmits the message. Each of the computations of a new RT include a randomization factor (RAND), which is a random number chosen with a uniform distribution between -0.1 and +0.1. The randomization factor is included to minimize synchronization of messages. The algorithm for choosing a random number does not need to be cryptographically sound. The algorithm SHOULD produce a different sequence of random numbers from each invocation. RT for the first message transmission is based on IRT: RT = IRT + RAND*IRT RT for each subsequent message transmission is based on the previous value of RT: RT = 2*RTprev + RAND*RTprev MRT specifies an upper bound on the value of RT (disregarding the randomization added by the use of RAND). If MRT has a value of 0, there is no upper limit on the value of RT. Otherwise: if (RT > MRT) RT = MRT + RAND*MRT MRC specifies an upper bound on the number of times a sender may retransmit a message. Unless MRC is zero, the message exchange fails once the sender has transmitted the message MRC times. MRD specifies an upper bound on the length of time a sender may retransmit a message. Unless MRD is zero, the message exchange fails once MRD seconds have elapsed since the client first transmitted the message. If both MRC and MRD are non-zero, the message exchange fails whenever either of the conditions specified in the previous two paragraphs are met. If both MRC and MRD are zero, the client continues to transmit the message until it receives a response.

This section presents a table of values used to describe the message retransmission behavior of PANA requests (REQ_*) and PANA-PAA-Discover message (PDI_*). The table shows default values. Parameter Default Description ------------------------------------------------ PDI_IRT 1 sec Initial PDI timeout. PDI_MRT 120 secs Max PDI timeout value. PDI_MRC 0 Configurable. PDI_MRD 0 Configurable. REQ_IRT 1 sec Initial Request timeout. REQ_MRT 30 secs Max Request timeout value. REQ_MRC 10 Max Request retry attempts. REQ_MRD 0 Configurable. So for example the first RT for the PBR message is calculated using REQ_IRT as the IRT: RT = REQ_IRT + RAND*REQ_IRT

This section provides guidance to the Internet Assigned Numbers Authority (IANA) regarding registration of values related to the PANA protocol, in accordance with BCP 26 . The following policies are used here with the meanings defined in BCP 26: "Private Use", "First Come First Served", "Expert Review", "Specification Required", "IETF Consensus", "Standards Action". This section explains the criteria to be used by the IANA for assignment of numbers within namespaces defined within this document. For registration requests where a Designated Expert should be consulted, the responsible IESG area director should appoint the Designated Expert. For Designated Expert with Specification Required, the request is posted to the PANA WG mailing list (or, if it has been disbanded, a successor designated by the Area Director) for comment and review, and MUST include a pointer to a public specification. Before a period of 30 days has passed, the Designated Expert will either approve or deny the registration request and publish a notice of the decision to the PANA WG mailing list or its successor. A denial notice must be justified by an explanation and, in the cases where it is possible, concrete suggestions on how the request can be modified so as to become acceptable.

PANA uses one well-known UDP port number (, and ), which needs to be assigned by the IANA.

PANA uses one well-known IPv4 multicast address for which the scope is limited to be link-local by setting the TTL field to 255, and one well-known IPv6 link-local scoped multicast address ( and ), which need to be assigned by the IANA.

As defined in , the PANA header contains two fields that requires IANA namespace management; the Message Type and Flags field.

The Message Type namespace is used to identify PANA messages. Values 0-65,533 are for permanent, standard message types, allocated by IETF Consensus . This document defines the Message Types 1-10. See through for the assignment of the namespace in this specification. The values 65,534 and 65,535 (hexadecimal values 0xfffe - 0xffff) are reserved for experimental messages. As these codes are only for experimental and testing purposes, no guarantee is made for interoperability between communicating PaC and PAA using experimental commands, as outlined in .

There are 16 bits in the Flags field of the PANA header. This document assigns bit 0 ('R'equest), bit 1 ('S'eparate) and bit 2 ('N'AP Authentication). The remaining bits MUST only be assigned via a Standards Action .

As defined in , the AVP header contains three fields that requires IANA namespace management; the AVP Code, AVP Flags and Vendor-Id fields where only the AVP Code and AVP Flags create new namespaces.

The AVP Code namespace is used to identify attributes. There are multiple namespaces. Vendors can have their own AVP Codes namespace which will be identified by their Vendor-ID (also known as Enterprise-Number) and they control the assignments of their vendor-specific AVP codes within their own namespace. The absence of a Vendor-ID or a Vendor-ID value of zero (0) identifies the IETF IANA controlled AVP Codes namespace. The AVP Codes and sometimes also possible values in an AVP are controlled and maintained by IANA. AVP Code 0 is not used. This document defines the AVP Codes 1-18. See through for the assignment of the namespace in this specification. AVPs may be allocated following Designated Expert with Specification Required . Release of blocks of AVPs (more than 3 at a time for a given purpose) should require IETF Consensus. Note that PANA defines a mechanism for Vendor-Specific AVPs, where the Vendor-Id field in the AVP header is set to a non-zero value. Vendor-Specific AVPs codes are for Private Use and should be encouraged instead of allocation of global attribute types, for functions specific only to one vendor's implementation of PANA, where no interoperability is deemed useful. Where a Vendor-Specific AVP is implemented by more than one vendor, allocation of global AVPs should be encouraged instead.

There are 16 bits in the AVP Flags field of the AVP header, defined in . This document assigns bit 0 ('V'endor Specific) and bit 1 ('M'andatory). The remaining bits should only be assigned via a Standards Action .

Certain AVPs in PANA define a list of values with various meanings. For attributes other than those specified in this section, adding additional values to the list can be done on a First Come, First Served basis by IANA .

As defined in , the Algorithm field of MAC AVP (AVP Code 1) defines the value of 1 (one) for HMAC-SHA1. All remaining values are available for assignment via IETF Consensus .

As defined in , the Protection-Capability AVP (AVP Code 5) defines the values 0 and 1. All remaining values are available for assignment via a Standards Action .

As defined in , the Termination-Cause AVP (AVP Code 6) defines the values 1, 4 and 8. All remaining values are available for assignment via IETF Consensus .

As defined in and the Result-Code AVP (AVP Code 7) defines the values 2001, 3001-3002, 3008-3009, 4001, 5001-5009 and 5011-5019. All remaining values are available for assignment via IETF Consensus .

As defined in , the Post-PANA-Address-Configuration AVP (AVP Code 17) defines the bits 0 ('N': no configuration), 1 ('D': DHCP), 2 ('A' stateless autoconfiguration), 3 ('T': DHCP with IPsec tunnel mode) and 4 ('I': IKEv2). All remaining values are available for assignment via a Standards Action .

The PANA protocol defines a UDP-based EAP encapsulation that runs between two IP-enabled nodes on the same IP link. Various security threats that are relevant to a protocol of this nature are outlined in . Security considerations stemming from the use of EAP and EAP methods are discussed in . This section provides a discussion on the security-related issues that are related to PANA framework and protocol design. An important element in assessing security of PANA design and deployment in a network is the presence of lower-layer (physical and link-layer) security. In the context of this document, lower-layers are said to be secure if they can prevent eavesdropping and spoofing of packets. Examples of such networks are physically-secured DSL networks and 3GPP2 networks with crytographically-secured cdma2000 link-layer. In these examples, the lower-layer security is enabled even before running the first PANA-based authentication. In the absence of such a pre-established secure channel, one needs to be created in conjunction with PANA using a link-layer or network-layer cryptographic mechanism (e.g., IPsec).

PANA provides multiple mechanisms to secure a PANA session. Since PaC and PAA are on the same IP link, a simple TTL check on the received PANA messages prevents off-link attacks. PANA messages carry sequence numbers, which are monotonically incremented by 1 with every new request message. These numbers are randomly initialized at the beginning of the session, and verified against expected numbers upon receipt. A message whose sequence number is different than the expected one is silently discarded. In addition to accomplishing orderly delivery of EAP messages and duplicate elimination, this scheme also helps prevent an adversary spoof messages to disturb ongoing PANA and EAP sessions unless it can also eavesdrop to synchronize on the expected sequence number. The PANA framework defines EP which is ideally located on a network device that can filter traffic from the PaCs before the traffic enters the Internet/intranet. A set of filters can be used to discard unauthorized packets, such as a PANA-Start-Request message that is received from the segment of the access network where only PaCs are supposed to be connected. The protocol also provides authentication and integrity protection to PANA messages when the used EAP method can generate cryptographic session keys. A PANA SA is generated based on the AAA-Key exported by the EAP method. This SA is used for generating per-packet MAC to protect the PANA header and payload (including the complete EAP message). The cryptographic protection prevents an adversary from acting as a man-in-the-middle, injecting messages, replaying messages and modifying the content of the exchanged messages. Any packet that fails to pass the MAC verification is silently discarded. The earliest this protection can be enabled is when the very first PANA-Bind-Request or PANA-FirstAuth-End-Request that signals a successful authentication is generated. Starting with these messages, any subsequent PANA message until the session gets torn down can be cryptographically protected. The PANA SA enables authenticated and integrity protected exchange of the device ID information between the PaC and PAA. This ensures there were no man-in-the-middle during the PANA authentication. The lifetime of the PANA SA is bounded by the AAA-authorized session lifetime with an additional tolerance period. Unless PANA state is updated by executing another EAP authentication, the PANA SA is removed when the current session expires. The ability to use cryptographic protection within PANA is determined by the used EAP method, which is generally dictated by the deployment environment. Insecure lower-layers necessitate use of key-generating EAP methods. In networks where lower-layers are already secured, cryptographic protection of PANA messages is not necessary.

The discovery and handshake phase is vulnerable to spoofing attacks as these messages are not authenticated and integrity protected. In order to prevent very basic denial-of service attacks an adversary should not be able to cause state creation by sending discovery messages to the PAA. This protection is achieved by using a cookie-based scheme (similar to which allows the responder (PAA) to be stateless in the first round of message exchange. A return-routability test does not provide additional protection as PANA traffic is not routed but simply forwarded on-link. It is difficult to prevent this threat entirely. In networks where lower-layers are not secured prior to running PANA, the capability discovery enabled through inclusion of Protection-Capability and Post-PANA-Address-Configuration AVPs in a PANA-Start-Request message is susceptible to spoofing leading to denial-of service attacks. Therefore, usage of these AVPs during the discovery and handshake phase in such insecure networks is NOT RECOMMENDED. The same AVPs are delivered via an integrity-protected PANA-Bind-Request upon successful authentication.

Eavesdropping EAP packets might cause problems when the EAP method is weak and enables dictionary or replay attacks or even allows an adversary to learn the long-term password directly. Furthermore, if the optional EAP Identity payload is used then it allows the adversary to learn the identity of the PaC. In such a case a privacy problem is prevalent. To prevent these threats, suggests using proper EAP methods for particular environments. Depending on the deployment environment an EAP authentication which supports user identity confidentiality, protection against dictionary attacks and session key establishment must be used. It is therefore the responsibility of the network operators and users to choose a proper EAP method.

The PANA design allows running two separate EAP sessions for the same PaC in a single authentication phase: one with the NAP, and one with the ISP. The process of arriving at the resultant authorization, which is a combination of the individual authorizations obtained from respective service providers, is outside the scope of this protocol. In the absence of lower-layer security, both authentications MUST be able to generate a AAA-Key, leading to generation of a PANA SA. The resultant PANA SA cryptographically binds the two AAA-Keys together, hence it prevents man-in-the-middle attacks.

When the EAP method exports a AAA-Key, this key is used to produce a PANA SA with PANA_MAC_KEY with a distinct key ID. The PANA_MAC_KEY is unique to the PANA session, and takes PANA-based nonce values into computation to cryptographically separate itself from the AAA-Key. The PANA_MAC_KEY is solely used for authentication and integrity protection of the PANA messages within the designated session. Two AAA-Keys may be generated as a result of separate NAP and ISP authentication. In that case, the AAA-Key used with the PANA SA is the combination of both keys. The PANA SA lifetime is bounded by the AAA-Key lifetime. Another execution of EAP method yields in a new AAA-Key, and updates the PANA SA, PANA_MAC_KEY and key ID. Upon PaC's movement to a another PAA (new PAA) and request to perform a context transfer based optimization, the current PAA computes a AAA-Key-int based on the AAA-Key, ID of new PAA, and the session ID. This AAA-Key-int is delivered to the new PAA, and used in the computation of AAA-Key-new, which further takes a pair of nonce values into account. After this point on, the AAA-Key-new becomes the AAA-Key between the PaC and the new PAA. When link-layer or network-layer ciphering is enabled as a result of successful PANA authentication, a separate PaC-EP master key is generated based on the AAA-Key, session ID, key ID, and EP ID. The lifetime of PaC-EP master key is bounded by the lifetime of the PANA SA. This key may be used with a secure association protocol to produce further cipher-specific and transient keys.

Networks that are not secured at the lower-layers prior to running PANA can rely on enabling per-packet data traffic ciphering upon successful PANA session establishment. The PANA framework allows generation of a PaC-EP master key from AAA-Key for using with a per-packet protection mechanism, such as link-layer or IPsec-based ciphering . In case the master key is not readily useful to the ciphering mechanism, an additional secure association protocol may be needed to produce the required keying material. These mechanisms ultimately establish a cryptographic binding between the data traffic generated by and for a client and the authenticated identity of the client. Data traffic must be minimally data origin authenticated, replay and integrity protected, and optionally encrypted.

The PANA framework allows separation of PAA from EP(s). SNMPv3 is used between the the PAA and EP for provisioning authorized PaC information on the EP. This exchange MUST be always physically or cryptographically protected for authentication, integrity and replay protection. It MUST also be privacy-protected when PaC-EP master key for per-packet ciphering is transmitted to the EP. The PaC-EP master key MUST be unique to the PaC and EP pair. The session ID and EP's device ID are taken into computation for achieving this effect . Compromise of an EP does not automatically lead to compromise of another EP or the PAA.

A PANA session is associated with a session lifetime. The session is terminated unless it is refreshed by a new round of EAP authentication before it expires. Therefore, at the latest a disconnected client can be detected when its session expires. A disconnect may also be detected earlier by using PANA ping messages. A request message can be generated by either PaC or PAA at any time and the peer must respond with an answer message. A successful round-trip of this exchange is a simple verification that the peer is alive. This test can be engaged when there is a possibility that the peer might have disconnected (e.g., after the discontinuation of data traffic for an extended period of time). Periodic use of this exchange as a keep-alive requires additional care as it might result in congestion and hence false alarms. This exchange is cryptographically protected when a PANA SA is available in order to prevent threats associated with the abuse of this functionality.

The mobility optimization described in involves the previous PAA (holding AAA-Key) providing a AAA-Key-new to the current PAA of the PaC. There are security risks stemming from potential compromise of PAAs. Compromise of the current PAA does not yield compromise of the previous PAA, as AAA-Key cannot be computed from a compromised AAA-Key-new. But a compromised previous PAA along with the intercepted nonce values on the current link leads to the compromise of AAA-Key-new. Operators should be aware of the potential risk of using this optimization. An operator can reduce the risk exposure by forcing the PaC to perform an EAP-based authentication immediately after the PaC gains access to new link via the optimized PANA execution. EAP-based authentication can be run in parallel with IP data packet transmission.

Even though the IP-Address AVP in a PANA-Update-Request can be cryptographically protected by the MAC AVP, there is not way to prove the ownership of the IP address presented by the PaC. Hence an authorized PaC can launch a redirect attack by spoofing a victim's IP address.

The PANA protocol supports the ability for both the PaC and the PAA to transmit a tear-down message before the session lifetime expires. This message causes state removal, a stop of the accounting procedure and removes the installed per-PaC state on the EP(s). This message is cryptographically protected when PANA SA is present.

We would like to thank Jari Arkko, Mohan Parthasarathy, Julien Bournelle, Rafael Marin Lopez, Pasi Eronen, Randy Turner, Erik Nordmark, Lionel Morand, Avi Lior, Susan Thomson, Giaretta Gerardo and all members of the PANA working group for their valuable comments to this document.