Internet-Draft | CoAP-EAP | October 2021 |
Marin-Lopez & Garcia-Carrillo | Expires 28 April 2022 | [Page] |
This document specifies an authentication service that uses the Extensible Authentication Protocol (EAP) transported employing Constrained Application Protocol (CoAP) messages. As such, it defines an EAP lower layer based on CoAP called CoAP-EAP. One of the primer goals is to authenticate a CoAP-enabled IoT device (EAP peer) that intends to join a security domain managed by a Controller (EAP authenticator). Secondly, it allows deriving key material to protect CoAP messages exchanged between them based on Object Security for Constrained RESTful Environments (OSCORE), enabling the establishment of a security association between them.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 28 April 2022.¶
Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
This document specifies an authentication service (application) that uses the Extensible Authentication Protocol (EAP) [RFC3748] and is built on top of the Constrained Application Protocol (CoAP) [RFC7252] called CoAP-EAP. CoAP-EAP is an application that allows authenticating two CoAP endpoints by using EAP, and to establish an Object Security for Constrained RESTful Environments (OSCORE) security association between them.¶
More specifically, this document specifies how CoAP can be used as a constrained, link-layer independent, reliable EAP lower layer [RFC3748] to transport EAP messages between a CoAP server (acting as EAP peer) and a CoAP client (acting as EAP authenticator) using CoAP messages. The CoAP client has the option of contacting a backend AAA infrastructure to complete the EAP negotiation as described in the EAP specification [RFC3748].¶
EAP methods transported in CoAP MUST generate cryptographic material [RFC5247] for this specification. This way, CoAP messages are protected after the authentication. After CoAP-EAP's operation, an OSCORE security association is established between endpoints of the service.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
Readers are expected to be familiar with the terms and concepts of described in CoAP [RFC7252], EAP [RFC3748] and OSCORE [RFC8613].¶
Figure 1 illustrates the architecture defined in this document. Basically, an IoT device, acting as the EAP peer, wants to be authenticated by using EAP to join a domain that is managed by a Controller acting as EAP authenticator. The IoT device will act a CoAP server for this service, and the EAP authenticator as a CoAP client. The rationale behind this decision, as expanded later, is that EAP requests go always from the EAP server to the EAP peer. Accordingly, the EAP responses go from the EAP peer to the EAP server.¶
It is worth noting that the CoAP client (EAP authenticator) MAY interact with a backend AAA infrastructure when EAP pass-through mode is used, which will place the EAP server in the AAA server that contains the information required to authenticate the EAP peer.¶
CoAP-EAP uses reliable delivery in CoAP ([RFC7252],[RFC8323]). Therefore, EAP retransmission time is set to infinite as mentioned in [RFC3748]. To keep ordering guarantee, CoAP-EAP uses Hypermedia as the Engine of Application State (HATEOAS). In particular, each step during the EAP authentication is represented as a new resource in the EAP peer (CoAP server). The previous resource is removed once the new resource is created indicating the resource that will process the next expected step of the EAP authentication.¶
It is assumed that any EAP method that exports cryptographic material is valid. For example, EAP-MD5 cannot be used since it does not export key material. It is RECOMMENDED a light EAP method for the case of constrained link and constrained IoT devices.¶
In CoAP-EAP, the IoT device (EAP peer/CoAP server) will only have one authentication session with a specific Controller (EAP authenticator/ CoAP client) and it will not process any other EAP authentication in parallel (with the same Controller). That is, a single ongoing EAP authentication is permitted for the same IoT device and same Controller. Moreover, EAP is a lock-step protocol ([RFC3748]). The article [eap-framework], highlights the benefits of the EAP framework.¶
To access the authentication service, this document defines the well-known URI "/.well-known/coap-eap" (to be assigned by IANA). This URI is referring to the authentication service that is present in the Controller so that IoT device can start the service.¶
Prior to the CoAP-EAP exchange takes place, the IoT device needs to discovers the Controller or the entity that will enable the exchange between the IoT and the Controller (e.g., an intermediary such as a proxy).¶
The discovery process is out of scope of this document. This document provides the specification using the mechanisms provided by CoAP to discover the Controller for CoAP-EAP.¶
The CoAP-EAP application is designated by the well-known URI "coap-eap" for the trigger message (Step 0). The CoAP-EAP service can be discovered by asking directly the services offered. This information can be also available in the resource directory.¶
Implementation Notes: On the methods on how the IPv6 address of the Controller or intermediary entity can be discovered, there can be different methods depending on the specific deployment. For example, on a 6LoWPAN network, the Border Router will typically act as the Controller, hence after receiving the Router Advertisement (RA) messages from the Border Router, the IoT device may engage on the CoAP-EAP exchange. Different protocols can be used to discover the IP of the Controller. Example of this protocols are Multicast DNS (mDNS) [RFC6762] or DHCPv6 [RFC8415].¶
Figure 2 shows the general flow of operation for CoAP-EAP to authenticate using EAP and establish an OSCORE security context. The flow does not show a specific EAP method. Instead we represent the chosen EAP method by using a generic name (EAP-X). The flow assumes that the IoT device knows the Controller implements the CoAP-EAP service. The specific mechanism of discovery is out-of-scope of this document. Some comments about Controller discovery is in Section 3.1.¶
The steps for the operation happens as follows:¶
When the CoAP-EAP state is close to expire, the IoT device MAY want to start a re-authentication process to renew the state. Since the initial authentication is usually taxing, it is assumed to be done only once over a long period of time (e.g. based on Session-Lifetime). Certain EAP methods such as EAP-NOOB [I-D.ietf-emu-eap-noob] or EAP-AKA' [RFC5448] provides fast reconnect for quicker re-authentication. The EAP re-authentication protocol (ERP) [RFC6696] MAY be also used for avoiding the repetition of the entire EAP exchange.¶
The message flow for the reauthentication will be exactly the same as the one shown in Figure 2. Nevertheless, two different CoAP-EAP states will be active during the re-authentication: the current CoAP-EAP state and the new CoAP-EAP state, which will be created once the reauthentication has finished with success. Once the reauthentication is completed successfully, the current CoAP-EAP state is deleted and the new CoAP-EAP becomes the current one. If by any reason, the reauthentication fails to complete, the current CoAP-EAP state will be available until it expires or it is renewed in another try of re-authentication.¶
If the reauthentication fails, it is up to the IoT device decide when to restart a reauthentication before the current EAP state expires.¶
The IoT device and the Controller keep a state during the CoAP-EAP negotiation. The CoAP-EAP state includes several important parts:¶
Once created, the Controller MAY choose to delete it as described in Figure 3. On the other hand, the IoT device may need to renew the CoAP-EAP state because the key material is close to expire, as mentioned in Section 3.3.¶
There are situations where the current CoAP-EAP state might need to be removed. For instance due to its expiration or a forced removal if the IoT device needs to be expelled from the security domain. This exchange is illustrated in Figure 3.¶
If the Controller deems necessary the removal of the CoAP-EAP state from the IoT device before it expires, it can send a DELETE command in a request to the IoT device, referencing the last CoAP-EAP state resource given by the CoAP server, whose identifier will be the last one received (e.g. '/a/w' in Figure 2). This message is protected with the OSCORE security association to prevent forgery. Upon reception of this message, the CoAP server sends a response to the Controller with the Code '2.02 Deleted', which is also protected with the OSCORE security association. If a response from the IoT device does not arrive after EXCHANGE_LIFETIME the Controller will remove the state from its side.¶
This section elaborates how different errors are handled, from EAP authentication failure, a non-responding endpoint, lost messages or initial POST message arriving out of place.¶
EAP authentication MAY fail for different situations (e.g. wrong credentials). The result is that the Controller will send an EAP failure as a result of the EAP authentication (Step 7 in Figure 2). In this case, the IoT device MUST send a response '4.01 Unauthorized' in Step 8. As a consequence, Step 7 and Step 8 are not protected in this case because no MSK is exported and the OSCORE security context is not generated.¶
If the EAP authentication fails during the re-authentication and the Controller sends an EAP failure, the current CoAP-EAP state will be still usable until it expires.¶
If, by any reason, one of the entities becomes non-responding, the CoAP-EAP state SHOULD be kept only for a period of time before it is removed. The removal of the CoAP-EAP state in the Controller assumes that the IoT device will need to authenticate again. According to CoAP, EXCHANGE_LIFETIME considers the time it takes until a client stops expecting a response to a request. A timer is reset every time a message is sent. If EXCHANGE_LIFETIME has passed waiting for the next message, both entities will delete the CoAP-EAP state if the authentication process has not finished correctly.¶
The reception of the trigger message in Step 0 containing /.well-known/coap-eap needs some additional considerations, as the resource is always available in the EAP authenticator.¶
If a trigger message (Step 0) arrives to the Controller during an ongoing authentication, the Controller MUST silently discard this trigger message.¶
If an old "POST /.well-known/coap-eap" (Step 0) arrives to the Controller and there is no authentication ongoing, the Controller may understand that a new authentication process is requested. Consequently, the Controller will start a new EAP authentication. However, the IoT device did not start any authentication and therefore it has not selected any resource for the EAP authentication. Thus, IoT device sends a '4.04 Not found' in the response (Figure 4).¶
The CoAP-EAP operation is intended to be compatible with the use of intermediary entities between the IoT device and the Controller, when direct communication is not possible. In this context, CoAP proxies can be used as enablers of the CoAP-EAP exchange.¶
This specification is limited to use standard CoAP [RFC7252] as well as standardized CoAP options [RFC8613]. It does not specify any addition in the form of CoAP options. This is expected to ease the integration of CoAP intermediaries in the CoAP-EAP exchange.¶
There is a consideration that needs to be taken into account, when using proxies in the CoAP-EAP, as the exchange contains a role-reversal process at the beginning of the exchange. In the first message, the IoT device acts as a CoAP client, and the Controller as the CoAP server. After that, remaining exchanges the roles are reversed, being the IoT device, the CoAP server and the Controller, the CoAP client.¶
In the CoAP-EAP exchange, there is information that needs to be exchanged between the two entities. Examples of these are the ciphersuites that need to be negotiated or authorization information (Session-lifetime). There may be also a need of extending the information that has to be exchanged in the future. This section specifies the CBOR [RFC8949] data structure to exchange information between the IoT device and the Controller in the CoAP payload.¶
Next, is the specification of the CBOR Object to exchange information in CoAP-EAP¶
The parameters contain the following information: ciphersuite: It contains a array with the list of the proposed or selected CBOR algorithms for OSCORE. If the field is carried over a request, the meaning is the proposed ciphersuite, if it is carried over a response, corresponds to the response.¶
OSCORE runs after the EAP authentication, using the ciphersuite selected in the ciphersuite negotiation (Step 1 and 2). To negotiate the ciphersuite, CoAP-EAP follows a simple approach: the Controller sends a list, in decreasing order or preference, with the identifiers of the supported ciphersuites (Step 1). In the response to that message (Step 2), the IoT device sends a response with the choice.¶
This list is included in the payload after the EAP message with a CBOR array that contains the ciphersuites. An example of how the fields are arranged in the CoAP payload can be seen in Figure 6. An example of the exchange with the ciphersuite negotiation is shown in Figure 7, where can be appreciated the disposition of both EAP-Request/Identity and EAP-Response/Identity, followed by the CBOR object defined in Section 4, containing in the ciphersuite field the CBOR array for the ciphersuite negotiation.¶
In case there is no CBOR array stating the ciphersuites, the default ciphersuites are applied. If the Controller sends a restricted list of ciphersuites that is willing to accept, and the ones supported by the IoT device are not in that list, the IoT device will respond with a '4.00 Bad Request', expressing in the payload the ciphersuites supported. Figure 8 illustrates this exchange.¶
The ciphersuite requirements are inherited from the ones established by OSCORE. By default, the HKDF algorithm is SHA-256 and the AEAD algorithm is AES-CCM-16-64-128. Both are mandatory to implement. The other ciphersuites supported and negotiated in the ciphersuite negotiation are the following:¶
This specification uses the (HMAC)-based key derivation function (HKDF) defined in [RFC5869] to derive the necessary key material. Since the key derivation process uses the MSK, which is considered fresh key material, we will use the HKDF-Expand function, which we will shorten here as KDF.¶
The derivation of the security context for OSCORE allows securing the communication between the IoT device and the Controller once the MSK has been exported providing, confidentiality, integrity, key confirmation (Step 7 and 8) and detecting a downgrading attack.¶
The Master Secret can be derived by using the chosen ciphersuite and the KDF. The Master Secret can be derived as follows:¶
Master Secret = KDF(MSK, CS | "OSCORE MASTER SECRET", length)¶
where:¶
The Master Salt, similarly to the Master Secret, can be derived as follows:¶
Master Salt = KDF(MSK, CS | "OSCORE MASTER SALT", length)¶
where:¶
Since the MSK is used derive the Master Key, the correct verification of the OSCORE protected request (Step 7) and response (Step 8) confirms the Controller and the IoT device have the same Master Secret, achieving key confirmation.¶
To prevent a downgrading attack, the content of the ciphersuites negotiation (which we refer to here as CS) is embedded in the Master Secret derivation. If an attacker changes the value of the ciphersuite negotiation, the result will be different OSCORE security contexts, that ends up with a failure in Step 7 and 8.¶
This section discusses the suitability of the CoAP protocol as EAP lower layer, and reviews the requisites imposed by the EAP protocol to any protocol that transports EAP. What EAP expects from its lower layers can be found in section 3.1 of [RFC3748], which is elaborated next:¶
Unreliable transport. EAP does not assume that lower layers are reliable but it can benefit for a reliable lower layer. In this sense, CoAP provides a reliability mechanism (e.g. through the use of Confirmable messages).¶
Lower layer error detection. EAP relies on lower layer error detection (e.g., CRC, Checksum, MIC, etc.). CoAP goes on top of UDP/TCP which provides a checksum mechanism over its payload.¶
Lower layer security. EAP does not require security services from the lower layers.¶
Minimum MTU. Lower layers need to provide an EAP MTU size of 1020 octets or greater. CoAP assumes an upper bound of 1024 for its payload which covers the requirements of EAP.¶
Ordering guarantees. EAP relies on lower layer ordering guarantees for correct operation. Regarding message ordering, every time a new message arrives at the authentication service hosted by the IoT device, a new resource is created and this is indicated in a "2.01 Created" response code along with the name of the new resource via Location-Path or Location-Query. This way the application indicates that its state has advanced. Although the [RFC3748] states: "EAP provides its own support for duplicate elimination and retransmission", EAP is also reliant on lower layer ordering guarantees. In this regard, [RFC3748] talks about possible duplication and says: "Where the lower layer is reliable, it will provide the EAP layer with a non-duplicated stream of packets. However, while it is desirable that lower layers provide for non-duplication, this is not a requirement". CoAP is providing a non-duplicated stream of packets and accomplish the "desirable" non-duplication. In addition, [RFC3748] says that when EAP runs over a reliable lower layer "the authenticator retransmission timer SHOULD be set to an infinite value, so that retransmissions do not occur at the EAP layer."¶
Regarding the impact that an EAP lower layer will have to the total byte size of the whole exchange, there is a comparison with another network layer based EAP lower layer, PANA [RFC5191], in [coap-eap]. Comparing the EAP lower layer (alone) and taking into account EAP. On the one hand, at the EAP lower layer level, the usage of CoAP gives important benefits. On the other hand, when taking into account the EAP method overload, this reduction is less but still significant if the EAP method is lightweight (authors of [coap-eap] used EAP-PSK as a representative of a lightweight EAP method). If the EAP method is very taxing, the impact of the reduction in size of the EAP lower layer is less significant. This leads to the conclusion that possible next steps in this field could be designing new EAP methods that can be better adapted to the requirements of IoT devices and networks.¶
However, the impact of the EAP lower layer itself cannot be ignored, hence the proposal of using CoAP as lightweight protocol for this purpose. Other EAP methods such as EAP-AKA'[RFC5448] or new lightweight EAP methods such as EAP-NOOB [I-D.ietf-emu-eap-noob] or EAP-EDHOC [I-D.ingles-eap-edhoc] that can benefit, as well as new ones that may be proposed in the future with IoT constraints in mind, from a CoAP-based EAP lower layer.¶
There are some aspects to be considered such as how authorization is managed, the use of MSK as keying material and how the trust in the Controller is established. Additional considerations such as EAP channel binding as per [RFC6677] are also discussed here.¶
Authorization is part of bootstrapping. It serves to establish whether the node can join and the set of conditions it has to adhere. The authorization data will be gathered from the organization that is responsible for the IoT device and sent to the EAP authenticator in case of AAA infrastructure is deployed.¶
In standalone mode, the authorization information will be in the Controller. If the pass-through mode is used, authorization data received from the AAA server can be delivered by the AAA protocol (e.g. RADIUS or Diameter). Providing more fine-grained authorization data can be with the transport of SAML in RADIUS [RFC7833].¶
After bootstrapping, additional authorization information to operate in the security domain, e.g., access services offered by other nodes, can be taken care of by the solutions proposed in the ACE WG.¶
In CoAP-EAP there is no nonce exchange to provide freshness to the keys derived from the MSK. The MSK and Extended Master Session Key (EMSK) keys according to the EAP Key Management Framework [RFC5247] are fresh key material. Since only one authentication is established per EAP authenticator, there is no need for generating additional key material. In case a new MSK is required, a re-authentication can be done, by running the process again, or using a more lightweight EAP method to derive additional key material as elaborated in Section 3.3.¶
According to the [RFC6677], channel binding related with EAP, is sent through the EAP method that supports it.¶
To satisfy the requirements of the document, we need to send the EAP lower layer identifier (To be assigned by IANA), in the EAP Lower-Layer Attribute if RADIUS is used.¶
Other security-related concerns can be how to ensure that the node joining the security domain can in fact trust the Controller. This issue is elaborated in the EAP Key Management Framework [RFC5247]. In particular, the constrained node knows it can trust the Controller because the key that is used to establish the security association is derived from the MSK. If the Controller has the MSK, it is clear the AAA Server of the node trusted the Controller, which can be considered as a trusted party.¶
Considerations for IANA regarding this document:¶
We would like to thank as the reviewers of this work: Carsten Bormann, Mohit Sethi, Benjamin Kaduk, Christian Amsuss, Alexandre Petrescu, Pedro Moreno-Sanchez and Eduardo Ingles-Sanchez.¶
We would also like to thank Gabriel Lopez-Millan for the first review of this document and we would like to thank Ivan Jimenez-Sanchez for the first proof-of-concept implementation of this idea.¶
And thank for their valuables comments to Alexander Pelov and Laurent Toutain, especially for the potential optimizations of CoAP-EAP.¶
CoAP-EAP makes possible to derive a PSK for (D)TLS to allow PSK-based authentication between the IoT device and the Controller.¶
Figure 9 shows the last steps of the operation for CoAP-EAP when (D)TLS is used to protect the communication between the IoT device and the Controller using the cryptographic material exported by the EAP authentication. The general flow is essentially the same as in the case of OSCORE, except that DTLS negotiation is established in Step 7). Once DTLS negotiation has finished successfully the IoT device is granted access to the domain. Step 7 MUST be interpreted by the IoT device as an alternate success indication, which will end up with the MSK and the DTLS_PSK derivation for the (D)TLS authentication based on PSK.¶
According to [RFC8446] the provision of the PSK out-of-band also requires the provision of the KDF hash algorithm and the PSK identity. To simplify the design in CoAP-EAP, the KDF hash algorithm can be included in the list of ciphersuites exchange in Step 1 and Step 2 if DTLS wants to be used instead of OSCORE. For the same reason, the PSK identity is derived from Sender ID (SID) and Recipient ID (SID) as defined in Appendix A.2.¶
It is also possible to derive a pre-shared key for DTLS to establish a DLTS security association after a successful EAP authentication. Analogously to how the ciphersuite is negotiated for OSCORE Section 5.1, the Controller sends a list, in decreasing order of preference, with the identifiers of the ciphersuites supported (Step 1). In the response, the IoT device sends the choice.¶
This list is included in the payload after the EAP message with a CBOR array that contains the ciphersuites. This CBOR array is enclosed as one of the elements of the CBOR Object used for transporting information in CoAP-EAP (See Section 4. An example of how the fields are arranged in the CoAP payload can be seen in Figure 6.¶
In case there is no CBOR array stating the ciphersuites, the default ciphersuite is applied. If the Controller sends a restricted list of ciphersuites that is willing to accept, and the ones supported by the IoT device are not in that list, the IoT device will respond with a '4.00 Bad Request', expressing in the payload the ciphersuites supported. Figure 8 illustrates this exchange.¶
The ciphersuites are the following:¶
To enable DTLS after an EAP authentication using the key material generated, we define the Identity and the PSK for DTLS. The Identity in this case is generated by concatenating the exchanged Sender ID and the Recipient ID.¶
PSK Identity = SID || RID¶
It is also possible to derive a pre-shared key for DTLS [RFC6347], refereed to here as "DTLS PSK", from the MSK between both IoT device and Controller if required. The length of the DTLS PSK will depend on the ciphersuite. To have a cryptographic material with sufficient length a key of 32 bytes is derived that can be later truncated if needed:¶
DTLS PSK = KDF(MSK, "DTLS PSK", length).¶
where:¶
For a IoT device to act as a trustworthy entity within a security domain, certain key material is needed to be shared between the IoT device and the Controller.¶
Next, we elaborate on examples of different use case scenarios about the usage of CoAP-EAP. Generally, we are dealing with 4 entities:¶
Generally, any IoT device wanting to join the domain managed by the Controller MUST perform a CoAP-EAP authentication with the Controller (C). This authentication MAY involve an external AAA server. This means that A and B, once deployed, will run CoAP-EAP once, as a bootstrapping phase, to establish a security association with C. Moreover, any other entity, which wants to join and establish communications with nodes under C's domain must also do the same. By using EAP, we can have the flexibility of having different types of credentials. For instance, if we have a device that is not battery dependent, and not very constrained, we could use a heavier authentication method. With varied IoT devices and networks we might need to resort to more lightweight authentication methods (e.g., EAP-NOOB[I-D.ietf-emu-eap-noob], EAP-AKA'[RFC5448], EAP-PSK[RFC4764], EAP-EDHOC[I-D.ingles-eap-edhoc], etc.) being able to adapt to different types of devices according to organization policies or devices capabilities.¶
In ACE, the process of Client registration and provisioning of credentials to the client is not specified. The process of Client registration and provisioning can be achieved using CoAP-EAP. Once the process of authentication with EAP is completed, fresh key material is shared between the IoT device and the Controller. In this instance, the Controller and the Authorization Server (AS) of ACE can be co-located.¶
Next, we exemplify how CoAP-EAP can be used to perform the Client registration in a general way, to allow two IoT devices (A and B) to communicate and interact after a successful client registration.¶
Node A wants to communicate with node B (e.g. to activate a light switch). The overall process is divided into three phases. Let's start with node A. In the first phase, the node A (EAP peer) does not yet belong to Controller C's domain. Then, it communicates with C (EAP authenticator) and authenticates with CoAP-EAP, which, optionally, communicates with the AAA server to complete the authentication process. If the authentication is successful, a fresh MSK is shared between C and node A. This key material allows node A to establish a security association with the C. Some authorization information may be also provided in this step. In case EAP is used in standalone mode, the AS itself having information about the devices can be the entity providing said authorization information. If authentication and authorization are correct, node A is enrolled in controller C's domain for a period of time. In particular, [RFC5247] recommends 8 hours, though the the entity providing the authorization information can establish this lifetime. In the same manner, B needs to perform the same process with CoAP-EAP to be part of the controller C's domain.¶
In the second phase, when node A wants to talk with node B, it contacts controller C for authorization to access node B and obtain all the required information to do that securely (e.g. keys, tokens, authorization information, etc.). This phase does NOT require the usage of CoAP-EAP. The details of this phase are out of the scope of this document, and the ACE framework is used for this purpose [I-D.ietf-ace-oauth-authz].¶
In the third phase, the node A can access node B with the credentials and information obtained from the controller C in the second phase. This access can be repeated without contacting the controller, while the credentials given to A are still valid. The details of this phase are out of scope of this document.¶
It is worth noting that first phase with CoAP-EAP is required to join the controller C's domain. Once it is performed with success, the communications are local to the controller C's domain and there is no need to perform a new EAP authentication as long as the key material is still valid. When the keys are about to expire, the IoT device can engage in a reauthentication as explained in Section 3.3, to renew the key material.¶
We assume we have a device (A) of the domain acme.org, which uses a specific kind of credential (e.g., AKA) and intends to join the um.es domain. This user does not belong to this domain, for which first it performs a client registration using CoAP-EAP. For this, it interacts with the controller's domain acting as EAP authenticator, which in turn communicates with a AAA infrastructure (acting as AAA client). Through the local AAA server to communicate with the home AAA server to complete the authentication and integrate the device as a trustworthy entity into the domain of controller C. In this scenario, the AS under the role of the Controller receives the key material from the AAA infrastructure¶
A University Campus, we have several Faculty buildings and each one has its own criteria or policies in place to manage IoT devices under an AS. All buildings belong to the same domain (e.g., um.es). All these buildings are managed with a AAA infrastructure. A new device (A) with credentials from the domain (e.g., um.es) will be able to perform the device registration with a Controller (C) of any building as long as they are managed by the same general domain.¶
In another case, without a AAA infrastructure, we have a Controller that has co-located the EAP server and using EAP standalone mode we can manage all the devices within the same domain locally. Client registration of a node (A) with Controller (C) can also be performed in the same manner.¶
One of the first steps for an IoT device life-cycle is to perform the authentication to gain access to the network. To do so, the device first has to be authenticated and granted authorization to gain access to the network. Additionally, security parameters such as credentials can be derived from the authentication process allowing the trustworthy operation of the IoT device in a particular network by joining the security domain. By using EAP, we are able to achieve this with flexibility and scalability, because of the different EAP methods available and the ability to rely on AAA infrastructures if needed to support multi-domain scenarios, which is a key feature when the IoT devices deployed under the same security domain, belong to different organizations. Given that EAP is also used for network access control, we can adapt this service for other technologies. For instance, to provide network access control to very constrained technologies (e.g., LoRa network). Authors in [lo-coap-eap] provide an study of a minimal version of CoAP-EAP for LPWAN networks with interesting results. In this specific case, we could leverage the compression by SCHC for CoAP [RFC8824].¶
It is not uncommon that the infrastructure where the device is deployed and the services of the IoT device are managed by different organizations. Therefore, in addition to the authentication for network access control, we have to consider the possibility of a secondary authentication to access different services. This process of authentication, for example, will provide with the necessary key material to establish a secure channel and interact with the entity in charge of granting access to different services. In 5G, for example, consider a primary and secondary authentication using EAP [TS133.501].¶