Identity for WebRTC 1.0

2. Conformance

As well as sections marked as non-normative, all authoring guidelines, diagrams, examples, and notes in this specification are non-normative. Everything else in this specification is normative.

The key words MAY, MUST, MUST NOT, and SHOULD are to be interpreted as described in [RFC2119].

This specification defines conformance criteria that apply to a single product: the user agent that implements the interfaces that it contains.

Conformance requirements phrased as algorithms or specific steps may be implemented in any manner, so long as the end result is equivalent. (In particular, the algorithms defined in this specification are intended to be easy to follow, and not intended to be performant.)

Implementations that use ECMAScript to implement the APIs defined in this specification MUST implement them in a manner consistent with the ECMAScript Bindings defined in the Web IDL specification [WEBIDL-1], as this specification uses that specification and terminology.

3. Terminology

The EventHandler interface, representing a callback used for event handlers, and the ErrorEvent interface are defined in [HTML51].

The concepts queue a task, fire a simple event and networking task source are defined in [HTML51].

The terms event, event handlers and event handler event types are defined in [HTML51].

performance.timeOrigin and performance.now() are defined in [HIGHRES-TIME].

The terms MediaStream, MediaStreamTrack, and MediaStreamConstraints are defined in [GETUSERMEDIA]. Note that MediaStreamTrack is extended in the MediaStreamTrack section in this document.

The term Blob is defined in [FILEAPI].

The term media description is defined in [RFC4566].

The term media transport is defined in [RFC7656].

The term generation is defined in [TRICKLE-ICE] Section 2.

The terms RTCStatsType, stats object and monitored object are defined in [WEBRTC-STATS].

When referring to exceptions, the terms throw and create are defined in [WEBIDL-1].

The term "throw" is used as specified in [INFRA]: it terminates the current processing steps.

The terms fulfilled, rejected, resolved, pending and settled used in the context of Promises are defined in [ECMASCRIPT-6.0].

The terms bundle, bundle-only and bundle-policy are defined in [JSEP].

The OAuth Client and Authorization Server roles are defined in [RFC6749] Section 1.1.

The terms RTCPeerConnection, target peer identity, setRemoteDescription, createOffer, createAnswer, addTrack and RTCError are defined in [WEBRTC].

4. Identity Provider Interaction

WebRTC offers and answers (and hence the channels established by RTCPeerConnection objects) can be authenticated by using a web-based Identity Provider (IdP). The idea is that the entity sending an offer or answer acts as the Authenticating Party (AP) and obtains an identity assertion from the IdP which it attaches to the session description. The consumer of the session description (i.e., the RTCPeerConnection on which setRemoteDescription is called) acts as the Relying Party (RP) and verifies the assertion.

The interaction with the IdP is designed to decouple the browser from any particular identity provider; the browser need only know how to load the IdP's JavaScript, the location of which is determined by the IdP's identity, and the generic interface to generating and validating assertions. The IdP provides whatever logic is necessary to bridge the generic protocol to the IdP's specific requirements. Thus, a single browser can support any number of identity protocols, including being forward compatible with IdPs which did not exist at the time the browser was written.

[Global,
 Exposed=RTCIdentityProviderGlobalScope]
interface RTCIdentityProviderGlobalScope : WorkerGlobalScope {
    readonly attribute RTCIdentityProviderRegistrar rtcIdentityProvider;
};

5. Registering an IdP Proxy

An IdP proxy implements the RTCIdentityProvider methods, which are the means by which the user agent is able to request that an identity assertion be generated or validated.

Once instantiated, the IdP script is executed. The IdP MUST call the register() function on the RTCIdentityProviderRegistrar instance during script execution. If an IdP is not registered during this script execution, the user agent cannot use the IdP proxy and MUST fail any future attempt to interact with the IdP.

[Exposed=RTCIdentityProviderGlobalScope]
interface RTCIdentityProviderRegistrar {
    void register(RTCIdentityProvider idp);
};

dictionary RTCIdentityProvider {
    required GenerateAssertionCallback generateAssertion;
    required ValidateAssertionCallback validateAssertion;
};

callback GenerateAssertionCallback = Promise<RTCIdentityAssertionResult> (DOMString contents,
                                                                          DOMString origin,
                                                                          RTCIdentityProviderOptions options);

callback ValidateAssertionCallback = Promise<RTCIdentityValidationResult> (DOMString assertion,
                                                                           DOMString origin);

dictionary RTCIdentityAssertionResult {
    required RTCIdentityProviderDetails idp;
    required DOMString                  assertion;
};

dictionary RTCIdentityProviderDetails {
    required DOMString domain;
             DOMString protocol = "default";
};

dictionary RTCIdentityValidationResult {
    required DOMString identity;
    required DOMString contents;
};

6. Requesting Identity Assertions

The identity assertion request process is triggered by a call to createOffer, createAnswer, or getIdentityAssertion. When these calls are invoked and an identity provider has been set, the following steps are executed:

1. The RTCPeerConnection instantiates an IdP as described in Identity Provider Selection and Registering an IdP Proxy. If the IdP cannot be loaded, instantiated, or the IdP proxy is not registered, this process fails.

2. If the RTCPeerConnection was not constructed with a set of certificates, and one has not yet been generated, wait for it to be generated.

3. The RTCPeerConnection invokes the generateAssertion method on the RTCIdentityProvider methods registered by the IdP.

   The RTCPeerConnection generates the contents parameter to this method as described in [RTCWEB-SECURITY-ARCH]. The value of contents includes the fingerprint of the certificate that was selected or generated during the construction of the RTCPeerConnection. The origin parameter contains the origin of the script that calls the RTCPeerConnection method that triggers this behavior. The usernameHint value is the same value that is provided to setIdentityProvider, if any such value was provided.

4. The IdP proxy returns a Promise to the RTCPeerConnection. The IdP proxy is expected to generate the identity assertion asynchronously.

   If the user has been authenticated by the IdP, and the IdP is able to generate an identity assertion, the IdP resolves the promise with an identity assertion in the form of an RTCIdentityAssertionResult.

   This step depends entirely on the IdP. The methods by which an IdP authenticates users or generates assertions is not specified, though they could involve interacting with the IdP server or other servers.

5. If the IdP proxy produces an error or returns a promise that does not resolve to a valid RTCIdentityAssertionResult (see 8. IdP Error Handling), then assertion generation fails.

6. The RTCPeerConnection MAY store the identity assertion for use with future offers or answers. If a fresh identity assertion is needed for any reason, applications can create a new RTCPeerConnection.

7. If the identity request was triggered by a createOffer() or createAnswer(), then the assertion is converted to a JSON string, base64-encoded and inserted into an a=identity attribute in the session description.

If assertion generation fails, then the promise for the corresponding function call is rejected with a newly created OperationError.

7. Verifying Identity Assertions

Identity assertion validation happens when setRemoteDescription is invoked on RTCPeerConnection. The process runs asynchronously, meaning that validation of an identity assertion might not block the completion of setRemoteDescription.

The identity assertion request process involves the following asynchronous steps:

1. The RTCPeerConnection awaits any prior identity validation. Only one identity validation can run at a time for an RTCPeerConnection. This can happen because the resolution of setRemoteDescription is not blocked by identity validation unless there is a target peer identity.

2. The RTCPeerConnection loads the identity assertion from the session description and decodes the base64 value, then parses the resulting JSON. The idp parameter of the resulting dictionary contains a domain and an optional protocol value that identifies the IdP, as described in [RTCWEB-SECURITY-ARCH].

3. If the identity assertion is malformed, or if protocol includes the character '/' or '\', this process fails.

4. The RTCPeerConnection instantiates the identified IdP as described in 4.1 Identity Provider Selection and 5. Registering an IdP Proxy. If the IdP cannot be loaded, instantiated or the IdP proxy is not registered, this process fails.

5. The RTCPeerConnection invokes the validateAssertion method registered by the IdP.

   The assertion parameter is taken from the decoded identity assertion. The origin parameter contains the origin of the script that calls the RTCPeerConnection method that triggers this behavior.

6. The IdP proxy returns a promise and performs the validation process asynchronously.

   The IdP proxy verifies the identity assertion using whatever means necessary. Depending on the authentication protocol this could involve interacting with the IdP server.

7. If the IdP proxy produces an error or returns a promise that does not resolve to a valid RTCIdentityValidationResult (see 8. IdP Error Handling), then identity validation fails.

8. Once the assertion is successfully verified, the IdP proxy resolves the promise with an RTCIdentityValidationResult containing the validated identity and the original contents that are the payload of the assertion.

9. The RTCPeerConnection decodes the contents and validates that it contains a fingerprint value for every a=fingerprint attribute in the session description. This ensures that the certificate used by the remote peer for communications is covered by the identity assertion.

   Note

   A user agent is required to fail to communicate with peers that offer a certificate that doesn't match an a=fingerprint line in the negotiated session description.

   Note

   The user agent decodes contents using the format described in [RTCWEB-SECURITY-ARCH]. However the IdP MUST treat contents as opaque and return the same string to allow for future extensions.

10. The RTCPeerConnection validates that the domain portion of the identity matches the domain of the IdP as described in [RTCWEB-SECURITY-ARCH]. If this check fails then the identity validation fails.

11. The RTCPeerConnection resolves the peerIdentity attribute with a new instance of RTCIdentityAssertion that includes the IdP domain and peer identity.

12. The user agent MAY display identity information to a user in its UI. Any user identity information that is displayed in this fashion MUST use a mechanism that cannot be spoofed by content.

If identity validation fails, the peerIdentity promise is rejected with a newly created OperationError if it is not settled. Then, if there is no target peer identity, set peerIdentity to a new unresolved promise. This permits the use of renegotiation (or a subsequent answer, if the session description was a provisional answer) to resolve or reject the identity.

If identity validation fails and there is a target peer identity for the RTCPeerConnection, the promise returned by setRemoteDescription is rejected with the same DOMException.

8. IdP Error Handling

Errors in IdP processing will - in most cases - result in the failure of the procedure that invoked the IdP proxy. This will result in the rejection of the promise returned by getIdentityAssertion, createOffer, or createAnswer. An IdP proxy error causes a setRemoteDescription promise to be rejected if there is a target peer identity; IdP errors in calls to setRemoteDescription where there is no target peer identity cause the peerIdentity promise to be rejected instead.

If an error occurs these promises are rejected with an RTCError if an error occurs in interacting with the IdP proxy. The following scenarios result in errors:

• An RTCPeerConnection might be configured with an identity provider, but loading of the IdP URI fails. Any procedure that attempts to invoke such an identity provider and cannot load the URI fails with an RTCError with errorDetail set to "idp-load-failure" and the httpRequestStatusCode attribute of the error set to the HTTP status code of the response.

• If the IdP loads fails due to the TLS certificate used for the HTTPS connection not being trusted, it fails with an RTCError with errorDetail set to "idp-tls-failure". This typically happens when the IdP uses certificate pinning and an intermediary such as an enterprise firewall has intercepted the TLS connection.

• If the script loaded from the identity provider is not valid JavaScript or does not implement the correct interfaces, it causes an IdP failure with an RTCError with errorDetail set to "idp-bad-script-failure".

• An apparently valid identity provider might fail in several ways. If the IdP token has expired, then the IdP MUST fail with an RTCError with errorDetail set to "idp-token-expired". If the IdP token is not valid, then the IdP MUST fail with an RTCError with errorDetail set to "idp-token-invalid". If an identity provider throws an exception or returns a promise that is ultimately rejected, then the procedure that depends on the IdP MUST also fail. These types of errors will cause an IdP failure with an RTCError with errorDetail set to "idp-execution-failure".

• The user agent SHOULD limit the time that it allows for an IdP to 15 seconds. This includes both the loading of the IdP proxy and the identity assertion generation or validation. Failure to do so potentially causes the corresponding operation to take an indefinite amount of time. This timer can be cancelled when the IdP proxy produces a response. Expiration of this timer cases an IdP failure with an RTCError with errorDetail set to "idp-timeout".

• If the identity provider requires the user to login, the operation will fail RTCError with errorDetail set to "idp-need-login" and the idpLoginUrl attribute of the error set to the URL that can be used to login.

• Even when the IdP proxy produces a positive result, the procedure that uses this information might still fail. Additional validation of an RTCIdentityValidationResult value is still necessary. The procedure for validation of identity assertions describes additional steps that are required to successfully validate the output of the IdP proxy.

Any error generated by the IdP MAY provide additional information in the idpErrorInfo attribute. The information in this string is defined by the IdP in use.

9. RTCPeerConnection Interface Extensions

The Identity API extends the RTCPeerConnection interface as described below.

partial interface RTCPeerConnection {
    void               setIdentityProvider(DOMString provider,
                                           optional RTCIdentityProviderOptions options);
    Promise<DOMString> getIdentityAssertion();
    readonly attribute Promise<RTCIdentityAssertion> peerIdentity;
    readonly attribute DOMString?                    idpLoginUrl;
    readonly attribute DOMString?                    idpErrorInfo;
};

dictionary RTCIdentityProviderOptions {
    DOMString protocol = "default";
    DOMString usernameHint;
    DOMString peerIdentity;
};

[Constructor(DOMString idp, DOMString name),
 Exposed=Window]
interface RTCIdentityAssertion {
    attribute DOMString idp;
    attribute DOMString name;
};

10. Media Stream API Extensions for Network Use

partial dictionary MediaStreamConstraints {
    DOMString peerIdentity;
};

partial interface MediaStreamTrack {
    readonly attribute boolean      isolated;
             attribute EventHandler onisolationchange;
};

11. Identity Examples

The identity system is designed so that applications need not take any special action in order for users to generate and verify identity assertions; if a user has configured an IdP into their browser, then the browser will automatically request/generate assertions and the other side will automatically verify them and display the results. However, applications may wish to exercise tighter control over the identity system as shown by the following examples.

pc.setIdentityProvider('example.com');

pc.setIdentityProvider('example.com', {
  usernameHint: 'alice@example.com',
  peerIdentity: 'bob@example.net'
});

async function consumeIdentityAssertion() {
  const identity = await pc.peerIdentity;
  console.log('IdP = ', identity.idp, 'identity =', identity.name);
}

12. Change Log

This section will be removed before publication.

Changes since June 21, 2018

1. This document was split from the [WEBRTC] specification.
2. Editors were changed to Cullen Jennings and Martin Thomson.