OpenID Connect
SanteDB supports OpenID Connect on the https://servername:8443/auth endpoint.
Discovery / Configuration
The configuration of the SanteDB server can be obtained via a GET to the openid-configuration endpoint, as illustrated below.
This results in an OpenID configuration response, illustrated in below.
For more information see OpenID Connect Discovery documentation.
Grant Types
SanteDB supports four different grant types described in this session. The grant types and flows are not described here. For a good reference on OpenID Connect flows see RFC67849.
You can limit the types of grants each device, application, and security role are allowed to authenticate by denying the following policies on those objects:
Policy
Grant
Description
OAUTH Login (1.3.6.1.4.1.33349.3.1.5.9.2.1.0.0)
Any
When denied, the specified object cannot authenticate using OAUTH or OpenID Connect.
OAUTH password flow (1.3.6.1.4.1.33349.3.1.5.9.2.1.0.0.2)
Password
When denied, the specified object is prohibited from using the password grant flow.
OAUTH client_credentials flow (1.3.6.1.4.1.33349.3.1.5.9.2.1.0.0.1)
Client Credentials
When denied, the specified object is prohibited from using the client_credentials grant flow.
OAUTH authorization code flow (1.3.6.1.4.1.33349.3.1.5.9.2.1.0.0.3)
Authorization Code
When denied, the specified object is prohibited from using the authorization_code grant flow.
OAUTH password reset flow (1.3.6.1.4.1.33349.3.1.5.9.2.1.0.0.4)
x_challenge
When denied, the specified object is prohibited from using the extended TFA challenge authentication for password resets.
Password Grant
The password grant is a grant whereby a trusted user agent (mobile app, website, etc.) collects both the username and password for the user which are then sent to the IdP server.
This grant is useful for:
Authenticating users from an offline application context (periodic authentication where the UA collects passwords anyways)
Authenticating users from a first party application or trusted application
The process is as follows:
User Agent uses OIDC discovery to determine appropriate scopes, grants, and configuration of the remote target.
User agent collects username and password (not specified how)
User agent requests an access token from the IdP
The IdP responds with an id_token, auth_token, refresh_token for subsequent use.
The user agent continues to access the protected resource.
The request for a token using the password is illustrated below:
The server responds with a token response (documented below).
Client Credentials Grant
The client credentials grant creates an application principal based on the device/application pair presented on the authentication request.
This grant is useful for:
A background task from a trusted application on a trusted device (like synchronization)
HIE traffic between nodes / applications where dual-PKI infrastructure is prohibitively complex.
The process for this is:
User agent runs a discovery request (optional)
User agent sends a token request (client_credentials grant)
User agent receives an application authentication token
User agent accesses the resource
A client credentials grant is illustrated below:
SanteDB's IdP mandates that application principals can only be created if the device is known. The device identity can be established from client-certificates (if configured), or in a simpler deployment, using the device's identifier and secret in the X-Device-Authorization header.
Authorization Code Grant
The authorization code grant is a traditional OpenID connect grant flow you've probably seen used online. The authorization code grant does not expose the user name or password to the client application. Rather, the user is redirected to the SanteDB IdP server where their credentials are collected by SanteDB and the resulting authorization code is sent back to the UA.
This grant is helpful when:
Integrating third party applications where you may or may not trust the application
SSO where you wish to use cookies to auto-login users.
The process for this is:
User is directed to the login screen on SanteDB's IdP (established from OIDC discovery)
User enters their username and password into the IdP
IdP generates an authorization code and redirects the user's browser back to the application requesting the token
Application uses the authorization code provided to exchange for auth tokens
IdP validates the authorization code and responds with token.
Application accesses the protected resource.
In this flow the first step is a redirection to the IdP supplying the minimum parameters:
The parameters for this initial request are:
Parameter
Description
Values
redirect_uri
The URI where the IdP should redirect the user's browser
client_id
The client identification of the requestor (note: must have grant
to do authorization code flow)
scope
The scope of the grant . Must contain openid, optionally can contain additional policies being requested)
openid
claim
Claims made about the user separated
state
A unique state identifier which can be used to maintain state across requests
response_type
The type of response being asked for.
code => The redirect is supplied a code which can be exchanged
token =>
code or token
response_mode
The mode of response.
query => The redirect will be made via a GET operation with the authorization code being the supplied as a query parameter.
post => The redirect will be made via a POST operation with the authorization code being supplied as a form parameter.
form_post or query
nonce
A unique NONCE which is passed back to the requestor to ensure tampering has not occurred.
The user will be presented with the defined IdP login screen (specified in the applets manifest).
If successful, the browser will be redirected back to the caller.
When an response_type = code and response_mode = query, a simple redirect is performed
When response_type = code and response_mode = form_post, a form post is returned
If using response_type = token and response_mode = form_post, a form post with an auth token is directly sent back to the UA for use with the assigned resource. This may be disabled based on your configuration.
Identity Token Contents
The id_token of the OpenID connect token response contains a series of claims about the user, device and application used to authenticate the user. The identity token is in the JWT token (jwt.io). The token, when processed contains claims similar to those shown below:
The claims provided by the identity provider are:
Claim
Description
Example
unique_name
The user name of the user, application or device which has been authenticated.
demoadmin
role
The role(s) to which the user belongs.
ADMINISTRATORS
authmethod
The grant flow which resulted in the identity token being generated.
Password
nameid
The SID of the user for which the token was generated.
UUID
actort
Identifies the classification of the actor (human, system, etc.)
The primary e-mail address for use for security correspondence of the user.
appid
The SID of the application which is granted use of the application
scope
jti
The unique identifier of the session from the issuer.
iat
Issued at
exp
Expiration time
sub
The primary subject of the session. Note: This may be a user's SID (if the session is a password grant) or an application SID (if the session is a client_credentials grant).
iss
The issuer of the token.
nbf
The effective time of the session.
Client Claims
SanteDB's OpenID connect provides an extended header option which can be used by clients to append claims about the user they would like reflected in the session. The authentication server will validate these claims against the authenticated principal and may return either an HTTP 200 (authentication success) or an HTTP 400 (authentication fail).
Client claims are passed via the X-SanteDBClient-Claim header. For example, this particular series of claims indicates the user is breaking the glass because of emergency access.
The types of claims which may be sent to the OpenID Connect service are listed below:
Claim
Description
Example
PolicyOverride
When true or 1, indicates that the user is re-authenticating for the purpose of violating/bypassing a policy.
1
PurposeOfUse
The codified purposes of use, the reason why the policy override is being executed.
EMERG
PAT
ResourceId
When specified, indicates the UUID of a specific resource for which the user is overriding the policy system.
UUID
SubjectRole
The role which the user is playing (if the user plays multiple roles). For example, if a user is both a Physician and a Surgeon they may specify which role they are authenticating as.
SubjectOrganizationID
The organization within which the user is authenticating themselves. For example, if a user works both at Good Health Hospital and Health Clinic Inc. this would allow specification of the organization.
Language
If the user wishes to switch messages from the server to a different language, the language of messages being returned from the server.
en, fr
When performing a consent override, you may also specify the policies which are being overridden in the source scope property. For example, the Elbonia MPI defines a directive "SUPER SECRET DISCLOSURE" , if the user Allison wished to override the default consent decision, the client could send a password grant illustrated below:
Related Topics
Last updated