Enterprise Single Sign-On for All

Delegate Authentication

CAS can act as a client using the pac4j security engine and delegate the authentication to:

  • A CAS server
  • A SAML identity provider
  • An OAuth2 provider: Facebook, Twitter, Google, LinkedIn, Yahoo and several other providers.
  • An OpenID provider
  • An OpenID Connect identity provider

Support is enabled by including the following dependency in the WAR overlay:


The client issuing the authentication request can be of any type (SAML, OAuth2, OpenID Connect, etc) and is allowed to submit the authentication request using any protocol that the CAS server supports and is configured to understand. This means that you may have an OAuth2 client using CAS in delegation mode to authenticate at an external SAML2 identity provider, another CAS server or Facebook and in the end of that flow receiving an OAuth2 user profile. The CAS server is able to act as a proxy, doing the protocol translation in the middle.

Register Providers

An identity provider is a server which can authenticate users (like Google, Yahoo…) instead of a CAS server. If you want to delegate the CAS authentication to Twitter for example, you have to add an OAuth client for the Twitter provider, which will be done automatically for you once provider settings are taught to CAS.

To see the relevant list of CAS properties, please review this guide.

Notice that for each OAuth provider, the CAS server is considered as an OAuth client and therefore should be declared as an OAuth client at the OAuth provider. After the declaration, a key and a secret is given by the OAuth provider which has to be defined in the CAS configuration as well.

User Interface

All available clients are automatically displayed on the login page as clickable buttons. CAS does allow options for auto-redirection of the authentication flow to a provider, if only there is a single provider available and configured.

Authenticated User Id

After a successful delegated authentication, a user is created inside the CAS server with a specific identifier: this one can be created only from the technical identifier received from the external identity provider (like 1234) or as a “typed identifier” (like FacebookProfile#1234), which is the default.

To see the relevant list of CAS properties, please review this guide.

Returned Payload

Once you have configured (see information above) your CAS server to act as an OAuth, CAS, OpenID (Connect) or SAML client, users will be able to authenticate at a OAuth/CAS/OpenID/SAML provider (like Facebook) instead of authenticating directly inside the CAS server.

In the CAS server, after this kind of delegated authentication, users have specific authentication data.

The Authentication object has:

  • The attribute AuthenticationManager.AUTHENTICATION_METHOD_ATTRIBUTE set to org.apereo.cas.support.pac4j.authentication.handler.support.ClientAuthenticationHandler
  • The attribute clientName set to the type of the provider used during authentication process.

The Principal object of the Authentication object has:

  • An identifier which is the profile type + # + the identifier of the user for this provider (i.e FacebookProfile#0000000001)
  • Attributes populated by the data retrieved from the provider (first name, last name, birthdate…)

Profile Attributes

In CAS-protected applications, through service ticket validation, user information are pushed to the CAS client and therefore to the application itself.

The identifier of the user is always pushed to the CAS client. For user attributes, it involves both the configuration at the server and the way of validating service tickets.

On CAS server side, to push attributes to the CAS client, it should be configured in the expected service:

  "@class" : "org.apereo.cas.services.RegexRegisteredService",
  "serviceId" : "sample",
  "name" : "sample",
  "id" : 100,
  "description" : "sample",
  "attributeReleasePolicy" : {
    "@class" : "org.apereo.cas.services.ReturnAllowedAttributeReleasePolicy",
    "allowedAttributes" : [ "java.util.ArrayList", [ "name", "first_name", "middle_name" ] ]

Server Configuration

Remember that this feature requires session affinity (and optionally session replication) in clustered deployments, as the requests and responses throughout the login flow are stored via server-backed session storage mechanisms.

You will need to configure your deployment environment and load-balancers accordingly.


To enable additional logging, configure the log4j configuration file to add the following levels:

<AsyncLogger name="org.pac4j" level="debug" additivity="false">
    <AppenderRef ref="console"/>
    <AppenderRef ref="file"/>