keycloak-scim/docbook/auth-server-docs/reference/en/en-US/modules/multi-tenancy.xml

74 lines
4 KiB
XML
Raw Normal View History

2016-02-03 10:20:22 +00:00
<!--
~ Copyright 2016 Red Hat, Inc. and/or its affiliates
~ and other contributors as indicated by the @author tags.
~
~ Licensed under the Apache License, Version 2.0 (the "License");
~ you may not use this file except in compliance with the License.
~ You may obtain a copy of the License at
~
~ http://www.apache.org/licenses/LICENSE-2.0
~
~ Unless required by applicable law or agreed to in writing, software
~ distributed under the License is distributed on an "AS IS" BASIS,
~ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
~ See the License for the specific language governing permissions and
~ limitations under the License.
-->
<section id="multi_tenancy">
<title>Multi Tenancy</title>
<para>
Multi Tenancy, in our context, means that one single target application (WAR) can be secured by a single (or clustered) Keycloak server, authenticating
its users against different realms. In practice, this means that one application needs to use different <literal>keycloak.json</literal> files.
For this case, there are two possible solutions:
<itemizedlist>
<listitem>
The same WAR file deployed under two different names, each with its own Keycloak configuration (probably via the Keycloak Subsystem).
This scenario is suitable when the number of realms is known in advance or when there's a dynamic provision of application instances.
One example would be a service provider that dynamically creates servers/deployments for their clients, like a PaaS.
</listitem>
<listitem>
A WAR file deployed once (possibly in a cluster), that decides which realm to authenticate against based on the request parameters.
This scenario is suitable when there are an undefined number of realms. One example would be a SaaS provider that have only one deployment
(perhaps in a cluster) serving several companies, differentiating between clients based on the hostname
(<literal>client1.acme.com</literal>, <literal>client2.acme.com</literal>) or path (<literal>/app/client1/</literal>,
<literal>/app/client2/</literal>) or even via a special HTTP Header.
</listitem>
</itemizedlist>
This chapter of the reference guide focus on this second scenario.
</para>
<para>
Keycloak provides an extension point for applications that need to evaluate the realm on a request basis. During the authentication
and authorization phase of the incoming request, Keycloak queries the application via this extension point and expects the application
to return a complete representation of the realm. With this, Keycloak then proceeds the authentication and authorization process,
accepting or refusing the request based on the incoming credentials and on the returned realm.
For this scenario, an application needs to:
<itemizedlist>
<listitem>
Add a context parameter to the <literal>web.xml</literal>, named <literal>keycloak.config.resolver</literal>.
The value of this property should be the fully qualified name of the class extending
<literal>org.keycloak.adapters.KeycloakConfigResolver</literal>.
</listitem>
<listitem>
A concrete implementation of <literal>org.keycloak.adapters.KeycloakConfigResolver</literal>. Keycloak will call the
<literal>resolve(org.keycloak.adapters.spi.HttpFacade.Request)</literal> method and expects a complete
<literal>org.keycloak.adapters.KeycloakDeployment</literal> in response. Note that Keycloak will call this for every request,
so, take the usual performance precautions.
</listitem>
</itemizedlist>
</para>
<para>
An implementation of this feature can be found in the examples.
</para>
</section>