3 Using the ArisID API

This chapter describes the architecture and key functionality of the Identity Governance Framework ArisID API (ArisID API). The ArisID API provides enterprise developers and system architects a library for building identity-enabled applications using multiple identity protocols. The ArisID API enables developers to specify requirements for identity attributes, roles, and search filters by using Client Attribute Requirements Markup Language (CARML). This chapter contains the following topics:

Section 3.1, "About ArisID API"
Section 3.2, "Configuring ArisID API"
Section 3.3, "Design Recommendations"
Section 3.4, "Generating ArisID Beans"
Section 3.5, "Sample Application Using IDX User and Role Beans"
Section 3.6, "OpenLDAP Considerations"

3.1 About ArisID API

The Identity Governance Framework ArisID API represents a common core service through which all identity information exchanged should be passed. While not an official name, the ArisID API is often referred to as Identity Beans by developers.

The 11g Release 1 (11.1.1) release of the ArisID API is a subset of the configuration proposed at:

http://www.openliberty.org/wiki/index.php/ArisID_Configuration.

If you have installed Oracle WebLogic Server and Oracle Identity Management, all the necessary jar files for developing applications with this API are already installed on your computer.

3.2 Configuring ArisID API

The Identity Governance Framework ArisID supports the basic development process Create > Modify > Test > Deploy. Creation requires CARML XML files and modifying them to suit your environment. Testing the application can be done in Oracle WebLogic Server embedded LDAP directory server.

This section contains the following topics:

Configuring CARML Files
Configuring the Identity Repository
Configuring the Mapping File

3.2.1 Configuring CARML Files

Determine whether the existing ArisID beans meet your application's needs by examining the CARML files idxuserrole.xml (read-only operations) and userrole.xml (read-only and read/write operations). These files are located in DOMAIN_HOME/config/fmwconfig/carml.

3.2.2 Configuring the Identity Repository

The identity repository to be used by the ArisID beans must be available. You can use the Oracle WebLogic Server embedded LDAP-based directory server or any LDAP directory supported by 11g Oracle Virtual Directory. The ArisID API is integrated with Oracle Platform Security Services. It automatically connects to the LDAP-based identity store configured in Oracle Platform Security Services. For more information about the identity stores supported by Oracle Platform Security Services, see "System Requirements and Certification".

For more information about Oracle Platform Security Services, see Oracle Fusion Middleware Application Security Guide.

If you must use a different identity store from the Oracle Platform Security Services identity store, then set the following system property:

igf.ovd.config.dir=DOMAIN_HOME/config/fmwconfig/arisidprovider/conf

Next, edit the adapters.os_xml file to include the host, port and credentials of the directory to be connected to. The igf.ovd.config.dir property can be set to any other directory containing adapaters.os_xml and other configuration files with the right settings.

Directory Limitations

The following LDAP directory limitations apply:

When using OpenLDAP 2.2 or Novell eDirectory, there is no paging support. Paging APIs, for example, searchUsersbyPage(), searchRolesbyPage(), etc., do not work. There is no Multiple Language Support (MLS).
When using Oracle WebLogic Server embedded LDAP-based directory server, there is no Multiple Language Support (MLS).

3.2.3 Configuring the Mapping File

When a CARML file is created a corresponding mapping file is created in the same location. The default mapping file has attribute details specific to Oracle WebLogic Server embedded directory server, which is the Oracle Platform Security Services default identity store. If you are using a default CARML file and the Oracle Platform Security Services identity store, you do not need to configure mapping. The configuration parameters in Oracle Platform Security Services override the parameters in the mapping file.

3.3 Design Recommendations

The default CARML and mapping files make certain assumptions about the deployment scenario. You may need to modify these details depending on your deployment requirements. The configuration parameters that can be modified are discussed in this section.

This section contains the following topics:

Choose a LoginID
Choose a UniqueKey
Specify Multiple Language Support
Handle Large Results
Secure the Application
Configure the Timeout Interval

3.3.1 Choose a LoginID

In the default configuration, email is used as a unique identifier for identifying user entries. When you are searching for a user, the default attribute expected for search is email. For example:

SearchUser( String uniqueid,  Map<String, Object>)

For performance reasons, the attribute used as a unique identifier must be a searchable attribute in the backend. The mapping between the application's choice of uniquekey and the backend attribute is handled at configuration time. This is a configuration in Oracle Virtual Directory mapping. The HashMap is used to provide the optional context information to be used while performing the operation. In the current release it supports the following options:

The Principal user that performs the search - (ArisIdConstants.APP_CTX_AUTHUSER, (Principal)user)
The language constraint if any - (ArisIdConstants.APP_CTX_LOCALE, "fr")
Pagination support if any - (ArisIdConstants.APP_CTX_PAGESIZE, 10)

3.3.2 Choose a UniqueKey

An application occasionally stores the entries accessed from the identity repository's backend in their own application-specific repository. In such cases, you must carefully consider which attribute should be persisted. For instance, if the backend is an LDAP-based repository, you should use the GUID attribute as the persisting attribute because this is the only unique key on the LDAP-based backend. All other LDAP attributes are modifiable.

If the backend is a relational database, choose an attribute on which uniqueness constraint is enforced as the unique key. You can specify this in the ArisID mapping property file. The method to search for a user based on the unique key is:

searchUserOnUniqueKey(String UniqueKey, Map<String,Object>)

The HashMap is used to provide the optional context information to be used while performing the operation. In the current release it supports the following options:

The Principal user that performs the search - (ArisIdConstants.APP_CTX_AUTHUSER, (Principal)user)
The language constraint if any - (ArisIdConstants.APP_CTX_LOCALE, "fr")
Pagination support if any - (ArisIdConstants.APP_CTX_PAGESIZE, "10")

3.3.3 Specify Multiple Language Support

Multiple Language Support (MLS) is provided for applications that need locale-specific results. The attributes and the appropriate MLS code are stored in the ArisID properties file in the multiLanguageAttributes element.

<multiLanguageAttributes>…</multiLanguageAttribute>

Because displayname is the most commonly used multiple language attribute, it is configured by default as a multi-language attribute. Other attributes can be added as needed in the ArisID mapping file.

Restrictions

Any API to which locale is specified as an argument will return the locale-specific values for all the attributes listed in the ArisID properties file as <multiLanguageAttributes> that have locale-specific values. For all other attributes it returns the default values stored.

In the backend system, the data is returned in a form conforming to ISO-3166. For example, if there is a French locale (in addition to English), it is stored as cn,:fr for the cn attribute. The locale for the client applications should be specified in the properties HashMap as ArisIdConstants.APP_CTX_LOCALE, "fr" and the ArisID properties file should contain cn as multiLanguageAttribute and map this attribute.

3.3.4 Handle Large Results

When applications access identity data, the result set for a search is frequently too large to be handled by the application. In such cases you have the option of dividing the result into manageable sized pages. You do this by defining the number of objects to be returned in the page.

The following example shows a typical usage pattern:

RoleManager rm = new RoleManager(env); 
  List<PropertyFilterValue> attrFilters = new  ArrayList<PropertyFilterValue>(); 
  attrFilters.add(new PropertyFilterValue(Role.NAME, "admin", AttributeFilter.OP_CONTAINS)); 

  HashMap<String,Object> map = new HashMap<String,Object>(); 
  map.put("ArisIdConstants.APP_CTX_PAGESIZE","2"); 
  SearchResults<Role> sr = rm.searchRolesbyPage(attrFilters, map); 

  while(sr.hasMore()) 
  { 
     List<Role> roles = sr.getNextSet(); 

     for (int i=0; i<roles.size(); i++) 
       //do the operations with roles.get(i)
}

3.3.5 Secure the Application

Two security scenarios are available for executing create, read, update, and delete (CRUD) operations on the target system. They are:

Domain level credentials
Application level credentials

Proxy authentication is not supported in this release.

3.3.5.1 Domain Level Credentials

In this scenario, all applications in a domain use common credentials to connect to the target system and perform operations with those credentials. The application does not maintain a footprint in the target system.

The LDAP Adapter's configuration file, adapters.os_xml, contains credentials to connect to the backend directory, along with the host and port details. If you do not provide any other credentials during initialization, the application connects to the target system using the credentials in the LDAP Adapter's configuration file.

If proxy user (logged in user id) is not specified in the API's application context, ArisID operation will be executed with the credentials that are in LDAP Adapter's configuration file.

If your application connects using common credentials, you must build security into the application itself so that it displays or modifies data only for an authorized user.

Example 3-1 Code Sample: adapters.os_xml

The LDAP Adapter's configuration file adapters.os_xml is configured with domain level userid and encrypted password to connect to backend directory. The following is a snippet of adapters.os_xml.

    <binddn>cn=admin</binddn>
    <bindpass>{OMASK}C2QXW1Nmf+s=</bindpass>

While initializing the ArisID API do not provide any credentials.

Map env = new HashMap();
// Do not set UserManager.SECURITY_PRINCIPAL & SECURITY_CREDENTIALS
UserManager uMgr  = new UserManager(env);
…
…
// Search Operation (with no proxy user in app context)
List<PropertyFilterValue> attrFilters = new ArrayList<PropertyFilterValue>();
attrFilters.add(new PropertyFilterValue("User.FIRSTNAME", "app1", AttributeFilter.OP_CONTAINS));
attrFilters.add(new PropertyFilterValue("User.LASTNAME", "user1", AttributeFilter.OP_BGNSWITH));
Map<String, Object> appCtx = null;
users = um.searchUsers(attrFilters, appCtx);

3.3.5.2 Application Level Credentials

In this scenario, each application uses application level credentials to connect to the target system and performs CRUD operations with those credentials.

In this case you provide the application's user id and password while initializing the ArisID API. When you do that, the application connects to the target system using those credentials.

If no proxy user is specified in the API's application context then ArisID operation will be executed with the application's credentials.

This scenario has the following features:

Each application has different privileges to view and update the data in the target system
You can audit the modifications performed by each application in the target system

Example 3-2 Code Sample: adapters.os_xml

The LDAP Adapter's configuration file adapters.os_xml is configured with domain level userid and encrypted password to connect to backend directory. The following is a snippet of adapters.os_xml.

<binddn>cn=admin</binddn>
    <bindpass>{OMASK}C2QXW1Nmf+s=</bindpass>

While initializing the ArisID API, provide the application user credentials.

Map env = new HashMap();
env.put(UserManager.SECURITY_PRINCIPAL, "cn=app1_user,cn=users,dc=example,dc=com");
env.put(UserManager.SECURITY_CREDENTIALS, "mypassword");
UserManager uMgr  = new UserManager(env);
…
// Search Operation (with no proxy user in app context)
List<PropertyFilterValue> attrFilters = new ArrayList<PropertyFilterValue>();
attrFilters.add(new PropertyFilterValue("User.FIRSTNAME", "app1", AttributeFilter.OP_CONTAINS));
attrFilters.add(new PropertyFilterValue("User.LASTNAME", "user1", AttributeFilter.OP_BGNSWITH));Map<String, Object> appCtx = null;
users = um.searchUsers(attrFilters, appCtx);

3.3.6 Configure the Timeout Interval

The default connect/read timout is configured to be 15 seconds. For example, if the LDAP operation on IdentityStore is taking more than 15 seconds, the operation will timeout and throw the following IGF exception:

org.openliberty.arisid.stack.ConnectionException

If the IdentityStore has a many entries and the applications are using filters with "contains" and search with paging/sorting, those queries may timeout.

The recommendation is to set the timeout value to 0 (no timeout) and increase the pool size to 20. If the preference is for the application has some timeout interval, set the value to greater than 0.

To configure the timeout interval:

Run the following WLST command to list all adapters.
```
listAdapters() 
```
Run the following command for each adapter to set timeout and maxpoolsize.
1. modifyLDAPAdapter('<ADAPTER NAME>', 'OperationTimeout', 0)
2. modifyLDAPAdapter('<ADAPTER NAME>', 'MaxPoolSize', 20)
Restart WebLogic Server.

3.4 Generating ArisID Beans

Generate ArisID beans as follows:

java BeanGenerator [-genmap] <package name> <output dir> [<relationship file>] <carml file>

where:

-genmap: Generates the CARML-ArisIDStack map file in the same location where the CARML file is.
package name: Is the package name for the generated bean classes.
output dir: Location to write the generated bean classes.
relationship file: File containing the relationship definitions between entities.
carml file: The CARML declaration file.

To build ORG beans from a CARML file:

Create a CARML file name org.xml with the appropriate attributes/interactions for ORG entity.

Generate org beans (OrgManager.java and Org.java) using the Bean Generator. The build.xml file should resemble the following sample:

<path id="ArisIDBeans.classpath">
            <pathelement location="MW_HOME/oracle_common/modules/velocity-dep-1.4.jar" />
            <pathelement location="MW_HOME/oracle_common/modules/oracle.jrf_11.1.1/jrf.jar" />
    </path>
    <property name="BeanGeneratorClassPath" refid="ArisIDBeans.classpath"/>
    <target name="generatebeans" description="generate arisid beans">
        <java classname="org.openliberty.arisidbeans.BeanGenerator" dir="${generatedsource.dir}" fork="true">
            <arg value="${generatedbean.userrole.packagename}"/>
            <arg value="."/>
            <arg value="${carml.dir}/org.xml"/>
            <classpath>
                <pathelement path="${BeanGeneratorClassPath}"/>
            </classpath>
            <sysproperty key="org.openliberty.arisid.policy.wspolicy.class"
                         value="org.openliberty.arisid.policy.neethi.PolicyImpl" />
        </java>
    </target>

Compile the generated java files Org.java and OrgManager.java.
Edit the generated mapping file (igf-map-config-.xml) to update basesearch, objectclass, OVD attribute names with their values.
An application can use the generated ORG APIs for the interactions defined in the CARML file (org.xml). After the application is deployed in application server:
1. Copy the mapping file under DOMAIN_HOME/config/fmwconfig/arisidprovider/conf, and
2. Copy the CARML file under DOMAIN_HOME/config/fmwconfig/carml

3.5 Sample Application Using IDX User and Role Beans

The following sample application uses IDX User/Role Beans.

3.5.1 SearchUsers.jsp

<%@ page language="java" contentType="text/html; charset=ISO-8859-1" pageEncoding="ISO-8859-1"%>
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<%@page import="org.openliberty.arisid.*"%>
<%@page import="org.openliberty.arisidbeans.*"%>
<%@page import="oracle.igf.userrole.*"%>
<%@page import="java.util.*"%>
<%@page import="java.net.URI"%>
<%!public static UserManager uMgr = null;
{
        try {
                uMgr = new UserManager(null);
        } catch (Exception e) {
                e.printStackTrace();
        }
 
}
%>
<html>
<head>
<title>Search Users</title>
<%
 
 
String firstname = request.getParameter("firstname");
String lastname = request.getParameter("lastname");
String telephone = request.getParameter("telephone");
 
 
List<PropertyFilterValue> attrFilters = new ArrayList<PropertyFilterValue>();
attrFilters.add(new PropertyFilterValue("firstname", firstname, AttributeFilter.OP_BGNSWITH));
attrFilters.add(new PropertyFilterValue("lastname", lastname, AttributeFilter.OP_BGNSWITH));
attrFilters.add(new PropertyFilterValue("telephone", telephone, AttributeFilter.OP_CONTAINS));
 
List<User> subjs = uMgr.searchUsers(attrFilters);
 
%>
</head>
<body>
 
<a href="SearchUsers.html">Home</a>
<center>List of Users with FirstName starting with "<%=firstname%>", LastName
starting with "<%=lastname%>" and TelephoneNumber containing
"<%=telephone%>"</center>
 
<%
Iterator<User> sIter = subjs.iterator();
while (sIter.hasNext()) {
        User subj = sIter.next();
 
        Map<String, IAttributeValue> vals = subj.getAllAttributes();
        Iterator<IAttributeValue> iter = vals.values().iterator();
%>
<table border="0">
        <tr>
                <th>Item</th>
                <th>Value</th>
        </tr>
        <%
                while (iter.hasNext()) {
                        IAttributeValue val = iter.next();
                        String name = val.getNameIdRef();
                        String value = null;
                        if (val.size() > 0)
                                value = val.get(0);
if (value != null)
{
        %>
        <tr>
                <td><%=name%></td>
                <td><%=value%></td>
        </tr>
        <%
}
                }
        %>
</table>
<%
        }
%>
<br>
<br>
<br>
<a href="SearchUsers.html">Home</a>
</body>
</html>

3.5.2 SearchUsers.html

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML>
<HEAD><TITLE>Search Users</TITLE></HEAD>
<BODY>
<FORM METHOD=POST ACTION="SearchUsers.jsp">
 
First Name Starting with <INPUT TYPE=TEXT NAME=firstname SIZE=30><BR><BR>
Last Name Starting with <INPUT TYPE=TEXT NAME=lastname SIZE=30><BR><BR>
Telephone Number containing <INPUT TYPE=TEXT NAME=telephone SIZE=15><BR><BR>
<P><INPUT TYPE=SUBMIT>
</FORM>
</BODY>
</HTML>

3.6 OpenLDAP Considerations

For OpenLDAP, Role.MEMBER is a mandatory attribute for the following APIs:

createRole(List<PropertyValue> attrVals, Map<String,Object> appCtxMap)
createRole(List<PropertyValue> attrVals)

If the Role.MEMBER is not included in the input attrVals list, role creation will fail.