DSML gives you the power to access your LDAP information as XML

The Directory Service Markup Language adds XML functionality to your directory services

In today's e-business environment, effective and efficient data management is crucial. As such, two technologies prove vital to proper data management: directory services and XML. Directory services allow you to store and manage data, and are thus a necessary prerequisite for conducting online business. XML provides an effective way to present and transfer data. With that in mind, there's a clear need to bridge the two technologies. The solution may be the new Directory Service Markup Language (DSML) standard, designed to make directory services more dynamic by employing XML.

DSML, in conjunction with technologies like XSL and EJB, provides a unique way to solve many business problems, such as supply chain management and customer support. The technology was developed out of work done by Bowstreet and completed by the DSML Working Group; the latter submitted the technology to the W3C and OASIS standards groups in December 1999.

Numerous products now support DSML, including Sun's recently released DSML Service Provider 1.2 technology, released for use with J2EE 1.2, and Exolab's Castor API, which provides services to extract information from LDAP as DSML, as well as . With the DSML Service Provider, which implements the javax.naming.directory.DirContext interface, you can access a DSML document, manipulate and update its contents by using the JNDI APIs, and then re-export the contents in DSML format.

In this article, we will discuss how to create dynamic content and manipulate a directory server with DSML. We will do this using Exolab's Castor, servlets, and XSL. We will not cover information on how to use XSL or servlets, but for more information you can refer to Michael Ball's earlier article, "XSL Gives Your XML Some Style" (JavaWorld, June 30, 2000).

Download the source code associated with this article.

The DSML specification

The DSML specification provides both a DTD and a schema for reference (see Resources for download information).

The DSML namespace is: http://www.dsml.org/DSML; it should be referenced in all your DSML documents, as in:

<dsml:dsml xmlsn:dsml:"http://www.dsml.org/DSML"/>

A DSML document can describe directory entries, a directory structure, or both. The root node in a DSML document is always dsml. Children can be directory-schema and/or directory-entries.

<dsml:dsml xmlns:dsml=http://www.dsml.org/DSML>

A directory-entries node has one or more directory-entry nodes as its children. Each of these directory-entry nodes contains the object class information as well as attributes for the information, as seen below:

  <dsml:directory-entry dn="uid=mball, ou=consultant, 
    <dsml:attr name="cn"><dsml:value>Michael 

If an attribute is multivalued, there would be many <dsml:value> nodes under the <dsml:attr> node.

The directory-schema node contains information about the object classes and the attributes. As an example, take a look at this sample XML pulled from the DSML specification:

<dsml: class
  <dsml:attribute ref="#sn" required="true"/>
  <dsml:attribute ref="#cn" required="true"/>
  <dsml:attribute ref="#userPassword" required="false"/>
  <dsml:attribute ref="#telephoneNumber" required="false"/>
  <dsml:attribute ref="#seeAlso" required="false"/>
  <dsml:attribute ref="#description" required="false"/>

Using DSML with Exolab's Castor API

Exolab, according to its homepage, works on the development of open source enterprise software projects, of which the Castor API is one. DSML is a subset of Castor. Be aware that the documentation for the DSML portion of Castor is scarce. With that in mind, we'll try to highlight the basic functionality of the API.

Note: In this article, we used Castor version 0.8.5.

Castor's DSML API

Castor can import and export directory information. Meanwhile, the DSML specification describes producers (used for exporting) and consumers (used for importing). Both of these services are divided into types 1 and 2 based on the document levels that they can handle. The four document levels are:

  • Level 1: A document containing no schema information, just directory entries
  • Level 2: A document containing directory entry information and a link to an external schema
  • Level 3: A document containing only a directory schema
  • Level 4: A document containing a directory schema and directory entries

A type 1 producer can handle the level 1 documents, while a type 2 producer can handle document levels 2, 3, and 4.

Meanwhile, a type 1 consumer must handle all four document levels. A type 2 consumer must handle all four document levels and make use of schema information.

Castor supports type 1 producers and consumers.

Producers (exporting)

Castor can provide level 1 and level 4 documents. The class org.exolab.castor.dsml.Exporter retrieves directory information from a directory server.

The Exporter class uses a org.exolab.castor.dsml.SearchDescriptor to define search parameters. The SearchDescriptor can be configured by calling setters on it. Table 1 defines some of the important methods available in the SearchDescriptor class.

The methods available to class SearchDescriptor

void SearchDescriptor()Constructor
void addReturnAttr(java.lang.String attrName)Causes exporter to only return attributes added to the SearchDescriptor
void setBaseDN(java.lang.String baseDN)Sets the base domain name to search on
void setFilter(java.lang.String filter)Sets a filter to search on (i.e., uid=lpoe)
void setScope(int scope)1= , 2= , 3=
java.lang.String getBaseDN()Gets the base domain name to search on
java.lang.String getFilter()Gets the filter used to search
java.lang.String[] getReturnAttrs()Gets an array of the attributes to retrieve
int getScope()Gets the scope
java.util.Enumeration listReturnAttrs()Gets an Enumeration of the attributes to retrieve

To use XML to define the search configuration, Exporter employs the following methods:

  • void readSearchDescriptor(java.io.InputStream input)
  • void readSearchDescriptor(java.io.Reader input)

The XML used would look like:

   <search base="o=javaworld.com" scope="subtree" 

Exporter uses the SearchDescriptor to determine what information to get from the directory server. To invoke exporting, Exporter includes two method calls:

  • void export(java.io.OutputStream output, boolean serverSchema, boolean importPolicy)
  • void export(java.io.Writer output, boolean serverSchema, boolean importPolicy)

Both methods require two flags, although we yielded the same result regardless of the their settings (no documentation on the flags' meaning was forthcoming).

Consumers (importing)

Castor can consume level 1 documents. The org.exolab.castor.dsml.Importer class puts directory information into a directory server.

Importer uses the base constructor and defines three methods for importing DSML:

  • void importDocument(java.io.InputStream stream)
  • void importDocument(org.xml.sax.Parser parser, org.xml.sax.InputSource input)
  • void importDocument(java.io.Reader reader)

These methods take the DSML that you want to import. In addition to the DSML, you can also define an ImportDescriptor. The ImportDescriptor defines the location to where you are importing and configures restrictions on deleting or updating attributes. The following methods, which take XML as input, are used to set up the ImportDescriptor:

  • void readImportDescriptor(java.io.InputStream input)
  • void readImportDescriptor(java.io.Reader input)

The XML used with the readImportDescriptor() method would look something like this:

     <import-policy dn="o=javaworld.com" delete-empty="true"
                    replace-attr="true" refresh-only="false"/>

The example

In a directory server, it's useful to have users separated into groups. As such, in our example we will create a dynamic way to distribute our users into groups through a Web interface. By retrieving a list of all the users in our organization, we can select the users we want to be in a group and load the new group into our directory server.

This example will illustrate two uses of DSML: data extraction for presentation purposes and directory-service entry creation with XSL and Castor.

To accomplish our task, we'll employ a servlet that uses the Castor API to retrieve data from a directory server. Then we will transform the resulting DSML into HTML using XSL.

Retrieve a list of users

The servlet below -- DisplayUser.java -- retrieves DSML from a DirectoryServer using Castor's Exporter class. It then strips the namespace off the DSML because the Xalan processor cannot handle namespaces in source XML. Finally, the servlet uses XSL to style the DSML in order to create HTML. Figure 1 shows a static HTML page used to perform the search of a directory server.

Here's the DisplayUser.java servlet's code:

import java.io.*;
import javax.servlet.*;
import javax.servlet.http.*;
import javax.naming.directory.InitialDirContext;
import javax.naming.Context;
import netscape.ldap.LDAPConnection;
import java.util.Hashtable;
import org.exolab.castor.dsml.Exporter;
import org.exolab.castor.dsml.jndi.JNDIExporter;
import org.exolab.castor.dsml.SearchDescriptor;
import RemoveNamespace;
public class DisplayUser extends HttpServlet
  InitialDirContext ctx;
  public void init()
      LDAPConnection conn;
      Hashtable env;
        //connect to LDAP Server
        env = new Hashtable();
        env.put(Context.SECURITY_AUTHENTICATION, "simple");
        env.put(Context.PROVIDER_URL, "ldap://localhost:389");
        env.put(Context.SECURITY_PRINCIPAL, "cn=Directory Manager");
        env.put(Context.SECURITY_CREDENTIALS, "password");
        ctx = new InitialDirContext(env);
      catch(Exception ex)
   public void doPost(HttpServletRequest req, HttpServletResponse res)
        RemoveNamespace remover = new RemoveNamespace();
        Search search = new Search(ctx);
        StringBuffer xsl = new StringBuffer();
        String results = "";
        String dsml = "";
        //get parameters
        String attributeName = req.getParameter("attributeName");
        String searchValue = req.getParameter("searchValue");
        //search directory
        results = search.doSearch(attributeName,searchValue);
        //remove the namespace from dsml
        dsml = remover.remove(results);
        //style search results
        XSLUtil xslUtil = new XSLUtil();
        File file = new File("DisplayUser.xsl");
        String html = xslUtil.styleXSL(file, dsml);
        PrintWriter out = res.getWriter();
      catch(Exception e)
   public String doSearch(String attribute, String searchValue)
      Exporter exporter = new JNDIExporter(ctx);
      SearchDescriptor sd = new SearchDescriptor();
      StringWriter writer = new StringWriter();
        //configure the search descriptor
        //set the domain name to search
        //set the scope of the search
        sd.setScope(3);  //scope 1 = base, 2 = onelevel, 3 = subtree
        //setFilter for Directory Server
        //in our case the attribute is departmentnumber
        //assign search descriptor
        //perform the search of the Directory Server
      catch(Exception e)
      //return String (dsml) of search results
      return writer.toString();

Next, let's take a look at the XSL code used to transform the DSML pulled from LDAP into HTML:

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="1.0"
<xsl:template match="/">
   <head><title>Search Results</title></head>
    <form action="CreateGroup" method="post">
     <table border="1" width="100%">
       <th> Employee Name</th>
       <th> User ID</th>
      <xsl:for-each select="dsml/directory-entries/entry">
         <xsl:value-of select="attr[@name='cn']/value"/>
         <xsl:value-of select="attr[@name='uid']/value"/>
         <xsl:value-of select="attr[@name='mail']/value"/>
         <xsl:value-of select="attr[@name='telephonenumber']/value"/>
         <xsl:element name="input">
          <xsl:attribute name="type">checkbox</xsl:attribute>
          <xsl:attribute name="name">member</xsl:attribute>
          <xsl:attribute name="value"><xsl:value-of
     <hr width="100%" noshade="true"></hr>
        Group Name:
        <input type="text" size="20" name="GroupName"/>
        <input type="submit" value="Submit"/>
        <input type="reset" value="Reset"/>

In Figure 2, we see the HTML that results after the XSL styles the DSML.

Figure 2. The HTML after the DSML-to-HTML transformation.

Create a group in DSML

Next, the CreateGroup.java servlet will take the HTTPRequest and format the parameters into XML. For example:


would be translated into:


The XSL stylesheet below then takes the XML and styles it into DSML:

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet version="1.0"
<xsl:template match="/">
<xsl:element name="dsml">
  <xsl:element name="directory-entries">
   <xsl:element name="entry">
     <xsl:attribute name="dn">cn=<xsl:value-of
     <xsl:element name="objectclass"><xsl:element
     <xsl:element name="attr">
      <xsl:attribute name="name">cn</xsl:attribute>
      <xsl:element name="value">
       <xsl:value-of select="data/GroupName"/>
     <xsl:element name="attr">
      <xsl:attribute name="name">uniquemember</xsl:attribute>
      <xsl:for-each select="data/member">
       <xsl:element name="value"><xsl:value-of

The XSL stylesheet above creates the following DSML:

<?xml version="1.0" encoding="UTF-8"?>
  <entry dn="cn=Developers,ou=Groups,o=aici.com">
   <attr name="cn">
   <attr name="uniquemember">

Now that we've created the DSML, we can load it into the directory server using the Importer class:

public String doCreate(String dsml)
      Importer importer = new JNDIImporter(ctx);
      PrintImportListener printer;
      StringWriter out = new StringWriter();
        StringWriter writer = new StringWriter();
        //set writer
        printer = new PrintImportListener( new PrintWriter( out ) );
        //set event listener
        importer.setImportEventListener( printer );
        //set the import policy
        importer.readImportDescriptor(new FileReader("import.xml"));
        //insert records into LDAP
    importer.importDocument(new StringReader(dsml));
      catch(Exception e)
      return out.toString();

We can now confirm that the group entry, as well as its two new members, was added to LDAP, as illustrated in Figure 3.

Figure 3. Now the new group is visible in our console.

Taking from the example general concepts such as data extraction and object generation, we can open up a world of opportunities to use and manage the data stored in your directory-service.


Directory services are powerful tools common in most of today's architectures. They store information about people, hardware, and even objects. But as directory-service use becomes more pervasive, a standard way to manipulate the information inside of them and transfer data between them becomes necessary. By taking advantage of the power of XML, DSML provides a powerful interface into directory servers. As XML tools become more and more common, using them to access DSML will become easier, and your directory servers will become more useful and dynamic.

Lanette Poe and Michael Ball are both senior consultants with Software Architects. Lanette is a Sun-certified Java programmer, and Mike is a Sun-certified Java developer and IBM-certified XML developer. Over the past six months, Mike and Lanette have been working on a project utilizing DSML, XSL, and servlets.

Learn more about this topic

Join the discussion
Be the first to comment on this article. Our Commenting Policies
See more