Home > Articles > Programming > Java

JBossNS The JBoss JNDI Naming Service

  • Print
  • + Share This
  • 💬 Discuss
Learn how to use Java Name Directory Interface (JNDI) with the JBoss naming service JBossNS and find out what exactly is in a name.

In this chapter

  • An Overview of JNDI
  • J2EE and JNDI—The Application Component Environment
  • The JBossNS Architecture
  • Additional Naming MBeans

This chapter discusses the JBoss JNDI-based naming service as well as the role of JNDI in JBoss and J2EE. An introduction to the basic JNDI API and common usage conventions is covered, and the JBoss specific configuration of J2EE component naming environments defined by the standard deployment descriptors is also addressed. The configuration and architecture of the JBoss naming service component, JBossNS, is examined, as well.

The JBoss naming service is an implementation of the Java Naming and Directory Interface (JNDI). JNDI plays a key role in J2EE because it provides a naming service that allows a user to map a name onto an object. This is a fundamental need in any programming environment because developers and administrators want to be able to refer to objects and services by recognizable names. A good example of a pervasive naming service is the Internet Domain Name System (DNS). The DNS service allows you to refer to hosts using logical names, rather than their numeric Internet addresses. JNDI serves a similar role in J2EE by enabling developers and administrators to create name-to-object bindings for use in J2EE components.

An Overview of JNDI

JNDI is a standard Java API that is bundled with JDK1.3 and higher. JNDI provides a common interface to a variety of existing naming services: DNS, LDAP, Active Directory, RMI registry, COS registry, NIS, and file systems. The JNDI API is divided logically into a client API that is used to access naming services, and a service provider interface (SPI) that allows the user to create JNDI implementations for naming services.

The SPI layer is an abstraction that naming service providers must implement to enable the core JNDI classes to expose the naming service using the common JNDI client interface. An implementation of JNDI for a naming service is referred to as a JNDI provider. JBossNS is an example JNDI implementation, based on the SPI classes. Note that the JNDI SPI is not needed by J2EE component developers.

This introduction to JNDI covers the basic concepts and JNDI client API usage.


For a thorough introduction and tutorial on JNDI, which covers both the client and service provider APIs, see the Sun tutorial at http://java.sun.com/products/jndi/tutorial/.


The main JNDI API package is the javax.naming package. It contains five interfaces, 10 classes, and several exceptions. There is one key class, InitialContext, and two key interfaces, Context and Name.


The notion of a name is of fundamental importance in JNDI. The naming system determines the syntax that the name must follow. The syntax of the naming system allows the user to parse string representations of names into its components. A name is used with a naming system to locate objects. In the simplest sense, a naming system is a collection of objects with unique names. To locate an object in a naming system, you provide a name to the naming system, and the naming system returns the object store under the name.

As an example, consider the Unix file system's naming convention. Each file is named from its path relative to the root of the file system, with each component in the path separated by the forward slash character ("/"). The file's path is ordered from left to right. The pathname, /usr/jboss/readme.txt, for example, names a file readme.txt in the directory jboss, under the directory usr, located in the root of the file system. JBossNS uses a Unix-style namespace as its naming convention.

The javax.naming.Name interface represents a generic name as an ordered sequence of components. It can be a composite name (one that spans multiple namespaces), or a compound name (one that is used within a single hierarchical naming system). The components of a name are numbered. The indexes of a name with N components range from 0 up to, but not including, N. The most significant component is at index 0. An empty name has no components.

A composite name is a sequence of component names tha span multiple namespaces. An example of a composite name would be the hostname+file commonly used with Unix commands like scp. For example, this command copies localfile.txt to the file remotefile.txt in the tmp directory on host ahost.someorg.org:

scp localfile.txt ahost.someorg.org:/tmp/remotefile.txt

The ahost.someorg.org:/tmp/remotefile.txt is a composite name that spans the DNS and Unix file system namespaces. The components of the composite name are ahost.someorg.org and /tmp/remotefile.txt. A component is a string name from the namespace of a naming system. If the component comes from a hierarchical namespace, that component can be further parsed into its atomic parts by using the javax.naming.CompoundName class. The JNDI API provides the javax.naming.CompositeName class as the implementation of the Name interface for composite names.

A compound name is derived from a hierarchical namespace. Each component in a compound name is an atomic name, meaning a string that cannot be parsed into smaller components. A file pathname in the Unix file system is an example of a compound name.


The javax.naming.Context interface is the primary interface for interacting with a naming service. The Context interface represents a set of name-to-object bindings. Every context has an associated naming convention that determines how the context parses string names into javax.naming.Name instances. To create a name-to-object binding, you invoke the bind method of a Context and then specify a name and an object as arguments. The object can later be retrieved using its name using the Context lookup method. A Context will typically provide operations for binding a name to an object, unbinding a name, and obtaining a listing of all name-to-object bindings. The object you bind into a Context can itself be of type Context. The Context object that is bound is referred to as a subcontext of the Context on which the bind method was invoked.

As an example, consider a file directory with a pathname /usr, which is a context in the Unix file system. A file directory named relative to another file directory is a subcontext (commonly referred to as a subdirectory). A file directory with a pathname /usr/jboss names a jboss context that is a subcontext of usr. In another example, a DNS domain, such as org, is a context. A DNS domain named relative to another DNS domain is another example of a subcontext. In the DNS domain jboss.org, the DNS domain jboss is a subcontext of org because DNS names are parsed right to left.

Obtaining a Context Using InitialContext

All naming service operations are performed on some implementation of the Context interface. Therefore, you need a way to obtain a Context for the naming service you are interested in using. The javax.naming.IntialContext class implements the Context interface, and provides the starting point for interacting with a naming service.

When you create an InitialContext, it is initialized with properties from the environment. JNDI determines each property's value by merging the values from the following two sources, in order such as:

  • The first occurrence of the property from the constructor's environment parameter and (for appropriate properties) the applet parameters and system properties.

  • All jndi.properties resource files found on the classpath.

For each property found in both of these two sources, the property's value is determined as follows. If the property is one of the standard JNDI properties that specify a list of JNDI factories, all of the values are concatenated into a single, colon-separated list. For other properties, only the first value found is used. The preferred method of specifying the JNDI environment properties is through a jndi.properties file. The reason is that this allows your code to externalize the JNDI provider specific information, and changing JNDI providers will not require changes to your code; thus it avoids the need to recompile to be able to see the change.

The Context implementation used internally by the InitialContext class is determined at runtime. The default policy uses the environment property "java.naming.factory.initial", which contains the class name of the javax.naming.spi.InitialContextFactory implementation. You obtain the name of the InitialContextFactory class from the naming service provider you are using.

Listing 3.1 gives a sample jndi.properties file a client application would use to connect to a JBossNS service running on the local host at port 1099. The client application would need to have the jndi.properties file available on the application classpath. These are the properties that the JBossNS JNDI implementation requires. Other JNDI providers will have different properties and values.

Listing 3.1 Sample JBossNS jndi.properties File

### JBossNS properties
  • + Share This
  • 🔖 Save To Your Account


comments powered by Disqus