Chapter 11. A Quick Tour of EJB

Traditionally, distributed applications were built using protocols, most notably TCP/IP, designed to allow autonomous processes on separate machines to exchange data. A typical client-server application, for example, might comprise a client application that does the processing and display of data and a database server that does nothing but provide the data. The reverse situation is also possible; the client application may be responsible only for the presentation of the data with all processing done on the server.

Modern distributed applications, however, often do not just transfer data but in addition allow one machine to invoke procedures on a remote machine. A popular standard for remote procedure calls (RPC) was developed in the 1980s. The Common Object Request Broker Architecture (CORBA) extended this procedural approach to an object-oriented one by specifying an infrastructure and protocol for distributed objects; this standard allows programs running on different types of hardware and written using different programming languages, to interoperate. EJB, the J2EE distributed object model, is based on CORBA (as well as Java's Remote Method Invocation technology) and is, to some degree, interoperable with CORBA.

In addition to benefits such as better sharing of resources and security, remote object architectures such as CORBA and EJB enable us to build reusable software components that can interoperate with other software components without regard to their location on the network. EJB, because it is based on Java, improves the portability of distributed components across application servers and hardware platforms.