Is EJB dead in times of SOA?

In times of lightweight containers and Service Oriented Architecture (SOA), Enterprise Java Beans (EJB) is not the predominant development model anymore. Instead multiple technologies and frameworks are competing to be THE service development model of the future. For instance SCA, Spring, Axis, CXF and …

… probably EJB?

The Jbi4Ejb Binding Component allows to integrate Stateless Session EJBs into an ESB infrastructure, thus making them available to service consumers.

This is again a nice example of how open standards such as JBI help to protect the investments of the past.

SOA without BPM without SOA

Some people argue that SOA and BPM are two different things. That is probably right because it is possible to create a service landscape without having a business process representation. But if you look at the main reason for establishing a service oriented architecture the perspective changes.
The main reason for SOA is business agility. Every technology that helps to achieve that goal is welcome to be integrated in the SOA technology stack.
BPM makes perfect sense as it fosters better aligment on business and IT and helps to shorten development cycles.
Ok understood. But if that is true wouldn’t be be enough to use a BPM only without services? The answer is no because in order to create a flexible BPM solution it must be based on a sound service landscape. If that is not the case changing a business process would require new services to be created or existing to be altered. Sooner or later this would lead to service proliferation which would render the BPM approach useless over time. The result would be anything but hardly an agile system.

My conclusion: Using BPM as part of a SOA makes perfect sense if it is based on reliable and well designed services.