Enterprise Java Beans, Distributed Artificial Intelligence and Group Communication

Today's Middleware is facing increasing pressures to operate in more open, loosely coupled and complex environments. E-business systems require seamless transactions between multiple enterprises and software components are becoming more opaque in an effort to hide their complexity. In parallel, interactions between components are becoming increasingly sophisticated. This paper describes: - The integration of a well known standards compliant agent platform with a transactional J2EE application server to create a hybrid middleware able to make use of high level agent communication and other features. - An solution for Group Communication (GC) (a known Distributed Systems problem) using the hybrid middleware. - The extension of GC to illustrate how new challenges in distributed applications might be dealt with by a convergence of Distributed Systems theory, Distributed Artificial Intelligence and Middleware solutions. The purpose of the paper is not to eulogise any particular technology but to highlight the relevance of research in the areas of Distributed systems, Distributed Artificial Intelligence and Middleware and illustrate how they might co-evolve.