The architecture of distributed systems driven by autonomic patterns

The autonomic computing notion has introduced the concept of selfoptimizing, self-healing and auto-diagnosis applications. In this article we would like to present how this idea affects the building of distributed systems. As a reference base, we take advantage of the Data Management System (DMS), which has been developed within the scope of the PROGRESS project. DMS enables the creation of a grid environment capable of storing large amounts of data. The complex architecture of this system, which constitutes a model of loosely coupled components, involves a special approach to its maintenance and management. To address these problems, we have applied the autonomic computing patterns in the DMS architecture. Our solution was designed to be reused in any project dealing with the same issues. It can also act as an autonomic service for any other applications and services.