INTRODUCING INTELLIGENCE IN DISTRIBUTED SYSTEMS MANAGEMENT

In recent years, there has been growing use of distributed systems, and therefore a growing demand for tools able to manage them. However, these tools are not always well suited to their task, in part because of the lack of built-in intelligence. This paper discusses how to introduce intelligence to distributed systems management and control. To do so we analyse requirements for distributed systems management and focus on management tasks from the architecture and model points of view. Intelligence is introduced through distributed artificial intelligence concepts and multiagent systems. We illustrate the application of our techniques with an example based on traffic and congestion control in ATM networks.