Thinking autonomic for sensing devices

This paper presents a middleware that aims to respond to the increasing management needs of networked sensing devices, such as remote software deployment, dynamic configuration and real-time performance monitoring and tuning. These are essential functions for building adaptable systems providing high quality of service. Furthermore, autonomic management bears particular importance for large scale systems having timely response requirements. Our middleware allows creating highlevel policies that describe the actions to be autonomously performed when interesting events occur in the system. The policies are formulated as Event-Condition-Action (ECA) rules. They are created by using UML-like diagrams and can be deployed at any level of the distributed architecture of the system.