A reconfigurable middleware for context-aware applications in autonomic computing

The vision of autonomic computing involves distributed nodes capable of managing and preserving themselves. In practice, autonomic computing is also strongly connected to the concepts of mobility and platform heterogeneity since next generation networks assume different types of mobile nodes that operate inside ad hoc environments. In such cases where the context changes can be frequent, the capability of capturing the environmental changes is crucial. However, it cannot be assumed that all the nodes are equipped with all possible types of sensors in order to locally retrieve the required contextual information. Sensor information exchange between the nodes through efficient data dissemination mechanisms can further improve the overall awareness of the network. Another challenging task for autonomic computing concerns the self-reconfiguration of the autonomous nodes according to the sensed context. This paper presents the overall architecture design, the implementation and the evaluation of a middleware developed in the context of the Integrated Platform for Autonomic Computing (IPAC). This middleware is targeted to embedded devices and supports mobile context-aware applications that render the aforementioned desired behavior. The paper focuses on two aspects of autonomic computing: collaborative context-awareness and self-reconfiguration.