An architecture-centric approach for dynamic smart spaces

Abstract

The development of sound and reliable dynamic smart spaces is a complex task. Many researchers have already addressed the problem from different angles. The autonomic computing community has been focusing on super-imposed adaptation mechanisms by adding further dedicated components to the (software) architecture of the system. In contrast, bio-inspired solutions provide inherent support to self-organization but they fail to guarantee the desired level of reliability and control. This paper aims to blend the two views and proposes an architecturecentric solution that merges component-based control and bio-inspired (fireflies-based) mechanisms. Suitable abstractions help conceive selforganizing ad-hoc collaborations among the —virtual and physical— components of a space. An example public park is used throughout the paper to explain and exemplify the key features of the proposed solution.