An Architecture-Centric Approach for Dynamic Smart Spaces

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 architecture-centric solution that merges component-based control and bio-inspired (fireflies-based) mechanisms. Suitable abstractions help conceive self-organizing 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.