UPGRADE journal, 2011, Issue No. 1: 5-16. February 2011. ISSN: 1684-5285
Pervasive environments are highly dynamic with lots of heterogeneous devices which share information through increasingly interconnected networks. In this context semantic models can be used to describe the context that surrounds them in a very expressive manner, usually stored in centralized knowledge bases. The applications built on top of these knowledge bases use heterogeneous protocols to transmit their data, but do not capture the dynamicism of the network. The presented middleware facilitates the exchange of knowledge between different sensors and actuators in a highly distributed, decoupled and resource oriented manner following the Triple Space paradigm. This middleware has been tested on an stereotypical scenario, which illustrates how different peers can exchange data keeping them autonomous and yet with a reasonable footprint for devices with reduced computation capabilities.