In accordance with the previous chapter, a particular class of smart environments is created by Smart Spaces, where many devices participate using information-driven and ontology-oriented interaction. In this case, a smart space is developed based on models from multi-agent systems and knowledge manipulation technologies from the Semantic Web. In this chapter, we consider this particular approach for creating such smart environments. The M3 architecture (multidevice, multivendor, multidomain) aims at development of smart spaces that host advanced service-oriented applications. We introduce the theoretical background of the M3 architecture in respect to its open source implementation—the Smart-M3 platform. The latter forms a technology for creating M3-based smart spaces (M3 spaces) as heterogeneous dynamic multi-agent systems with multi-device, multi-vendor, multi-domain devices and services. We further consider the concept models of space computing that enable the studied class of smart spaces, derive the generic properties that an M3 space design requires, and describe the basic software components of M3 architecture that realize the generic design properties in accordance with the concept models.
