Design of Semantic Information Broker for Localized Computing Environments in the Internet of Things

Technologies of the Internet of Things (IoT) make all the devices of a spatial-limited physical computing environment interconnected as well as connected to the Internet. This ability leads to the consideration of notion of localized IoT-environments, which now appear in many places of everyday life. The next generation of software applications (smart applications) can be deployed in localized IoT-environments based on the smart spaces paradigm. We consider the M3 architecture (multi-device, multivendor, multi-domain) for smart spaces. Interaction of agents is performed via information sharing, rather than sending messages to one another directly. The information and its semantics are collected in a smart space using ontological representation models of the Semantic Web and forming a knowledge base for interoperable sharing among participants. A smart space is maintained by a Semantic Information Broker (SIB), which runs on a host accessible by any device of the localized IoT-environment. SIB maintains an RDF triplestore, which represents the smart space content and acts as an informational hub relating many data sources. SIB provides access and reasoning primitives for agents to operate over the collected information.

On one hand, the existing SIB implementations clearly showed elegant properties of the M3 architecture. On the other hand, the application development needs consideration of many technical aspects in order to achieve satisfactory operation in real-life settings. In this talk, we present a renewed SIB design with increased extensibility, dependability, and portability. The modular approach is applied for high extensibility. Interprocess communication in SIB is optimized along the path from agent request arrival to SIB till the result is formed for delivering to the agent, which leads to higher dependability. The pool of access operations is systematized based on the original Smart Spaces Access Protocol (SSAP) and its variants as well as extended with mechanisms previously introduced in Knowledge Sharing protocol (KSP). The C/C++ programming language and Qt framework are used for the SIB implementation for higher portability. The research done is a step towards an efficient open interoperability platform for the smart space application development.