Semantics based service orchestration in IoT

Internet of Things (IoT) presents a dynamic global revolution in the Internet where physical and virtual "things" communicate and share information. There arises a need to allow heterogeneous "things" to seamlessly interoperate, interact and exchange information and subsequently share services. Services are represented as functionalities that are offered by the "things". In IoT context, service orchestration refers to identifying which devices or software components are required to formulate the requested service. IoT devices can offer different services based on the context and semantics. Service orchestration provides an approach to integration and interoperability which decouples applications from each other, and enhances capabilities to centrally manage and monitor components. This paper investigates requirements for semantic interoperability and exposes current challenges in IoT interoperability as a means of facilitating services orchestration in IoT. The paper proposes a semantic platform that allows heterogeneous devices to collaborate thereby enabling dynamic service orchestration. The proposed platform provides a common ontology-based framework for representing semantics allowing for a consistent information exchange format. The approach used combines domain-specific ontologies to obtain device and service abstraction for the purposes of process mining and analysis of service orchestration in heterogonous IoT environments. This approach was evaluated using a prototype IoT environment and mining techniques developed in the Process Mining Framework (ProM). The results revealed that semantically enriching service annotation and device descriptions through ontologies can successfully enable interoperability in heterogeneous IoT environments. Semantically enriched descriptions enable multiple perspectives of analysis on service orchestration processes and expose the steps involved in service request and provision in an a typical IoT setup.