Dependable data sharing in dynamic IoT-systems: subject-oriented process design, complex event processing, and blockchains

Internet of Things (IoT) applications are increasingly used in domains critical for society and adapted to individual needs. Thereby, various physical entities, humans and computer programs are dynamically networked while being situated in different locations. For system stakeholders, reliable operation requires intelligible design and adaptability at run time. The more users can adjust IoT systems dynamically, the more transparent has to be how specific requirements can be met. Executable and role-specific process models are a viable starting point, as they represent the behavior of components and the flow of control. However, they need to be extended with effective data exchange mechanisms between IoT system elements. In this paper, we utilize the subject-oriented concept and notation for role-specific process modeling and communication design of IoT systems and enrich them with a data sharing facility, so-called shared input pools. In order to avoid a single point of failure in system operation, we propose to implement them as blockchains. IoT-elements can then interact and exchange data in a semantically correct and robust way, in particular when behavior changes may occur dynamically.