A service diagnosis and adaptation scheme in service-oriented IoT

IoT embeds heterogeneous sensors and actuators in the bottom layer, and the robustness and privacy of services and applications are becoming main concerns. This paper presents an across-layer service diagnosis and adaptation scheme for IoT using the service-oriented architecture. We exploit a decentralized diagnostic model to identify the faulty IoT services. We propose two indicators for services to estimate their fault probability based on sensing data distances and applications feedbacks. We model an office-building scenario as an example, to elaborate the diagnosis and adaptation procedures. We design the rules of service adaptation based on these indicators, and conduct experiments to evaluate the performance and applicability of the scheme.