Taming the interoperability challenges of complex IoT systems

The Internet of Things is characterised by extreme heterogeneity of communication protocols and data formats; hence ensuring diverse devices can interoperate with one another remains a significant challenge. Model-driven development and testing solutions have been proposed as methods to aid software developers achieve interoperability compliance in the face of this increasing complexity. However, current approaches often involve complicated and domain specific models (e.g. web services described by WSDL). In this paper, we explore a lightweight, middleware independent, model-driven development framework to help developers tame the challenges of composing IoT services that interoperate with one another. The framework is based upon two key contributions: i) patterns of interoperability behaviour, and ii) a software framework to monitor and reason about interoperability success or failure. We show using a case-study from the FI-WARE Future Internet Service domain that this interoperability framework can support non-expert developers address interoperability challenges. We also deployed tools built atop the framework and made them available in the XIFI large-scale FI-PPP test environment.