Abstract
We would like to share our research journey towards enabling DevOps for IoT applications, and how Open Source makes the journey feasible and fun. DevOps is widely adopted for developing cloud applications, which supports developers in continuously placing software changes directly to production. As companies are including IoT and Edge devices into their IT infrastructures, supporting DevOps for IoT is a must. However, IoT challenges some fundamental assumptions behind DevOps, such as the homogeneous infrastructure and centralized governance, and therefore, breaking-through research is needed. Funded by H2020, 30 people from 12 partners crossing academia and industry gathered to solve these fundamental challenges, which results in full-stack open source tools for automatic deployment, learning-based operation and security monitoring of IoT applications, and risk management of the development process. The tools are evaluated on industrial use cases in intelligent transportation, smart building, and eHealth. The mass open source tools and communities around IoT development provides the sound foundation for this design research and the opportunities for the further exploitation of the results. In particular, we are proud of spinning off a start-up to commercialize the risk management services in the open source + SaaS model.