Toward an event-driven infrastructure for in-situ monitoring systems 

Monitoring systems are used to understand and predict natural phenomena so that science-based wise decision can be opportunely taken to preserve natural resources and ecosystems. Today, thanks to IoT technologies, in-situ sensors allow to collect unprecedent number of observations of several environmental parameters at high frequency and at high spatial resolution so that fast evolving phenomena, that cannot be derived from satellite information, can be detected and analysed. The management of these data pose new challenges and requires a paradigm shift to fully exploit their potential by moving from time-based to event-based management. For this reason, we are redesigning istSOS, a long-lasting Open-Source project implementing the Sensor Observation Service: an interoperable Open Standards from the Open Geospatial Consortium. The new solution, that is currently under intense development, is based on the following individuated requirements: (i) Open Source software to guarantee the forever free usage for everyone with full open rights; (ii) Interoperability to offer a standard interface for data and metadata; (iii) reliability of adopted solutions to guarantees a solution to be used in production; (iv) microservices based, so that specialized tasks can be performed atomically using the best technologies; (v) reactive approach that enable real-time and asynchronous exploitation of information; (vi) choreographed architecture which permits single tasks to be autonomously executed without any external coordinator. To respond to the individuated requirements the new system is based on the OM:Observation standard as a core element to describe observations and Apache Kafka as reliable, scalable, durable and high performing solution for data stream management. Based on these two core elements, several microservices for observation consuming, archiving, processing and notification have been created and integrated in the architecture which is available as containerized composed solution. The solution is being tested in production environment for the management of the hydro meteorological monitoring data for the Canton Ticino in Southern Switzerland. With this work we share our results and our experience and discuss individuated challenges and barriers.