Monitoring of environment conditions on a farm using the concepts of Smart Farming is able to bring benefits, by providing analyses backed by data aggregated from previously unavailable breadth of sources. However, it introduces several technical challenges in terms of integrating very heterogeneous data (sensors, vehicles, etc.), validating and fusing these types of data, and turning them into useful information for farmers and agronomists. We describe an open integration architecture designed to address the above challenges and needs, including components we developed and used in validation demonstrators that provide sensor data integration (SensLog) and external systems interoperability (OGC SensorThings API connector).