Edge or Fog computing is a relatively new architectural deployment model, ideally fit for the unique requirements of the Internet of Things. This paper presents a novel solution, which leverages the architectural characteristics of edge computing for security reasons. Machine learning models (specifically Support Vector Machines) are employed on the edge of the cloud, to perform low footprint unsupervised learning and analysis of sensor data for anomaly detection purposes. To this end, a proof of concept system is developed, capable of detecting anomalies in real world vessel sensor streams (big data) in a smart port environment. We report on early results, that validate the potential of the solution. The quality and performance of the model is investigated in real world conditions.