Enhancing Transparency of MQTT Brokers For IoT Applications Through Provenance Streams

Systems based on Internet of Things (IoT) technologies may violate user privacy if personal data they produce and use become available to unauthorised agents. Recording provenance of IoT system behaviour may support assessment mechanisms ensuring compliance of system components with data access constraints. In this paper, we describe a prototype implementation of a provenance-enabled MQTT broker enhanced with the ability to generate provenance records describing the actual broker behaviour during message forwarding. The implementation utilises a semantic stream based approach for generating and analysing provenance data to discover message forwarding to untrusted agents. The initial evaluation demonstrates the feasibility of semantic solutions in this context.