MQfilTTr: Strengthening Smart Home Privacy Through MQTT Traffic Manipulation

We present the software design for an easily extendable architecture that prevents leakage of privacy-relevant data in messages sent from smart home environments to external servers outside of a smart home. Such problems arise since users are mostly unable to control which data leaves the local network unfiltered. We address this situation by a privacy-enforcing hub architecture located at the smart home gateway; our approach enables manipulation of e.g. MQTT messages in order to block or rewrite data before it leaves the smart home. We exemplify our architecture for information dissemination control using MQTT, since it is a well-established and frequently used communication protocol in Internet of Things applications. Our performance analysis demonstrates a comparably low overhead that can easily be tolerated. We also discuss how brown field (e.g. existing) devices can be integrated into such an approach which also highlights our extension of modeling the network protocol’s sender and receiver functions separately facilitating Node-RED; moreover we investigate the danger of circumventing such a hub architecture by exploiting metadata side channels; both these issues have not been sufficiently addressed in previous works, which we extend in this work.