Exploring Data Exhaust Under Different IoT Traffic Scenarios

This work explores the potential cases of user data exhaust in IoT traffic under different scenarios. Our proposed approach yields a new feature set that allows us to significantly outperform the known results of data exhaust identification in consumer IoT traffic. The results show that the proposed approach achieves high performance in data exhaust detection of voice commands of users as well as the types of environments of smart home assistants while waiting or executing user commands.