RiverIoT - a Framework Proposal for Fuzzing IoT Applications

This paper presents an integrated testing framework for Internet of Things (IoT) systems based on the open-source platform RIVER. Our objective is to leverage the existing methods for guided fuzzing using AI techniques to test end-to-end IoT deployed applications. The framework’s architecture is defined first, along with its interface to the user application. Then, the paper presents the technical details for leveraging the concolic execution methods in testing IoT applications. The testing architecture is modeled as a graph of connected processes deployed on physical devices. The methods used extend the capability of testing individual processes in isolation. The proposed new framework enables the user to test the interaction of processes as well as dynamic connections that could appear during the deployed application at runtime.