Securing Dynamic Firmware Updates of Mixed-Critical Applications

This work introduces a secure framework for runtime updating of firmware in Internet of Things devices that execute mixed-critical applications. Taking advantage of the capabilities of modern heterogeneous System-on-Chip devices to run cores in asymmetric multiprocessing (AMP) configuration, we developed a methodology to showcase dynamic updating of real-time applications in a novel secure way when executing on a Xilinx ZYNQ-based platform. As an exemplary implementation we demonstrate a bio-signal monitoring use case that reads accelerometer data to determine if a person has fallen, while a distant medical management system can dynamically perform firmware updates. Even sophisticated code injection or reuse attacks can be subverted with the proposed defenses that ensure a practically isolated environment for the critical firmware with negligible overhead to the device in terms of performance and cost.