Embedded Systems (ES) underlie society's critical cyberinfrastructure and comprise the vast majority of consumer electronics, making them a prized target for dangerous malware and hardware Trojans. Malicious intrusion into these systems present a threat to national security and economic stability as globalized supply chains and tight network integration make ES more susceptible to attack than ever. High-end ES like the Xilinx Zynq-7020 system on a chip are widely used in the field and provide a representative platform for investigating the methods of cybercriminals. This research suggests a novel anomaly detection framework that could be used to detect potential zero-day exploits, undiscovered rootkits, or even maliciously implanted hardware by leveraging the Zynq architecture and real-time device-level measurements of thermal side-channels. The results of an initial investigation showed different processor workloads produce distinct thermal fingerprints that are detectable by out-of-band, digital logic-based thermal sensors.
(Publisher abstract provided.)