Side-channel attacks (SCAs) are powerful techniques used to recover keys from electronic devices by exploiting various physical leakages, such as power, timing, and heat. Although heat is one of the less frequently analyzed channels due to the high noise associated with thermal traces, it poses a significant and growing threat to the security of very large-scale integrated (VLSI) microsystems, particularly system in package (SiP) technologies. Thermal side-channel attacks (TSCAs) exploit temperature variations, risking not only hardware damage from excessive heat dissipation but also enabling the extraction of sensitive data, like cryptographic keys, by observing thermal patterns. This dual threat underscores the need for a synergistic approach to thermal management and security in designing integrated microsystems. In response, this paper presents a novel approach that improves the early detection of abnormal thermal fluctuations in SiP designs, preventing cybercriminals from exploiting such anomalies to extract sensitive information for malicious purposes. Our approach employs a new concept called Thermal Digital Twin (TDT), which integrates two previously separate methods and techniques, resulting in successful outcomes. It combines the gradient direction sensor scan (GDSSCAN) to capture thermal data from the physical field programmable gate array (FPGA), which guarantees rapid thermal scan with a measurement period that could be close to 10 μs, a resolution of 0.5 ∘C, and a temperature range from −40 ∘C to 140 ∘C; once the data are transmitted in real time to a Digital Twin created in COMSOL Multiphysics® 6.0 for simulation using the Finite Element Method (FEM), the real time required by the CPU to perform all the necessary calculations can extend to several seconds or minutes. This integration allows for a detailed analysis of thermal transfer within the SiP model of our FPGA. Implementation and simulations demonstrate that the Thermal Digital Twin (TDT) approach could reduce the risks associated with TSCA by a significant percentage, thereby enhancing the security of FPGA systems against thermal threats.
