Reproducibility of Firmware Analysis: An Empirical Study

Firmware analysis methods are crucial for IoT security, yet their reproducibility-the ability to replicate results in subsequent research-has not been thoroughly examined. This study addresses this gap by empirically analyzing the reproducibility of three methods in two key applications of firmware analysis: third-party library identification and binary image base determination. We then evaluate the original studies on each of these methods, using two reproducibility assessment techniques, providing insights into the challenges and opportunities related to reproducibility in firmware analysis. Our findings highlight the current reproducibility status of these methods and offer guidance for improving the reliability of research in this field.