A Highly Efficient Profiled Power Analysis Attack Based on Power Leakage Fitting
Yuanzhi Li, Shanshan Xu, Yuling Luo, Sheng Qin, Shunsheng Zhang, Min Su · 2021
Power Analysis Attack (PAA) is one of the most effective Side Channel Attack (SCA) methods with good effect. A highly efficient profiled power analysis attack method named PLF is proposed in this paper. Nine average power traces are calculated according to the Hamming Weight (HW), and the Points of Interest (POIs) are extracted. Then the template is established by linear fitting. The information hidden in the attack traces is mapped to the points in the two-dimensional coordinate system, then the correct key is cracked by calculating the distance between the points and the line. The numerical problems caused by the irreversible covariance matrix and exponential operation of Template Attack (TA) are avoided in the PLF. The PLF is verified by attacking the Advanced Encryption Standard (AES) on the Atmel XMEGA128-D4 microcontroller. The experimental results show that the computational complexity is reduced in PLF compared to TA and Correlation Power Analysis (CPA), and the slope of PLA is 95.5% and 55.1% lower than TA and CPA respectively. Besides, the PLF requires fewer power traces to effectively crack the correct key, and it has a lower Partial Guessing Entropy (PGE). Moreover, the distinguishing level of PLF is 51% and 30% higher than TA and CPA respectively. Meanwhile, PLF only needs 17 attack traces to achieve 80% Partial Success Rate (PSR), which is 57.5% and 77.9% lower than CPA and TA in the number of attack traces, respectively.