Side-Channel Resistance Evaluation Method using Statistical Tests for Physical Unclonable Function

Abstract

To obtain the internet of things (IoT) security, physical unclonable functions (PUFs) have attracted attention. Regarding hardware security, in recent years, the risk of side-channel attacks (SCAs) for PUF is pointed out. Therefore, countermeasures against SCAs have been proposed, and field programmable gate array (FPGA) implementation evaluations have also been reported. On the other hand, the evaluation of PUFs with countermeasures needs actual modeling attacks using many side-channel information; therefore, costs increase. This study proposes a new PUF security evaluation method, which does not need actual modeling attacks. The proposed method verifies the existence of side-channel leakages by applying statistical tests to measured power consumption waveforms during PUF operations. In experiments using an FPGA, by using the proposed method, it was confirmed that there were significant differences in the PUF without countermeasure. Experiments also showed that significant differences did not appear in the PUF with countermeasure and the proposed method could evaluate the PUF security easily without actual modeling attacks.