An Efficient Post-Processing Method for SSL-PUF in MEC Security Authentication

Abstract

Computation offloading is a key technology in mobile edge computing (MEC) that addresses the performance and energy constraints faced by mobile devices when handling computationally intensive tasks. Identity authentication for computation offloading is a critical issue as it ensures data security and user identity legitimate verification. Semiconductor superlattice physical unclonable functions (SSL-PUFs) are unique physical characteristics based on semiconductor superlattice materials, which can be used for secure authentication and encrypted communication in edge computing with wide applications in security authentication. However, the adoption of SSL-PUF in computation offloading for MEC applications faces two practical challenges: insufficient alignment accuracy of response signals and poor stability of SSL-PUF response signals. To address these two issues, an efficient post-processing algorithm specifically designed for SSL-PUF has been proposed. This algorithm consists of two steps. The first step involves aligning the re-sponse signals of SSL-PUF using a sequence alignment algorithm based on preset sequence, which significantly reduces the intra-chip Hamming distance of SSL-PUF. Then, a data fusion algorithm combining time majority voting mechanism is used to filter out erroneous response data, thereby improving the accuracy of SSL-PUF response signals. Experimental results demonstrate that after applying the proposed post-processing algorithm, the randomness of the signals remains largely unaffected. The maximum bit error rate of SSL-PUF response signals is reduced by 34.33%, and the average intra-chip Hamming distance decreases from 12% to 4.9%. The reliability of SSL-PUF is significantly enhanced, making it promising for secure identity authentication in mobile edge computing.