A Novel Identity Authentication for FPGA Based IP Designs

Le résumé fourni par la source

With the rapid development of Internet-of-Things (IoT), hardware security becomes an important issue. Once a chip is attacked, the authentication always faces serious credibility crisis. In this work, a Hausdorff distance model-based scheme is proposed to authenticate circuit identity under IoT environment. Firstly, the structure of FPGA is analyzed and the positions of LUTs are abstracted into a collection of reconfigurable nodes. The depth-first searching algorithm is used to determine the unused LUT resources for inserting the copyright information. The Hausdorff distance matching function is utilized to reorder the random positions. These positions will be further mapped to satisfy specific constraints of the optimal watermark positions. As the identity of circuit requires real-time authentication, virtual positions are mapped to initial key file. Once decrypted, the identity can be authenticated by the mapping relationship of Hausdorff distance function. Lastly, we conduct a series of experiments to evaluate the ability of the proposed scheme against illegal attacks, such as removal attacks. Experiments show that, the proposed scheme has good randomness and secrecy. Besides, the resource overhead during authentication is encouraging.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.