Quality-driven Design Methodology for PUFs in FPGAs for Secure IoT

Xiangyun Wang1, Yicheng Song1, Katyayani Prakash1, Zeljko Zilic1, Tomas Langsetmo2
1McGill University, 2KNOX Industries Inc.


Abstract

The emergence of Internet of Things (IoT) has placed strong emphasis on security. Physical Unclonable Functions (PUFs) are a promising low cost security solution for IoT. Since Field-Programmable Gate Arrays (FPGAs) are more trusted than custom Integrated Circuits (ICs) in building secure systems, implementing PUFs in FPGAs is appealing. In this paper, we propose a methodology for improving the quality of delay-based PUFs on FPGAs. With our method, an asymmetric delay-based PUF has been predictably designed for temperature- and aging-independent operation. Furthermore, the resistance to modeling attacks is built into the design.