DevFing: Robust LCR Based Device Fingerprinting
Vijay Kumar, Kolin Paul · 2021
Ahstract- The continuous growth of IoT technologies raises several security concerns and challenges. Device identification and authentication are the key challenges due to the significant presence of malicious/counterfeited devices. The traditional approaches used to identify counterfeited devices, e.g., digital signatures, hardware watermarking, or techniques based on IPV4/IPV6, fail to be directly applied to the resource-constrained IoT end-nodes. The advancement in sophisticated counterfeiter tools and technologies makes these problems even more daunting. In this direction, we propose DevFing - A Device Fingerprinting approach based on Hardware Intrinsic Security (HIS) mechanism. DevFing employs an electronic board's intrinsic characteristics such that Inductance (L), Capacitance (C), Resistance (R), and Impedance (Z) to identify each device uniquely. We propose a dedicated daughterboard design to measure and generate robust device fingerprints from these intrinsic properties. The board's error- correcting module enhances the robustness by utilizing a fuzzy extractor for error mitigation and regeneration. We have carried out extensive experiments and simulations on the embedded system boards (Raspberry Pi-3B) and desktop to validate DevFing. Our approach improved the fingerprint matching accuracy for the 64 bits key from 44.2% to 100% with a fuzzy extractor/error- correctina module with the hamming distance of 5.