DRV-Fingerprinting: Using Data Retention Voltage of SRAM Cells for Chip Identification

Physical unclonable functions (PUFs) produce outputs that are a function of minute random physical variations. Promoted for low-cost authentication and resistance to counterfeiting, many varieties of PUFs have been used to enhance the security and privacy of RFID tags. To different extents, applications for both identification and authentication require a PUF to produce a consistent output over time. As the sensing of minute variations is a fundamentally noisy process, much effort is spent on error correction of PUF outputs. We propose a new variant of PUF that uses well-understood properties of common memory cells as a fingerprint. Our method of fingerprinting SRAM cells by their data retention voltage improves the success rate of identification by 28% over fingerprints based on power-up state.