Low-power and Computing-free Privacy Design for IoT Systems

Published in IEEE Computer Society Annual Symposium on VLSI (ISVLSI), 2024

This work is the first to leverage the tendency of hardware transistors to degrade under voltage drop as a source of noise for Local Differential Privacy (LDP). We propose SRAM_DP, a hardware-software co-design for achieving differential privacy on semi-structured data. Experiments on a custom SRAM design, using the WHO-released disease dataset, show that the method preserves individual privacy while retaining meaningful data distribution analysis.