In-Depth Review and Comparative Analysis of DRAM-Based PUFs

Dynamic Random-Access Memory Physical Unclonable Functions (DRAM PUFs) are gaining interest in hardware security, particularly for resource-constrained IoT devices such as smart sensors in the era of rapid digitalization. Since DRAM is present in most of the embedded devices, implementing DRAM PUFs does not require a lot of additional hardware components. In addition, DRAM consumes less power than SRAM and DRAM’s size is generally larger, which leads to larger challenge-response pair (CRP) space, enhancing security and robustness. Despite their potential, DRAM PUFs come with some limitations in device authentication. This paper reviews three types of DRAM PUFs, including start-up DRAM PUF, retention-based DRAM PUF, and latency-based DRAM PUF. It explores their various compositions and examines the individual characteristics of each type. Latency-based DRAM PUFs have been identified as one of the highly promising and reliable PUFs for authentication. We then analyze the research gap in this area to identify opportunities for improvement and provide insights that can help shape future research directions.