Prediction of variation aware FOSC in ring oscillators (ROs) to mitigate the impact of aging on RO-PUF

Abstract

We propose a methodology to predict device-level variability (including aging) impact on the oscillation frequency (Fosc) of an N-stage ring oscillator (RO). This task is accomplished by creating a tool in a Python environment that uses our own developed variability-aware timing models of the CMOS inverter. Moreover, we use the model to foretell the impact of aging on the logical effort (g) of a CMOS inverter. Using the modified g, we resize ROs in an RO-based physical unclonable function (PUF) in the pre-layout stage to mitigate the impact of aging on the reliability of RO-PUF. The simulation is performed in the Cadence AMS environment using STMicroelectronics (STM) 28 nm CMOS process technology. With a one-time SPICE/aging simulation, the proposed methodology eliminates SPICE/aging simulation overhead for the prediction of variability impact on Fosc of a given N-stage RO. This approach mitigates the impact of aging on the reliability of RO-PUF and provides a method for variability (including aging) aware design in the pre-layout stage.