In an increasingly digitized society and economy, security in all kinds of electronic systems has become one of the biggest concerns, especially when resources are scarce (as with ‘wearable’ devices, where there are severe restrictions in area and energy). Responding to this problem with conventional cryptography approaches is not a viable alternative. To avoid this shortcoming, lightweight (i.e., adapted to limited resources) cryptographic solutions have been developed. Among these are the ones using the concept of ‘Physical Unclonable Function‘ (PUF), which typically exploit the intrinsic variability of CMOS manufacturing (time-zero variability or TZV) to have circuits with unique and unpredictable behavior, essential properties in secure cryptographic systems. These circuits usually have low fabrication costs (inarea) and low power consumption and can even be implemented with circuitry that is already being used for other purposes. However, effects such as Random Telegraph Noise (RTN) or aging, which introduce a time-dependent variability (TDV) component, can seriously compromise not only the reliability of the PUF itself, but, and as a result, the security of data and communications as well. It follows that TDV is, from this point of view, a phenomenon that should be palliated. In this Project, however, we intend to explore the completely opposite view: how to exploit the RTN having in mind that this TDV phenomenon, in the same way that TZV does, provides unique and unpredictable behavior to the circuit. Thus, the RTN-SECURE Project will study methods, techniques, and circuits to exploit the RTN for security, both for unique identifier generation and for random number generation. Furthermore, it will also address how to palliate the effect of circuit aging and TZV on these new implementations. Two pillars will sustain this project: (1) an accurate model of the TDV effects and (2) efficient simulation techniques using such model. Both will be developed in RTN-SECURE and both will allow to convincingly address the goal of exploiting RTN for more robust and efficient hardware security primitives.
