A New Class of Digital Circuits for the Design of Entropy Sources in Programmable Logic

We propose a novel class of Digital Nonlinear Oscillators (DNOs) supporting complex dynamics, including chaos, suitable for the definition of high-performance and low-complexity entropy sources in Programmable Logic Devices (PLDs). We derive our proposal from the analysis of simplified models, investigated as non-autonomous nonlinear dynamical systems under different excitation conditions. The study lead the authors to the design of a fully digital entropy source consuming only two slices of a Xilinx FPGA, including post-processing, sufficient to define a class of TRNGs capable to pass the NIST standard tests for randomness in any worst case experimentally tested by the authors (6 chips, 96 generators). The solution has been compared with others published in the literature, confirming the validity of the proposal.