True Random Number Generator for Reliable Hardware Security Modules Based on a Neuromorphic Variation-Tolerant Spintronic Structure

The generation of true random numbers is one of the most important tasks in a hardware security module (HSM), particularly for cryptography applications. The stochastic behavior of electronic devices can be used to generate random numbers. In this paper, a reliable neuromorphic true random number generator (TRNG) relying on stochastic switching of the magnetic tunnel junction (MTJ) in the subcritical current regime is proposed. Thanks to the efficient structure of the proposed design as well as the fascinating features of the MTJs and carbon nanotube field-effect transistors (CNTFET), the proposed TRNG consumes low power. Moreover, the neuromorphic structure of the proposed design leads to variation tolerance and guarantees the truly random number generation even in the presence of the process variations. HSPICE simulations verify the functionality of the proposed TRNG. Furthermore, by considering the corners of the fabrication process, the randomness of the bitstream, generated by the proposed TRNG, is validated by the statistical randomness test provided by the U.S National Institute of Standards and Technology (NIST).