A New Entropy Source Design Based on NAND-XOR Ring Oscillators for Resource-Efficient and Ultra-High Throughput TRNG

抄録

True random number generator (TRNG) is the important hardware security primitive for modern internet of things (IoT) devices, while the entropy source (ES) serves as the most crucial component for TRNG. This paper explores the NAND-XOR ring oscillators (NXROs) structure to design a novel ES architecture for TRNG. The basic principle of the ES is to add a self-feedback NAND gate in XOR RO to generate a high-frequency signal, so as to apply the signal to induce a high-frequency change of XOR RO oscillation states and thus to achieve a more significant amplification for random clock jitter. In addition, our NXRO has a higher oscillation frequency and min-entropy than traditional ROs. We implement a TRNG with the new NXRO ES unit on both Xilinx Spartan-6 and Atrix-7 FPGAs. Our experiment results demonstrate that compared with the state-of-the-art TRNGs, the new TRNG achieves a higher throughput and lower hardware overhead in generating true random numbers successfully passing the NIST test and AIS31 test without post-processing.