On-the-Fly Evaluation of FPGA-Based True Random Number Generator

Many embedded security chips require a high-quality digital True Random Number Generator (TRNG). Recently, some new TRNGs have been proposed in the literature, innovating by their new architectures. Moreover, some of them don't need to use the post-processing unit usually required in TRNG constructions. As a result, the TRNG data rate is enhanced and the produced random bits only depend on the noise source and its sampling. However, selecting a TRNG can be a delicate problem. In a hardware context (e.g. Field-Programmable Gate Array (FPGA) or Application-Specific Integrated Circuit (ASIC) implementation), the design area and power consumption are important criterions. To the best of our knowledge, no effective comparison of several TRNGs appears in the literature. This paper evaluates the randomness behavior, the area and the power consumption of the latest TRNGs. These investigations are realized into real conditions, by implementing the TRNGs into FPGA circuits.