A Digital Random Number Generator Based on Irregular Sampling of Regular Waveform

There is an increasing demand for the security of information over the past few decades. Random Number Generators (RNGs) play a significant role in many cryptographic applications for generating unpredictable bit-streams. Despite the fact that Application Specific Integrated Circuit (ASIC) implementation of RNGs are usually preferred for their high performance, fully digital circuits implemented on Field Programmable Gate Array (FPGA) platforms are more attractive for designers considering their ability of rapid-reconfigurable prototyping, and their easy integration to other digital circuits. In this study, a new improvement based on irregular sampling of regular waveform is proposed for ring oscillator based RNGs. It is subjected to National Institute of Standards and Technology (NIST) 800-22 test suite and it passes full tests without any post-processing. An attack circuit is also implemented to see the robustness of the designed RNG against coupling-based attacks. No effective correlation can be obtained between the outputs of the target and the attack circuit; in spite of, their extremely close placement to each other inside the FPGA chip. The feasibility of the proposed method is given in experimental results.