Design And Analysis Of Digital True Random Number Generators

With the advent of the era of big data,the problem of information security and personal data leakage have become increasingly prominent.These bring to people's life great security risks and a lot of inconvenience.Cryptography provides effective security protection for these problems.High-quality random numbers play an important role in security systems.The digital true random number generators extract physical random phenomenon in the circuit.The unpredictability of the physical phenomenon guarantees the quality of the random numbers and provide strong security protection for cryptosystem.This paper discusses the entropy mechanism of multiple random number generators.Two new types of digital true random number generators are proposed.Based on the different entropy sources,the output characteristics are analyzed by different methods.On the one hand,this paper first analyzes mechanism and physical stochastic model of the entropy source based on transition effect ring oscillator circuit.In this paper,we implement the transition effect ring oscillator on FPGA and verify the physical stochastic model given by the proposers.Afterwards,a novel true random number generator based on the classical ring oscillator is proposed,which is based on the periodic restart mechanism in the transition effect ring oscillator.A physical stochastic model is given to estimate the entropy of the output and the design parameters required for enough entropy are given.The true random number generator is implemented on FPGA and the internal random numbers have passed AIS31 test published by German company BSI.At last,a scheme to improve the entropy rate for traditional entropy sources such as classical ring oscillator is presented.On the other hand,the research of chaos theory opens up a new direction for the design of true random number generators and provides new design ideas.According to the design requirements of chaotic Boolean networks,this paper proposes a novel true random number generator based on chaotic Boolean networks,analyzes the chaotic entropy mechanism and presents the sampling and post-processing circuit.The entropy source is simulated and the result shows that piecewise linear differential equation model is closer to the actual situation.The modeling can serve as reference while implementing the circuit in ASIC.The new chaotic true random number generator is implemented on multiple FPGAs.The output characteristics such as power spectral density,autocorrelation and Lyapunov exponent of the entropy source are analyzed.By comparing the output of the entropy source proposed in this paper with the same type of entropy source proposed by other authors,the result indicates that the entropy source proposed in this paper has stable oscillation characteristics and the network dynamics do not change with the number of nodes.At last,the randomness of the output is tested by SP800-22 standards published by National Institute of Standards and Technology,GM/T0005-2012 standards presented by the State Encryption Administration and AIS31 standards.The test results indicate that the output of the novel chaotic true random number generator has excellent statistical characteristics.The random numbers can be used in cryptosystem.The two novel digital real random number generators proposed in this paper are implemented on the development board.The analysis methods are given.This paper provides theoretical and experimental reference for the design of digital true random number generators.