Research And Design Of True Random Number Generator Based On Biphasic Sampling

With the rapid development of information technology,the problem of information security has penetrated into every corner of the society.National security,national defense,private life and personal property protection are inseparable from a strong and secure information security system.Information encryption and decryption technology,which plays an important role in information security system,has become the focus of scientific research in related fields.Cipher is the core input of encryption and decryption algorithms,and its quality as well as the performance of those algorithms,directly affect the security of information security systems.Therefore,random numbers,often used as cipher,is regarded as the key of encryption and decryption technology-related activities.In the current encryption system,deterministic mathematical algorithm is often used to generate pseudo-random numbers.Most of these type of random numbers have property of limited unpredictability,which will become more fragile even disappear with the increase of computing resources and improvement of cracking ability.Therefore random numbers with higher quality are required to ensure encryption system in working order.Non-repeatable and unpredictable random numbers are called true random numbers,which are obtained by transforming random physical phenomena.The typical hardware true random number generators used to generate true random numbers can be classified into four types accroding to its entropy source extraction scheme,including: 1.Direct-amplified noise based;2.Discrete-time chaos based;3.Oscillator sampling based;4.Metastability based.This dissertation mainly focuses on true random number generator based on oscillator sampling.Firstly,several factors that affect the performance of entropy source extraction block are sorted out: the ratio of low-frequency signal jitter standard deviation to high-frequency signal period,the sampling edge position and the duty cycle of high-frequency signal.By analyzing how high frequency signal duty cycle influence the output of entropy source extraction block,a biphasic oscillator sampling architecture is proposed,which can effectively reduce the high sensitivity of the uniphasic sampling architecture to the change of duty cycle according to simulation and calculation results.On the basis of the previous chapters,a biphasic oscillator sampling architecture-based true random number generator,of which the entropy extraction block is constructed by a low-frequency oscillator based on schmitt trigger and a new type of the coupling ring oscillator,is designed,as well as the low supply voltage bandgap refference circuit.Simulation data shows that,under the supply voltage of 1.2V,the designed true random number generator is in working order.The core circuit consumes 220.08?W and covers chip area of 0.007mm2,with a throughput rate of 6.8Mbps.Meanwhile,the output of our design can pass NIST SP 800-22 statistiscal test suite and autocorrelation test.