Font Size: a A A

Design Of High-performance Lowcost True Random Number Generator Based On Ring Resonator

Posted on:2021-01-15Degree:MasterType:Thesis
Country:ChinaCandidate:P WangFull Text:PDF
GTID:2518306557990219Subject:IC Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of the information society,communication security is becoming more and more important,which requires authentication and confidential transmission of data.Random number based cryptography can be used to meet these security applications.True Random Number Generator(TRNG)based on integrated circuit is required in So C solution for cryptographic technology,and the TRNG based on the ring oscillator entropy source can be implemented with pure digital logic cells,which is easy to implement in FPGA and semi-custom ASIC platforms.However,this type of entropy source generally has problems such as low throughput,large hardware cost and lack of a suitable evaluation strategy for formal security evaluation.This thesis proposed an entropy source based on Asynchronous feedback unit(AFU)and a TRNG evaluation strategy based on a stochastic model and on-line monitoring from the aspects of entropy source structure design and security evaluation strategy.Chaotic oscillation behavior caused by the random flipping time of the logic gate and the random delay in the asynchronous feedback ring oscillator and the metastable state of the sampling logic are utilized in the AFU-based entropy source to realize a pure digital entropy source,greatly reducing the jitter propagation delay and hardware cost under the condition of ensuring the randomness of the entropy source.The asynchronous feedback unit can be used as a general structural unit based on the ring oscillator entropy source,at the same time,the corresponding random Boolean network model is established to provide theoretical guidance for the selection of circuit feedback coefficients in the AFU-based entropy source,making the entropy source to be more reliably transplanted to a pure digital logic hardware implementation platform.Simultaneously,this thesis established a corresponding random Boolean network model to provide theoretical guidance for the selection of circuit feedback coefficients of the AFU-based entropy source.The asynchronous feedback unit can be used as a general structural unit of the entropy source based on ring oscillator,and can be more reliably transplanted to a pure digital logic hardware implementation platform.The TRNG evaluation strategy based on the stochastic model and online monitoring evaluates the unpredictability of TRNG with the minimum Shannon entropy to perform formal security evaluation on the original output sequence of TRNG,and a low cost uniformity monitoring logic is implemented on-chip to continuously monitor the uniformity of the original output sequence of the TRNG to ensure the normal work of TRNG.Finally,in this thesis,the proposed TRNG design based on AFU entropy source structure and online uniformity monitoring logic and corresponding random sequence acquisition platform are implemented on Xilinx Artix-7 FPGA.During the experiment,we collects the data to the PC end through a USB serial port,and uses Matlab for data analysis and processing.Compared with the entropy source based on the matrix feedback ring oscillator(MFRO)of the same array size,the proposed AFU-based entropy source reduces the hardware overhead by 37.5%and reduces the jitter feedback time by 38.2%.The stochastic model in the TRNG evaluation strategy takes platform parameters and design parameters as input,and outputs the minimum Shannon entropy to provide a formal security proof of the TRNG original output sequence.The hardware cost of the uniformity online monitoring logic is 51 LUTs.The TRNG proposed in this thesis passed all NIST test sets without any post processing,reaching a baud rate of at least 300Mbps(continuous mode)and 100Mbps(restart mode).Finally,this thesis completes the ASIC physical implementation of the TRNG integrated with AFU entropy source structure,uniformity online monitoring logic and AMBA APB interface protocol on the TSMC 40nm process,and the area cost is 3175.45 um~2(where the hardware cost of the entropy source is 64.42 um~2).
Keywords/Search Tags:true random number generator, entropy source, asynchronous feedback, chaos, stochastic model
PDF Full Text Request
Related items