基于28 nm RRAM的可重构真随机数发生器  

28 nm RRAM-based reconfigurable true random number generator

作　　者：宋长坤 郑彩萍 陈铖颖 SONG Changkun;ZHENG Caiping;CHEN Chengying(School of Opto-electronic and Communication Engineering,Xiamen University of Technology,Xiamen 361024,China)

机构地区：[1]厦门理工学院光电与通信工程学院,福建厦门361024

出　　处：《浙江大学学报（工学版）》2024年第7期1516-1523,共8页Journal of Zhejiang University：Engineering Science

基　　金：福建省自然科学基金引导性项目(2023H0052);厦门市重大科技项目(3502Z20221022)。

摘　　要：基于阻变存储器(RRAM)电流饥饿型环形振荡器的真随机数发生器(TRNG)方案,提出交替读写的操作模式,优化熵源重构机制.针对现有RRAM在多操作周期下跨导线性化问题,提出熵可配置电阻窗口钳位电路.通过减小电阻窗口获得RRAM非线性的最大化,所提电路能够有效避免RRAM出现过度置位、过度复位的现象,保证熵源稳定性.基于UMC 28 nm HKMG工艺对TRNG进行流片.输出数据统计性测试结果通过了NIST SP800-22所有测试集的真随机数标准测试.检测结果表明,在高斯分布的95%置信区间,所有统计数据的自相关函数值均落在-0.003~0.003,输出序列具有良好的随机性.An alternating read and write mode was proposed based on the current-starved ring oscillator(CSRO)true random number generator(TRNG)with resistance random access memory(RRAM),and the entropy source reconstruction scheme was optimized.An entropy configurable resistance window clamping circuit(ECRWCC)was proposed to address the trans conductor linearization issues of RRAM under multiple operating cycles.By reducing the resistance window to maximize the nonlinearity of RRAM,the phenomenon of over-set and over-reset was effectively avoided in ECRWCC,and the stability of the entropy source was ensured.The TRNG was tapeout with a UMC 28 nm HKMG process.The statistical test results of the output data passed the true random number standard test for all test sets of NIST SP800-22.The results showed that within the 95%confidence interval of the Gaussian distribution,the autocorrelation function,of all statistics were in the range of−0.003 and 0.003,and the output sequence had good randomness.

关 键 词：阻变存储器(RRAM) 真随机数发生器(TRNG) 熵源 电流饥饿型环形振荡器 跨导线性 

分 类 号：TN453[电子电信—微电子学与固体电子学]