A general method to evaluate the correlation of randomness tests

ISCAS OpenIR

	A general method to evaluate the correlation of randomness tests
	Fan, Limin (1); Chen, Hua (1); Gao, Si (1); Fan, L.(fanlimin@tca.iscas.ac.cn)
	2014
会议名称	14th International Workshop on Information Security Applications, WISA 2013
页码	52-62
会议日期	August 19, 2013 - August 21, 2013
会议地点	Jeju Island, Korea, Republic of
收录类别	EI
出版地	Springer Verlag
ISSN	3029743
ISBN	9783319051482
部门归属	(1) Trusted Computing and Information Assurance Laboratory, Institute of Software, Chinese Academy of Sciences, 100190 Beijing, China
摘要	This paper discusses the correlation of the randomness tests. In this paper, we propose a new general method to evaluate the correlation of randomness tests. Firstly, we deduce the distribution that independent randomness tests should obey, then evaluate whether the randomness tests are independent or not based on hypothesis test. Using this method, we research the correlation of some statistical tests included in the NIST SP 800-22 suits, which is a collection of tests for the evaluation of both true random and pseudorandom number generators for cryptographic applications. Our experiment results show that some correlations of dependent exist among the randomness tests, which is different from the declaration that the randomness tests are independent by NIST. Moreover, the method we proposed also can be used in the study of parameter selection in randomness tests. © 2014 Springer International Publishing Switzerland.; This paper discusses the correlation of the randomness tests. In this paper, we propose a new general method to evaluate the correlation of randomness tests. Firstly, we deduce the distribution that independent randomness tests should obey, then evaluate whether the randomness tests are independent or not based on hypothesis test. Using this method, we research the correlation of some statistical tests included in the NIST SP 800-22 suits, which is a collection of tests for the evaluation of both true random and pseudorandom number generators for cryptographic applications. Our experiment results show that some correlations of dependent exist among the randomness tests, which is different from the declaration that the randomness tests are independent by NIST. Moreover, the method we proposed also can be used in the study of parameter selection in randomness tests. © 2014 Springer International Publishing Switzerland.
语种	英语
内容类型	会议论文
URI标识	http://ir.iscas.ac.cn/handle/311060/16628
专题	中国科学院软件研究所
通讯作者	Fan, L.(fanlimin@tca.iscas.ac.cn)
推荐引用方式 GB/T 7714	Fan, Limin ,Chen, Hua ,Gao, Si ,et al. A general method to evaluate the correlation of randomness tests[C]. Springer Verlag,2014:52-62.