Improved multidimensional zero-correlation linear cryptanalysis and applications to LBlock and TWINE

ISCAS OpenIR

	Improved multidimensional zero-correlation linear cryptanalysis and applications to LBlock and TWINE
	Wang, Yanfeng (1); Wu, Wenling (1)
	2014
会议名称	19th Australasian Conference on Information Security and Privacy, ACISP 2014
页码	1-16
会议日期	July 7, 2014 - July 9, 2014
会议地点	Wollongong, NSW, Australia
收录类别	EI
出版地	Springer Verlag
ISSN	3029743
ISBN	9783319083438
部门归属	(1) Trusted Computing and Information Assurance Laboratory, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China; (2) State Key Laboratory of Computer Science, Institute of Software, Chinese Academy of Sciences, Beijing 100190, China; (3) Graduate University, Chinese Academy of Sciences, Beijing 100049, China
摘要	Zero-correlation linear cryptanalysis is a new method based on the linear approximations with correlation zero. In this paper, we propose a new model of multidimensional zero-correlation linear cryptanalysis by taking the equivalent relations of round keys into consideration. The improved attack model first finds out all the longest multidimensional zero-correlation linear distinguishers, then regards the distinguishers with the least independent guessed keys as the optimal distinguishers and finally chooses one optimal distinguisher to recover the secret key of cipher by using the partial-compression technique. Based on the improved attack model, we extend the original 22-round zero-correlation linear attack on LBlock and first evaluate the security of TWINE against the zero-correlation linear cryptanalysis. There are at least 8×8 classes of multidimensional zero-correlation linear distinguishers for 14-round LBlock and TWINE. After determining the corresponding optimal distinguisher, we carefully choose the order of guessing keys and guess each subkey nibble one after another to achieve an attack on 23-round LBlock, an attack on 23-round TWINE-80 and another attack on 25-round TWINE-128. As far as we know, these results are the currently best results on LBlock and TWINE in the single key scenario except the optimized brute force attack. © 2014 Springer International Publishing Switzerland.; Zero-correlation linear cryptanalysis is a new method based on the linear approximations with correlation zero. In this paper, we propose a new model of multidimensional zero-correlation linear cryptanalysis by taking the equivalent relations of round keys into consideration. The improved attack model first finds out all the longest multidimensional zero-correlation linear distinguishers, then regards the distinguishers with the least independent guessed keys as the optimal distinguishers and finally chooses one optimal distinguisher to recover the secret key of cipher by using the partial-compression technique. Based on the improved attack model, we extend the original 22-round zero-correlation linear attack on LBlock and first evaluate the security of TWINE against the zero-correlation linear cryptanalysis. There are at least 8×8 classes of multidimensional zero-correlation linear distinguishers for 14-round LBlock and TWINE. After determining the corresponding optimal distinguisher, we carefully choose the order of guessing keys and guess each subkey nibble one after another to achieve an attack on 23-round LBlock, an attack on 23-round TWINE-80 and another attack on 25-round TWINE-128. As far as we know, these results are the currently best results on LBlock and TWINE in the single key scenario except the optimized brute force attack. © 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/311060/16619
专题	中国科学院软件研究所
推荐引用方式 GB/T 7714	Wang, Yanfeng ,Wu, Wenling . Improved multidimensional zero-correlation linear cryptanalysis and applications to LBlock and TWINE[C]. Springer Verlag,2014:1-16.