Font Size: a A A

On The Linear Cryptanalysis Of DES

Posted on:2006-12-30Degree:MasterType:Thesis
Country:ChinaCandidate:S J YanFull Text:PDF
GTID:2178360182960495Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
The Date Encryption Standard (DES) of America in 20th century is one of the block ciphers which have been used and discussed extensively up to now. Linear cryptanalysis method and differential cryptanalysis method are the main attacks to DES, and we study linear cryptanalysis of DES systematically in this paper. We analyse the relationship between the the number of efficient key-bits and the success rate, then present an improved algorithm which could increase success rate efficiently. At the same time, the mean of success rate was generalized, and new formulas for calculating success rate were presented.Linear cryptanalysis with multiple approximations is a generalization of linear cryptanalysis, and how to attack DES efficiently by multiple linear cryptanalysis method is still a problem. In this paper we give another two good 14-round linear approximations of DES. Their correlation coefficients are 0.6 times and 0.8 times of the best linear approximation's respectively, and they involve the same key-bits and different plaintext-bits. We make use of them by multiple linear cryptanalysis method, present an algorithm and analyze its success rate theoretically. By combining the two good linear approximations with the best linear approximations we present improved algorithms which could increase the attacking speed.When we make use of linear cryptanalysis method, finding the best linear approximation of the encrypting algorithm is very important. Taking advantage of simulate annealing algorithm, we present an evolutionary searching algorithm to search the best linear approximation of DES. This algorithm is applicable to block ciphers with Feistel constructure, and it would still be very applicable when the length of block increases.
Keywords/Search Tags:DES, linear cryptanalysis, multiple linear approximations, best linear approximation, simulate annealing algorithm
PDF Full Text Request
Related items