How to crack a Linear Congruential Generator如何破解线性同余发生器.pdfVIP

How to crack a Linear Congruential Generator Haldir[RET] 22nd December 2004 1 Introduction The Linear Congruential Generator (LCG) is a common, but not secure way to generate random numbers for a given range. Definition 1 : xn = axn−1 + k1 modulo m for all n ≥ 1 and x0 = k0 Most common Pseudo Number Generators (PRNG) implemented in standard libraries use the LCG, so I presented in [1] a method to solve the linear Congruential Generator using recursive equations. The downside of that method was that you couldn’t easily find the modulus used, either you knew it in advance or you had to try a few values, but you just needed three consecutive values to solve it. In this paper I will present a method which will solve all values of the LCG including the modulus with six or more consecutive numbers of PRNG output. 2 Mathematical Background In [2] George Marsaglia analyzed Pseudo Random Number Generators. He found the flaw, that LCGs are mainly “falling into planes“, this means that any three consecutive integers (x, y, z) with multiplier a fall on the lattice of points generated by all linear combinations of the three points 2 (1, a, a ), (0, m, 0), (0, 0, m). Also any point (x,y) falls on the lattice generated by (1,a),(0,m). See [2] for more properties of this flaw. We use this to solve all three parameters of the LCG We take a list of consecutive integers: 207560540, 956631177, 2037688522, 1509348670, 1546336451, 429714088, 217250280 Four integers from the list form a 3x3 matrix:  207560540 956631177 1   956631177 2037688522 1   2037688522 1509348670 1  Thanks to the flaws listed above and in the references, the determinant of this matrix will be an intege