| With the rapid growth of the informational society, people pay more and more attention to the information security.The random number is the fundation of information security and acts as the key role, for every aspect of security needs random numbers to work well.In applications of the network, people usually take it for granted that when needing a random number, the application systems always generate the numbers which meet the requirements. But in fact, it is just the ideal instance.In this dissertation, we focused on the random number generators and the methods to test the randomness of number sequence.Talking about the random number generators, we discussed the details of the generating algorithm of the Middle-Square Method, the Linear Congruential Method, the Chaos-mapping Method, ANSI x9.17, DSA, and the Intel Generator. Meanwhile, some weaknesses of them and some attacks are represented in this dissertation, especially, when these algorithms are implemented in computer system. After that, a new method based on the noise to generate a random number sequence is put forward. This method reconstructs the noise signal by choicing some high frequencies of the original signal. Of course, the original signal is decomposed by wavelet packet as low frequence and high frequencies. We used the Matlab toolbox to demonstrate the method.To test the randomness, we introduced the "A Statistical Test Suite for Random and Pseudorandom Number generators for Cryptographic Applications", which was published by the NIST (National Institute of Standard and Technology) and as the criterion to test the candidates of AES. Some improper tests are presented here. The test suite of DIEHARD and methods described by D.E.Knuth are also included in this paper, for they are deemed to the other popular test suites of the randomness.There are two new methods to test randomness in the dissertation.One is called "Exclusive-OR matrix" and the other is called "the coefficients of wavelet... |
PDF Full Text Request