Standard encryption (AES) currently used all over the world such as mobile phones.When this AES was selected in 2001, the random number evaluation test NIST SP 800-22 was used as an evaluation tool.However, it was announced in 2003 that one of them, the "discrete Fourier test test" (abbreviated as DFT test), was theoretically wrong.
Since then, many institutions and researchers around the world have pursued the correct DFT test and made numerous revision proposals.However, these amendments are based on the assumption that a certain "pseudo-random number is a good random number", and the accuracy of the reference distribution does not depend on the evaluation that assumes the randomness of the random number to be evaluated. There was no complete amendment proposal that could prove mathematically independently.
Under these circumstances, a research group at Kyoto University has released a modified version of the DFT test that completely solves this problem.Unlike the approach that has been taken in the research methods so far, in which the reference distribution is approximately obtained under the assumption that "○○ random numbers are perfect random numbers", the correction is made by an approach based on a completely new idea. I tried.The result is an accurate DFT test with the correct reference distribution mathematically and independently, without the need for any assumptions.
This test can be directly applied to the evaluation of all cryptographic evaluations and the randomness of random numbers.Furthermore, in the selection of next-generation standard cryptography, which will be the successor to AES, more accurate randomness is required, so it is expected to be used as an important standard random number evaluation tool at that time.In the future, the research group will investigate and discuss the dependency of the results of this research with other random number test results, and aim to construct an effective random number evaluation system and standardize randomness.
Paper information:[IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences] Randomness Test to Solve Discrete Fourier Transform Test Problems
