Corrections of the NIST Statistical Test Suite for Randomness

TL;DR

This paper identifies and corrects errors in the NIST statistical test suite, specifically in the Discrete Fourier Transform and Lempel-Ziv tests, indicating that previous randomness evaluations may need re-assessment.

Contribution

The authors provide four specific corrections to the test settings in the NIST suite, improving the accuracy of randomness testing.

Findings

Four corrections to the test settings are proposed.

Previous test results may be invalid due to these errors.

Re-evaluation of algorithms using the suite is recommended.

Abstract

It is well known that the NIST statistical test suite was used for the evaluation of AES candidate algorithms. We have found that the test setting of Discrete Fourier Transform test and Lempel-Ziv test of this test suite are wrong. We give four corrections of mistakes in the test settings. This suggests that re-evaluation of the test results should be needed.