Cameron Foreman,Richie Yeung,Florian J. Curchod
Abstract:Random number generators (RNGs) are notoriously hard to build and test, especially in a cryptographic setting. Although one cannot conclusively determine the quality of an RNG by testing the statistical properties of its output alone, running numerical tests is both a powerful verification tool and the only universally applicable method. In this work, we present and make available a comprehensive statistical testing environment (STE) that is based on existing statistical test suites. The STE can be parameterised to run lightweight (i.e. fast) all the way to intensive testing, which goes far beyond what is required by certification bodies. With it, we benchmark the statistical properties of several RNGs, comparing them against each other. We then present and implement a variety of post-processing methods, in the form of randomness extractors, which improve the RNG's output quality under different sets of assumptions and analyse their impact through numerical testing with the STE.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper aims to solve the construction and testing problems of random number generators (RNGs), especially in the cryptographic environment. Specifically, the authors attempt to solve the following key issues:
1. **Quality assessment of random number generators**:
- The quality of random number generators (RNGs) is difficult to be fully verified by simple statistical tests. Nevertheless, running numerical tests is a powerful tool for verifying their output statistical properties and the only generally applicable method.
- The authors developed a comprehensive statistical test environment (STE), based on existing statistical test suites, and can adjust the test intensity as needed, from lightweight to high - intensity tests.
2. **Performance comparison of different random number generators**:
- Use the developed STE to benchmark multiple different random number generators (including pseudo - random number generators PRNG, true - random number generators TRNG, and quantum random number generators QRNG) to compare their statistical characteristics.
- Through these tests, it can be identified which RNGs perform well statistically and which have defects.
3. **Improving the output quality of random number generators**:
- Proposed multiple post - processing methods, namely randomness extractors, to improve the quality of RNG output. These methods are divided into four levels: deterministic extraction, seeded extraction, dual - source extraction, and physical randomness extraction.
- By using STE to analyze the impact of these post - processing methods on RNG output, evaluate their effectiveness.
4. **Addressing the limitations of existing test suites**:
- Existing statistical test suites (such as NIST, Dieharder, etc.) are powerful but complex to use and the output results are difficult to interpret. In addition, combining tests from different suites can conduct more in - depth statistical tests.
- The authors selected and parameterized specific statistical tests, creating a flexible and easy - to - use STE that can adjust the test intensity to different degrees, thus surpassing the capabilities of a single suite.
5. **Ensuring compliance with certification standards**:
- In cryptographic applications, many companies require RNGs to be certified by institutions such as the National Institute of Standards and Technology (NIST) or the Federal Office for Information Security (BSI). The authors' work helps to meet these certification requirements and provides a more stringent testing method.
### Summary
The core objective of the paper is to develop a flexible and powerful statistical test environment (STE) for evaluating and improving the output quality of random number generators. By comparing different types of RNGs and applying multiple randomness extraction methods, the authors show how to significantly improve the statistical performance of RNGs, thus providing strong support for cryptography and other fields that rely on high - quality randomness.