A Chaotic Cipher Mmohocc and Its Randomness Evaluation

Xiaowen Zhang, Ke Tang, Li ShuarXiv 2006 · OpenReview

TL;DR

This paper introduces Mmohocc, a novel chaotic stream cipher that leverages the inherent unpredictability and sensitivity of chaotic systems. By employing a unique "orbit-hopping" mechanism across multiple chaotic maps, Mmohocc generates keystreams that have rigorously passed two of the most stringent statistical randomness test suites (NIST and Diehard), demonstrating its potential for generating high-quality pseudorandom numbers.

What Problem Does This Paper Address?

Traditional cryptographic systems often rely on complex mathematical operations that can be computationally intensive. Chaotic dynamical systems, with their properties of mixing, unpredictability, and extreme sensitivity to initial conditions, offer an intriguing alternative for cryptography. However, a significant challenge in applying chaos to digital systems is that the desirable chaotic behaviors can degrade or disappear due to the discrete nature and limited resources of digital environments. This paper addresses this by proposing a method to accelerate these chaotic behaviors and then rigorously evaluating the randomness of the generated keystreams to ensure their cryptographic suitability.

Methods

The core of the Mmohocc cipher lies in its "orbit-hopping" mechanism. Instead of relying on a single chaotic map, Mmohocc utilizes multiple chaotic maps. A key scheduling process expands a given key into subkeys that control these maps. A "hopping mechanism," guided by a pseudorandom sequence (the hopping pattern), dictates how the cipher switches between different chaotic orbits generated by these maps. This hopping across multiple orbits is designed to achieve faster mixing and diffusion of information, mimicking the properties of a Vernam one-time pad.

For random number extraction, a chaotic orbit point (a real number) is first converted into a 32-bit integer. This integer is then split into four 8-bit integers. Two smaller pseudorandom integers are generated by XORing specific pairs of these 8-bit integers (u = a ⊕ c, v = b ⊕ d), further scrambling the bits and enhancing randomness.

To evaluate the quality of the generated keystreams, the authors employed two comprehensive batteries of statistical randomness tests: the NIST Suite and the Diehard Suite. These suites are considered highly stringent and include numerous independent tests designed to detect non-random patterns.

Datasets

The paper does not introduce new datasets for training. Instead, it focuses on evaluating the output of the Mmohocc cipher. For the NIST tests, the authors generated data files totaling over 1.19 billion bits, using different keys. These files were processed by the NIST suite, with 1000 bit sequences of 1 million random bits each being analyzed for each test. For the Diehard tests, 100 binary files, each 100 million bits long, were generated using 100 different keys.

Key Insights and Ideas

  • Accelerated Chaos for Cryptography: The paper demonstrates a practical approach to harness chaotic properties in digital cryptosystems by accelerating chaotic behaviors through techniques like orbit-hopping.
  • Orbit-Hopping for Enhanced Mixing: The novel "orbit-hopping" mechanism across multiple chaotic maps is a key innovation, designed to achieve faster and more robust confusion and diffusion than single-map chaotic systems.
  • Rigorous Statistical Validation: The commitment to using both the NIST and Diehard suites for randomness evaluation underscores the importance of thorough statistical testing for cryptographic primitives.
  • Statistical Success as a Foundation: While not a complete security analysis, passing these stringent tests is presented as a crucial foundational step, confirming the statistical quality of the generated pseudorandom numbers.

Key Results

The primary quantitative result is that the keystreams generated by the Mmohocc cipher successfully passed all tests within both the NIST Suite and the Diehard Suite.

  • NIST Suite: The paper presents detailed results for 16 NIST tests, showing mean p-values generally around 0.5 and variances around 0.08, indicating a uniform distribution of p-values. The proportions of passing sequences for each test fell within the acceptable range defined by NIST, confirming randomness with high confidence.
  • Diehard Suite: The results from the 18 Diehard tests also showed p-values that indicated randomness. For instance, a p-value larger than 0.01 and smaller than 0.99 signifies randomness with 99% confidence. All tests in the Diehard suite met this criterion.

Qualitatively, the authors conclude that Mmohocc can generate high-quality pseudorandom numbers from a statistical perspective.

Limitations and Future Work

  • Statistical vs. Cryptographic Security: The authors explicitly state that evaluating statistical characteristics of keystreams, while necessary, cannot replace a full security analysis for a stream cipher. The paper focuses solely on statistical randomness.
  • Parameter Tuning: The paper mentions that the statistical tests helped in improving adjustments for parameters and coefficients within the chaotic maps, hopping offsets, and hopping patterns. This suggests that the design process is iterative and relies on empirical validation.
  • Future Work: While not explicitly detailed, the successful statistical validation opens the door for further cryptographic analysis, including resistance to known attacks and performance optimization for practical implementation. The paper also hints at ongoing development of the Mmohocc cipher.

Summary updated

Abstract

After a brief introduction to a new chaotic stream cipher Mmohocc which utilizes the fundamental chaos characteristics of mixing, unpredictability, and sensitivity to initial conditions, we conducted the randomness statistical tests against the keystreams generated by the cipher. Two batteries of most stringent randomness tests, namely the NIST Suite and the Diehard Suite, were performed. The results showed that the keystreams have successfully passed all the statistical tests. We conclude that Mmohocc can generate high-quality pseudorandom numbers from a statistical point of view.
BibTeX
@misc{zhang2006chaotic,
  author = {Xiaowen Zhang and Ke Tang and Li Shu},
  title = {{A Chaotic Cipher Mmohocc and Its Randomness Evaluation}},
  year = {2006},
  eprint = {cs/0605066},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/cs/0605066},
}

Loading related content...

Figure