Information Theory Based Evaluation of the RC4 Stream Cipher Outputs.

Area of Science:

• Cryptography
• Information Theory
• Computer Science

Background:

• The RC4 stream cipher is widely used but has known vulnerabilities.
• Assessing the security of cryptographic algorithms is crucial.
• Information theory provides tools for quantifying information flow.

Purpose of the Study:

• To develop a criterion for measuring information leakage from RC4 outputs.
• To establish a vulnerability metric for RC4 based on its internal state and output sequences.
• To identify the most vulnerable RC4 outputs susceptible to iterative probabilistic attacks.

Main Methods:

• Utilizing information theory principles to estimate entropy.
• Calculating the sum of maximum plausible estimates of conditional entropies H(jt|zt).
• Analyzing probability distributions P(jt|zt) of output and internal state sequences.

Main Results:

• A novel test statistic based on entropy estimation is presented.
• Experimental validation demonstrates the statistic's ability to pinpoint vulnerable RC4 outputs.
• The proposed metric effectively quantifies RC4 output sequence vulnerability.

Conclusions:

• The information-theoretic criterion offers a robust method for RC4 vulnerability assessment.
• The developed metric can guide cryptanalysis efforts against RC4.
• This approach enhances understanding of RC4's security properties under probabilistic attacks.