A 2 parameter, 2 state Markov model of an RNG (Random Number Generator) can generate all possible combinations of \(\mu \) (mean), SCC (Serial Correlation Coefficient) in binary data. A formula for Min entropy of data from the model is derived, leading to an efficient online hardware entropy boundary test of non IID (Independent and Identically Distributed) data called the polygon test. In addition the model can produce labelled data with known entropy with which to validate entropy test algorithms. Significant shortcomings in the SP800-90B Non IID tests are shown and an example of using the generator to test tests.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Online Testing Entropy and Entropy Tests with a Two State Markov Model

  • David Johnston

摘要

A 2 parameter, 2 state Markov model of an RNG (Random Number Generator) can generate all possible combinations of \(\mu \) (mean), SCC (Serial Correlation Coefficient) in binary data. A formula for Min entropy of data from the model is derived, leading to an efficient online hardware entropy boundary test of non IID (Independent and Identically Distributed) data called the polygon test. In addition the model can produce labelled data with known entropy with which to validate entropy test algorithms. Significant shortcomings in the SP800-90B Non IID tests are shown and an example of using the generator to test tests.