Justification of methods for calculating and analyzing the properties of pseudorandom and random sequences based on DNA

Abstract

Listen

Translate article

An integral requirement for modern information systems is to provide users with services such as confidentiality, integrity, availability, and irrefutability. The quality of such services directly depends on cryptographic transformations, an important component of most of which is randomness. Therefore, the generation of pseudorandom and random sequences is one of the most relevant and important tasks in cryptography. Such sequences are generated based on physical and non-physical noise sources. Our previous studies indicate the theoretical possibility to use DNA as a noise source and, accordingly, as a source of random sequences. Any noise source "contains randomness", i.e. it has a certain amount of entropy, but a sample from such a source will not always have good properties. That is why there is a need for tools that can obtain sequences with good randomness properties of samples from the DS sequences with good randomness properties of samples from the noise source (by, for example, some kind of enhancement). Such sequences should satisfy the necessary conditions: be statistically indistinguishable, uniform, etc. NIST-approved DRBG designs can be used to guarantee this. Such constructions are most often based on strong crypto-primitives, such as hashes, HMACs, or block or stream ciphers in the required modes. This work is devoted to newly developed methods for obtaining pseudorandom and random sequences based on DNA sequences using the national standard for block symmetric encryption DSTU 7624:2014 in the counter (CTR) operation mode, as well as methods for comparing sequences (both DNA and binary). The work essentially opens the topic of obtaining random sequences based on DNA, since previous studies of DNA in cryptography have focused on the use of DNA in encryption and steganography. The paper presents the results of solving such issues as the development of methods for obtaining DNA-based sequences, the evaluation of statistical and stochastic properties of such sequences, and the evaluation of similarity based on k-mer and MinHash distances. The results obtained indicate the prospects and relevance of further research in this area.