Research of Statistical Properties of Stream Symmetric Ciphers

In recent years, numerous studies of stream symmetric ciphers in Ukraine are continuing, the main purpose of which is to argue the principles of creating a new cryptographic algorithm, which can be based on the national standard. One of the essential aspects in choosing from many alternatives is the statistical properties of the output pseudorandom sequence (key stream). In this paper, the results of comparative studies of statistical properties of output sequences, which are formed by various stream ciphers, in particular, by world-known algorithms Enocoro, Decim, Grain, HC, MUGI, Mickey, Rabbit, RC-4, Salsa20, SNOW2.0, Sosemanuk, Trivium and the Ukrainian cryptographic algorithm Strumok, that was developed in recent years, are presented. For comparative studies, the NIST STS method was used, according to which experimental studies are performed in 15 statistical tests, the purpose of which is to determine the randomness of the output binary sequences. Each of the tests is aimed at studying certain vulnerabilities of the generator, that is, points to the potential usage of different methods of cryptographic analysis. Although each of the considered streaming encryption algorithms has been studied, we have carried out a statistical test of the generated pseudorandom sequences under equal conditions and with identical initial parameters, that is, our results allow us to perform a comparative analysis of ciphers and to justify the best of statistical properties. The estimates presented in the article, as expected, confirmed the high statistical security indexes of modern ciphers. In addition, according to the results of experimental research, it was found that the new Ukrainian development - the stream cipher Strumok does not yield to the best world algorithms in the statistical properties of the initial sequences.