Evaluation with NIST Statistical Test on Pseudorandom Number Generators based on DMP-80 and DMP-128

The random numbers possess an essential role in cryptography as one can use them as keys, or to generate challenges, or even as IV values. One can generate random numbers from a pseudorandom number generator (PRNG). There are 2 categories of PRNG algorithms known to us, including the specially built algorithms and the algorithms based on existing cryptographic construction. One of the algorithms commonly used to generate output from a PRNG is a block cipher-based hash function. In this research, we evaluated a PRNG based on block cipher-based hash function by utilizing the NIST test. A certain data set was used to fit the required sample as a hash function produces outputs with fixed length. We used Davies-Meyer scheme based on PRESENT-80 and PRESENT-128 block cipher algorithms. From the 48 samples tested, there were 17 samples that failed the NIST test. Out of the 17 samples that failed, 12 samples failed in the non-overlapping template test.