| dc.description.abstract | Random Number Generator (RNG) is an important component in cryptography that is used to generate secret keys, encrypt messages, and support cryptographic protocols. Cryptographically Secure Random Number Generator (CSPRNG) is a PRNG algorithm designed for use in cryptography. One of the algorithms included in this CSPRNG is the Blum-Blum-Shub algorithm. However, this algorithm has drawbacks, such as being vulnerable to brute force attacks if the initial seed is known and vulnerable to dictionary attacks because it has a short period. This research proposes a modification of the Blum-Blum-Shub algorithm using multi-seed XOR and modulo order to overcome the shortcomings of the algorithm and generate random numbers that have good randomness and pattern stability and are resistant to predictability. The evaluation is done through probability tests, distribution tests, entropy tests, and NIST statistical tests. The research results show that the modified Blum-Blum-Shub algorithm is capable of generating random numbers with a probability that approaches the ideal balance between 0 bits and 1 bits, which amounts to 50%; a uniform distribution without a dominant pattern; an entropy value very close to the maximum value (15.9313 out of 16); and successfully passed 14 out of 15 NIST statistical tests (93.33%). In contrast, the classic Blum-Blum-Shub algorithm failed all testing parameters despite having significantly faster execution time. In conclusion, this modification significantly improves the quality of random number output, despite the slower execution time. The enhanced randomness quality makes the present algorithm a viable candidate for use as a CSPRNG in various cryptographic applications. | en_US |
