Known-plaintext attack and improvement of PRNG-based text encryption

Abstract

Cryptanalysis combines a set of techniques and heuristics to reveal existing security flaws in a certain cryptographic scheme, for the sake of improvement or sometimes violation. In this paper, we analyze a newly proposed text cryptosystem that uses three different schemes for a pseudorandom number generator (PRNG). We show that two of the suggested methods are not secure against known-plaintext attacks. Moreover, we propose an improved symmetric-stream encryption method based on PRNG. We implemented the proposed method in MATLAB and evaluated it for various types of attacks and different sensitivity tests. The improved method was designed to overcome the security issues in the other methods. Especially the resiliency to known-plaintext attack. Furthermore, the results of the sensitivity tests show excellent level of randomness ranging from 40% to 50%.