Effects of Runtime Generated S-Boxes to the DES Model

Abstract

The Data Encryption Standard (DES) is a cryptographic algorithm designed using various static tables. Among these static tables are substitution boxes, generally called S-Boxes. S-Boxes are viewed as an important aspect in block ciphers because cryptographically strong S-Boxes imply stronger security. This paper investigates the effects of introducing dynamic S-Boxes in the DES cipher which are generated at run time without following any specific order. These S-Boxes would alter the entries of the original S-Boxes, and would be used in any dynamic sequence. We present a method of populating S-Boxes based on the concept of logistic maps in chaos theory. That way, great qualities of dynamism are attained. Additionally, we exploit in-built functions in Java to randomize number generation before encryption and decryption. It will also be labeled in order to prevent it from being picked again. We tested and analyzed the performance of the new algorithm against the benchmark original DES. We tested and analyzed the performance of the new algorithm against the benchmark original DES. The results yielded were categorized into four, namely; CPU usage, memory usage, number of classes, and number of threads. Generally, the new DES out-performed the old DES. Initially we present the background and statement of the problem, then our findings from literature followed by our proposed methodology and finally the results and analysis which showed encryption and decryption will still be successful despite having dynamic S-Boxes instead of the static ones. Our work makes a computational contribution to the DES design, introducing dynamic runtime generated S-Boxes whose sequence and entries are also dynamic.