Statistical analysis on enhanced 3D-AES block cipher cryptographic algorithm / Nor Azeala Mohd Yusof

Randomness test is one of the measurement techniques which have been taken into consideration in the evaluation of the minimum security requirement of the block cipher algorithm. A non-random block cipher seems to be vulnerable to any type of attack. Many algorithms such as AES, LED, KTANTAN, KATAN, L-Block, SIMON, SPECK, RECTANGLE, and GRAIN-128 have performed randomness test using NIST Statistical Test Suite. Therefore, this research aims to analyse the randomness of 3D-AES, a local SPN-based block cipher algorithm. Specifically, the randomness test is performed using the NIST Statistical Test Suite which consists of 15 statistical tests. The output data sequences are generated from seven different data categories. Unfortunately, the failed test results for Cipher Block Chaining Mode (CBCM), Strict Key Avalanche (SKA), High Density Key (HDK), and Random Plaintext Random Key (RPRK) data categories indicate that 3D-AES produced non-random output binary sequences. The major failure is on the SKA data category which is used to evaluate the avalanche effect. On this basis, the enhancement on 3D-AES is proposed to achieve the ability to generate a random number and the modified version of 3D-AES named Enhanced 3D-AES. Two new functions, ConfuseK and ConfuseP have been injected into the 3D-AES. ConfuseK is a process of XORing the key element with its corresponding position number, whereas ConfuseP is a process of XORing the plaintext with its corresponding position number. These two new functions are based on the confusion method proposed by Shannon’s theory since 1945. The randomness of Enhanced 3D-AES then is re-evaluated and it has passed all 15 statistical tests for all seven data categories. These research findings portray the effectiveness of the modification that has been done towards the initial version of 3D-AES. It can be concluded that Enhanced 3D-AES meets the standard security requirements for block cipher design. It is strongly suggested to implement Enhanced 3D-AES in software applications to secure data transmission.