MODIFICATION OF DOT ALGORITHM ON PERMUTATION PROCESS TO IMPROVE SECURITY AGAINST DIFFERENTIAL CRYPTANALYSIS

• Main Author: Yuda Paripurna, Dian
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84845
• Institution: Institut Teknologi Bandung

The number of data breaches has increased significantly in recent times. Numerous cases have highlighted the importance of storing sensitive data in an encrypted state. Encryption can be implemented using block ciphers. Traditionally, block ciphers were designed to operate in environments with no limitation on computational resources, but the focus has now shifted towards lightweight block ciphers (LWBC). LWBCs are computationally more efficient as they process smaller data and key sizes compared to conventional block ciphers. Examples of LWBC algorithms include PRESENT, PUFFIN2, DoT, and ACT. In addition to providing a lightweight implementation, encryption algorithms must also ensure security. An LWBC is considered secure if it is proven to be resistant to cryptanalysis attacks. Although challenging and time-consuming, evaluating algorithms through cryptanalysis is a crucial step in assessing the security of an algorithm. An LWBC algorithm should at least demonstrate sufficient resistance against basic cryptanalysis methods, such as differential cryptanalysis (Chan et al., 2023). DoT is a lightweight block cipher with a 64-bit block size and a 128-bit key, claimed to be resistant to differential cryptanalysis with a probability of 2 ?70 (Patil dkk., 2019). However, in 2022, Kumar demonstrated that the maximum differential trail probability for DoT over 31 rounds is not 2 ?70 but 2 ?62 . Kumar (2022) subsequently proposed MDoT as a modification of DoT. Although this represents an improvement, the value still exceeds the security threshold of 2 ?128. Therefore, this research aims to further develop the DoT algorithm by replacing its permutation layer with a 64-bit permutation using PRESENT, PUFFIN, and ACT permutations. The research was conducted using the Design Science Research Methodology (DSRM), which includes designing a modified algorithm, analyzing the impact of algorithm components on differential cryptanalysis, reconstructing Kumar’s (2022) work as a validation step, and evaluating the security of the modified algorithm design against differential cryptanalysis. Based on the evaluation, the differential trail for 31 rounds in DoTPRE and DoTPUF was found to have a probability of 2 ?93 with a total of 31 active S-boxes. Meanwhile, the differential trail for 31 rounds in DoTACT had a probability of 2 ?133 with a total of 47 active S-boxes. Iterative patterns of one round were also found in the differential trails of DoTPRE and DoTPUF, while no such pattern was found in DoTACT. Based on these evaluation results, DoTACT provides the most significant v security improvement for DoT compared to the other two modifications. Therefore, DoTACT is recommended as a modification of the DoT algorithm that can enhance its security against differential cryptanalysis.