Intelligent Hybrid Encryption Selection: An AI-Driven Approach for Optimizing Security and Performance

Abstract

Hybrid encryption plays a critical role in enhancing data security and performance by combining the strengths of symmetric and asymmetric encryption techniques. This research proposes an AI-driven approach to selecting the most efficient hybrid encryption pair based on various parameters. The study is divided into four phases. In the first phase, three file types (text, binary, backup) of varying sizes (100KB, 100MB, 1GB) are generated as input data. The second phase applies nine different hybrid encryption combinations, including AES, CHACHA20, and DES with RSA, ECC, and DSA, to each file type and size. The third phase captures key encryption metrics such as file size, encryption combination, and total encryption time, storing the results in a structured dataset for further analysis. In the final phase, AI classification models are integrated to evaluate the collected data and predict the optimal hybrid encryption pair based on efficiency and performance. The results of this study aim to automate the encryption selection process, ensuring both enhanced security and computational efficiency for various file types and sizes.