Efficient Data Encryption for Securing HDFS Using DQN-Enhanced Deep Reinforcement Learning Techniques

In the era of big data, ensuring the security of large-scale distributed storage systems like the Hadoop Distributed File System (HDFS) is critical. Traditional encryption methods often struggle to balance robust security with system performance, leading to vulnerabilities and inefficiencies. This study presents the design and implementation of an efficient data encryption algorithm for securing HDFS using Deep Q-Network (DQN) enhanced Deep Reinforcement Learning (DRL) techniques. The proposed model dynamically optimizes encryption parameters by leveraging the adaptive capabilities of DQN, ensuring robust security while maintaining high system performance and scalability. Our approach addresses key limitations of existing encryption methods by integrating DQN with deep reinforcement learning to create a dynamic encryption framework that adjusts in real time based on data access patterns and threat levels. The results demonstrate significant improvements in security and efficiency compared to conventional encryption techniques. Specifically, the DQN-enhanced DRL algorithm consistently outperformed baseline methods regarding encryption strength, computational efficiency, latency, resource utilization, adaptability, and energy consumption. The contributions of this research include the development of a novel DQN-enhanced encryption algorithm tailored for HDFS, the creation of an adaptive encryption framework that leverages real-time data dynamics, and a thorough evaluation demonstrating the practical benefits of the proposed solution. This study paves the way for future research in intelligent encryption systems, offering a robust and efficient approach to securing large-scale distributed storage environments. Our findings underscore the potential of integrating advanced machine learning techniques into encryption processes to enhance security and performance, addressing the complex challenges modern data storage systems pose.