Tlade+: Intelligent Adaptive Encryption Using Reinforcement Learning For Secure Cloud Services
Keywords:
Cloud Security, TLADE+, QoS, Encryption, AES, ChaCha20, Decision Tree, Latency.Abstract
Cloud computing can offer efficient data storage and sharing solutions. However, securing sensitive data is a major concern. Traditional encryption methods lack adaptability and can cause extra overhead and poor security. Our solution is the Intelligent Adaptive Encryption system called TLADE+ (Time-oriented Latency Approximation-based Data Encryption Plus). Q-learning and Decision Tree techniques are used to determine appropriate encryption levels. The Q-learning agent takes into account how sensitive the file is, the file size, system latency, CPU Usage and threat level. Based on this it will decide whether to apply LOW, MEDIUM or HIGH levels of encryption. Rules have been set up so that if the file is very sensitive, or the threat levels are high, then LOW encryption can never be applied. After encryption of the file a reward will be calculated based on how secure, quick and how it fits within the systems requirements. This reward will be used to update the Q-table. The system also adapts by storing key features of the file and what encryption label was chosen. If there are enough entries with valid data then the Decision Tree can be retrained. The system also includes a Flask authentication system, OTP verification, encrypting files, storing them onto IPFS and viewing the history of the file. This system allows intelligent adaptive protection of files.

