A Unified Threat Intelligence Framework using FTLMS for Proactive IoT Security in Cloud Environments
T Kamalavalli, Barakkath Nisha U, Yasir Abdullah R.
Abstract
Adaptive Security remains a critical requirement as cloud-integrated Internet of Things deployments expose vast, fast-evolving attack surfaces that conventional intrusion-detection pipelines struggle to defend. This study introduces the Fortified Threat Learning and Mitigation System (FTLMS), a unified framework that fuses threat intelligence aggregation, hierarchical temporal modelling, and dynamic trust propagation to deliver proactive defence. At the core of FTLMS lies the Hierarchical Adaptive Threat Prediction Algorithm (HATPA), which forecasts risk trajectories through multi-stage recurrent embeddings and refines mitigation policies via reinforcement-driven feedback. The framework was implemented in MATLAB and evaluated against state-of-the-art baselines—LSTM-IDS, DNN-FEDSEC, and Federated SVM—over composite smart-grid, e-health, and industrial traffic traces. Experiments demonstrated that FTLMS-HATPA reduced average detection latency by 23 % and lifted attack-classification precision by 8 % while maintaining sub-second mitigation response, confirming its suitability for real-time, cloud-centric IoT security.