Revolutionizing IoT Security: A Blockchain and Federated Learning-Based Anomaly Detection System
Haya Alharthi, Suhair Alshehri, Manal Kalkatawi · 2024
The rapid development of the Internet and smart devices has led to a surge in network traffic, increasing the complexity and vulnerability of infrastructure to malicious activities.Intrusion Detection Systems (IDS) are crucial in safeguarding the security and privacy of Internet of Things (IoT) devices.In IDS, Machine Learning techniques are often employed to identify deviations from normal behavior, enabling effective threat detection.However, traditional approaches face challenges such as centralized architectures, which raise data privacy concerns and limited resources on large-scale datasets.Federated Learning (FL) aligns well with a privacy-preserving decentralized learning method that does not transfer data but trains models locally.Only model parameters are transferred to the centralized server.The inherent risk of relying on a central server for model aggregation is mitigated by incorporating Blockchain technology, which provides a verifiable ledger for recording model updates and ensures secure aggregation without needing a central authority.This research introduces a novel conceptual framework for IDS integrating FL with Blockchain to enhance privacy, scalability, and resilience in IoT networks.