HFTEA: A HIERARCHICAL FEDERATED TRUST-AWARE EDGE COMPUTING ARCHITECTURE FOR SECURE AND SCALABLE IOT SYSTEMS

Abstract

The convergence of the Internet of Things (IoT) and edge computing has created transformative opportunities for latency-sensitive, data-intensive applications. However, heterogeneous device ecosystems introduce profound security vulnerabilities and formidable scalability constraints that existing architectures inadequately address. This paper proposes HFTEA, a Hierarchical Federated Trust-Aware Edge Computing Architecture that addresses these challenges through four novel contributions: (i) a three-tier dynamic hierarchy comprising device, fog, and macro-edge layers with adaptive workload migration governed by a context-aware scheduling policy; (ii) a Federated Learning-based Intrusion Detection System (FL-IDS) that preserves data privacy while achieving distributed anomaly detection without centralizing sensitive IoT telemetry; (iii) a Lightweight Zero-Trust Adaptation Protocol (LZTAP) engineered for resource-constrained devices operating under intermittent connectivity; and (iv) a Blockchain-Anchored Trust Scoring Mechanism (BATSM) that maintains tamper-proof reputation ledgers for edge nodes using a DAG-based lightweight consensus. Simulation experiments on NS-3 with 1,200 heterogeneous IoT nodes demonstrate that HFTEA achieves a 94.7% intrusion detection accuracy with only 3.2% false positives, reduces end-to-end latency by 41% compared to centralized cloud architectures, and sustains scalable operation up to 10,000 concurrent devices with sub-linear overhead growth. Security analysis confirms resistance against seven critical attack vectors including Byzantine fault injection, Sybil attacks, and gradient poisoning in federated rounds.