Federated Learning-Based Model to Lightweight IDSs for Heterogeneous IoT Networks: State-of-the-Art, Challenges, and Future Directions

Abstract

A large number of Internet of Things (IoT) devices have been deployed in numerous applications (e.g., smart homes, healthcare, smart grids, manufacturing processes, and product supply chains). However, IoT networks’ wide range and heterogeneity make them prone to cyberattacks. Most IoT devices have limited resource capabilities (e.g., memory capacity, processing power, and energy consumption) to function as conventional intrusion detection systems (IDSs). Many research approaches to lightweight IDSs have been taken, namely energy-based IDSs, machine learning/deep learning (ML/DL)–based IDSs, and federated learning (FL)–based IDSs. FL has become a promising solution for IDSs in IoT networks because it reduces overhead in the learning process by engaging IoT devices during the training process. In this paper, we present a comprehensive survey on FL for IDSs in an IoT environment with resource-constrained devices. We investigate the existing studies of FL in different architectures, namely centralized (client-server), decentralized (device-to-device), and semi-decentralized. The study’s findings highlight the necessity for enhancing the FL framework to better suit IoT networks. This enhancement is crucial, particularly in addressing two key challenges: the need to lightweight FL client’s models to accommodate the resource constraints of IoT devices and having a design aggregation algorithm capable of effectively handling the heterogeneity and limited resources inherent in IoT devices. Finally, we discuss the open challenges and future directions for scientists and researchers interested in FL-based IDS for IoT environments.

Keywords

• Internet of Things
• intrusion detection system
• anomaly detection
• machine learning
• deep learning
• federated learning