Blockchain-driven machine learning-enabled intrusion-resilient authenticated key agreement protocol for edge-centric IoT systems

Abstract

Abstract The edge computing-based Internet of Things (IoT) system minimizes latency by processing data locally, reducing the distance it needs to travel. Processing data in proximity to its source enables rapid decision-making and real-time reactions. The edge-based IoT has several possible uses, including smart cities, smart healthcare, industrial automation & processing, smart farming, and many more. In this paper, we propose a blockchain-driven machine learning-enabled intrusion-resilient authenticated key agreement scheme for edge-centric IoT systems (in short, BMAS-EIoT), which is equipped with the features of authentication, key management, and machine learning-based intrusion detection. In BMAS-EIoT, we provide the network and threat models to enhance comprehension of the organization and deployment of devices and systems, as well as the potential threats to the system. BMAS-EIoT has been observed to possess protection against a variety of potential attacks during the security investigation. Moreover, it has been observed that BMAS-EIoT outperforms other present schemes in terms of performance comparison. A practical implementation of BMAS-EIoT is provided to evaluate the effectiveness of its key components, including intrusion detection and blockchain implementation. Furthermore, BMAS-EIoT possesses supplementary noteworthy capabilities and enhanced security attributes.