Blockchain assisted clustering with Intrusion Detection System for Industrial Internet of Things environment. (30th November 2022)

• Record Type: Journal Article
• Authors: Mansour, Romany F.
• Is Part Of: Expert systems with applications. Volume 207(2022)
• Publication Date: 2022-11-30
• Subjects: Industrial IoT -- Clustering -- Intrusion detection system -- Deep learning -- Blockchain -- Cluster Heads; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2022.117995
• Languages: English
• ISSNs: 0957-4174

Abstract

Abstract: Industry 4.0 has the potential to create new industrial revolution by offering intelligent, secure, independent, and self-adaptive Industrial Internet of Things (IIoT) networks. These IIoT nodes can be grouped via the clustering process into clusters to enhance networks' efficiency, lifetime, and scalability. However, it offers several benefits but still security is a primary concern, which can be resolved by Intrusion Detection Systems (IDS) and blockchain technology. Therefore, an effective Blockchain-Assisted Cluster-based Intrusion Detection System for IIoT, called the BAC-IDS technique has been devised. The proposed BAC-IDS model aims to cluster the IIoT devices for detecting the intrusions and enabling blockchain-based secure data transmission. BAC-IDS technique involves Harris Hawks Optimization (HHO)-based clustering technique to choose the Cluster Heads (CH) effectively and construct clusters accordingly. In addition, Chicken Swarm Optimization with Gated Recurrent Unit-based intrusion detection is also followed to determine the presence of intrusions in IIoT environment. Moreover, blockchain technology enables secure data transmission from Cluster Members (CMs) to CHs. The design of HHO clustering and CSO-based hyperparameter optimization for the IIoT environment represents the novelty of the proposed work. A wide range of simulations are performed under different aspects. From experimental results, it has been observed that the BAC-IDS approach has gained a TNR (99.97%), AUC (99.96%), precision (99.99%), recall (99.97%), accuracy (99.98%), F-score (99.98%) and error rate (0.03) on NSL-KDD2015 dataset. Further, on CICIDS 2017 dataset, it has obtained TNR (99.96%), AUC (99.98%), precision (99.96%), recall (99.98%), accuracy (99.97%), F-score (99.96%), and error rate (0.04). The experimental results highlight the supremacy of the proposed BAC-IDS technique over current state-of-art techniques.