Towards Enhancing the Robustness of Scale-Free IoT Networks by an Intelligent Rewiring Mechanism

Sensors

Sheraz Aslam¹  Ahmad Taher Azar¹  Zahoor Ali Khan²  Nadeem Javaid³  Syed Minhal Abbas³  Umar Qasim⁴ 

[1] Automated Systems & Soft Computing Lab (ASSCL), Prince Sultan University, Riyadh 12435, Saudi Arabia;Computer Information Science, Higher Colleges of Technology, Fujairah 4114, United Arab Emirates;Department of Computer Science, COMSATS University Islamabad, Islamabad 44000, Pakistan;Department of Computer Science, University of Engineering and Technology Lahore (New Campus), Lahore 54000, Pakistan;

关键词: scale-free IoT networks;    centrality measures;    edge rewiring;    malicious attacks;    network optimization;    random attacks;   

DOI  :  10.3390/s22072658

来源: DOAJ

【 摘 要 】

The enhancement of Robustness (R) has gained significant importance in Scale-Free Networks (SFNs) over the past few years. SFNs are resilient to Random Attacks (RAs). However, these networks are prone to Malicious Attacks (MAs). This study aims to construct a robust network against MAs. An Intelligent Rewiring (INTR) mechanism is proposed to optimize the network R against MAs. In this mechanism, edge rewiring is performed between the high and low degree nodes to make a robust network. The Closeness Centrality (CC) measure is utilized to determine the central nodes in the network. Based on the measure, MAs are performed on nodes to damage the network. Therefore, the connections of the neighboring nodes in the network are greatly affected by removing the central nodes. To analyze the network connectivity against the removal of nodes, the performance of CC is found to be more efficient in terms of computational time as compared to Betweenness Centrality (BC) and Eigenvector Centrality (EC). In addition, the Recalculated High Degree based Link Attacks (RHDLA) and the High Degree based Link Attacks (HDLA) are performed to affect the network connectivity. Using the local information of SFN, these attacks damage the vital portion of the network. The INTR outperforms Simulated Annealing (SA) and ROSE in terms of R by 17.8% and 10.7%, respectively. During the rewiring mechanism, the distribution of nodes’ degrees remains constant.

【 授权许可】

Unknown