Central attacks in complex networks: A revisit with new fallback strategy

• Published in: Physica A Vol. 549; p. 124347
• Main Authors: Lekha, Divya Sindhu, Balakrishnan, Kannan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2020
• Subjects: Attack vulnerability; Average shortest path length; Centrality; Closeness centrality; Complex networks; Attack vulnerability; Complex networks; Centrality; Closeness centrality; Average shortest path length
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2020.124347

In this work, we revisit central attacks in complex networks. We simulate simultaneous as well as sequential attacks on networks based on degree (DC), betweenness (BC) and closeness (CC) centralities. We observed the disintegration of giant components and updates in average geodesic distance, in order to assess the vulnerability of networks. There is ample literature depicting the high severity of BC and DC attacks. However, this severity is calculated based on the size of the giant component in residue networks. We show that CC attacks are also equally hazardous when the severity is calculated based on average geodesic distance. This confirms the relevance of CC in central attacks. Attacks on nodes with high BC and DC can lead to faster collapse of the network system. So, protecting these critical nodes can improve the robustness of system. If they are protected, then an attacker needs to find other targets which can impart an equal / near-to equal destruction to the network. We propose a fallback strategy (a plan B) to attack a network in such a protected environment. We show that this strategy is the best approach when the prominent nodes are protected. •We revisit three different kinds of centrality-based attacks on complex networks.•We estimate severity of attack based on the updates in average geodesic distance.•We observe that attacks based on closeness centrality are relevant.•We propose a fallback strategy to attack a protected network.