Complex network security using community structure and dynamical analysis: spectral clustering and VEIP-WQU model

Abstract The advancement of complex networks and systems offers a wide range of services, but security remains a significant challenge across different network types. Enhancing the security of complex networks requires a dual focus on both network structure and network dynamics. This paper explores...

Full description

Saved in:
Bibliographic Details
Main Authors: Morteza Jouyban, Soodeh Hosseini
Format: Article
Language:English
Published: SpringerOpen 2025-07-01
Series:Applied Network Science
Subjects:
Online Access:https://doi.org/10.1007/s41109-025-00717-8
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract The advancement of complex networks and systems offers a wide range of services, but security remains a significant challenge across different network types. Enhancing the security of complex networks requires a dual focus on both network structure and network dynamics. This paper explores the identification of highly interactive structures and the application of appropriate dynamic models to mitigate the spread of malware and cyberattacks, thereby enhancing network security. First, we leverage the spectral clustering algorithm to estimate the number of network communities and identify critical community structures. These high-density groups are vulnerable to the spread of infections, malware, and cyberattacks within the network. Second, we propose the VEIP-WQU (Vulnerable-Exposed-Infected-Patched with Warned-Quarantined-Immuned) model to identify and secure compromised nodes and help increase network security. Third, we investigate the impact of using the community structure and the proposed nonlinear mathematical model applied to network communities by analyzing the basic reproduction number and its effect on user dynamics and network security. We also evaluate the effectiveness and implementation of the proposed methodology as a security strategy through numerical simulations on six synthetic and real-world networks. The results demonstrate that the proposed approach effectively enhances network security, providing a robust framework for the development and enforcement of security protocols in complex networks.
ISSN:2364-8228