An Iterative Systematic Analytical Review of Modern Intrusion Detection Systems

• Version
• Download 11
• File Size 1.04 MB
• File Count 1
• Create Date 18 June, 2025
• Last Updated 18 June, 2025

Authors : Rahul V. Bambodkar, Amitabh Wahi, Ganesh Khekare

DOI : 10.46335/IJIES.2025.10.8.19

Abstract: With the increasing dependence on cloud computing and Internet of Things (IoT) environments, data integrity, confidentiality, and availability become significantly larger issues. The reviews of the existing IDSs have often failed to consider the optimization techniques, scalability, privacy preservation, and adaptability of such systems to real-world threats. Moreover, numerous reviews do not synthesize the relative performance of state-of-art ML and DL models designed for cloud IDS applications. Addressed those gaps as there is a comprehensive review regarding an advance methodology for the optimization-driven, IDS. Various ranges of methods are included SHO-DESNID; CIDF VAWGAN-GOA, and REPO Stack, that are optimized for novel anomaly detection using innovative      techniques such as Seahorse Optimization and Archerfish Hunting Optimizer, respectively. Models based on federated learning, such as LS2DNN with PBKA and blockchain-based architectures, like SecFedIDM-V1, are critically analysed for their privacy-preserving capabilities and their scalability. Besides, continuous learning frameworks like HFIN and synthetic data generation models such as CDAAE + CDAEE-KNN  are reviewed in terms of their effectiveness in dealing with dynamic and rare cyber threats. The results show that hybrid approaches combining ML, DL, and optimization techniques outperform traditional approaches with an accuracy of up to 99.9% and lower computational overhead. This review offers actionable insights into the design of robust, scalable, and adaptive IDS frameworks for diverse applications, ranging from cloud environments to IoT and IIoT landscapes. This work contributes to advancing cybersecurity solutions, identifying optimal models for specific scenarios that close critical gaps in the research process of cloud IDS.