Deep Learning-Driven Optimized Approaches for Network Anomaly Detection in IoT-Enabled Cloud Ecosystems: A Comprehensive Review

Abstract

The rapid proliferation of Internet of Things (IoT) devices within cloud environments has introduced unprecedented challenges in securing network infrastructures against anomalies and cyber threats. Traditional detection mechanisms often struggle to meet the dynamic and complex demands of these integrated ecosystems. This review paper focuses on the potential of deep learning (DL)-based optimized models for effective network anomaly detection in IoT-enabled cloud environments. It examines the fundamental role of DL techniques in addressing key challenges, including scalability, adaptability, and real-time threat identification. The paper systematically explores state-of-the-art models, highlighting their architectures, optimization strategies, and performance metrics. A comparative analysis is provided to underscore strengths, limitations, and suitability across diverse use cases. Furthermore, emerging trends, such as lightweight DL models and federated learning, are discussed in the context of resource-constrained IoT networks. The review aims to offer researchers and practitioners insights into current advancements while identifying gaps and future directions for research in enhancing security and reliability in IoT-cloud ecosystems.

This review highlights the role of deep learning in detecting network anomalies in IoT-integrated cloud environments, focusing on optimization strategies to handle challenges like scalability, heterogeneity, and real-time detection. We provide a concise review of existing approaches and optimization methods, identify challenges, and suggest directions for future research.