Article contents
Intelligent Anomaly Detection for Complex Cloud Systems: A Deep Learning Framework
Abstract
This article investigates the application of deep learning approaches for anomaly detection in complex distributed cloud environments. Traditional rule-based monitoring systems face significant limitations in modern cloud infrastructures characterized by massive scale, heterogeneity, concept drift, and cross-organizational dependencies. The article explores how Long Short-Term Memory (LSTM) autoencoders and Transformer models can effectively analyze time-series telemetry and log data, respectively, providing superior anomaly detection capabilities. LSTM autoencoders demonstrate exceptional performance in processing numerical metrics and capturing temporal dependencies across multiple time scales, while Transformer architectures excel at analyzing textual log data through their self-attention mechanism. The article further presents a hierarchical, distributed architecture for implementing these models at scale, incorporating edge preprocessing, specialized regional processing nodes, continuous model evaluation, and federated learning. This comprehensive article enables real-time anomaly detection with improved accuracy, reduced latency, and enhanced operational efficiency while respecting data sovereignty requirements.
Article information
Journal
Journal of Computer Science and Technology Studies
Volume (Issue)
7 (8)
Pages
526-532
Published
Copyright
Open access

This work is licensed under a Creative Commons Attribution 4.0 International License.