End-To-End Anomaly Detection of Service Function Chains in Cloud-Native Systems Using a Self-Guided Quantum Generative Adversarial Network

Abstract

Cloud-native technology enables Network Functions Virtualization (NFV) to dynamically provide and deploy network services in the Industrial Internet of Things (IIoT). However, compared to traditional hardware solutions, Service Function Chains (SFCs) are more vulnerable to faults in complex and dynamically changing cloud environments, requiring advanced anomaly detection techniques. Existing methods often struggle with accuracy, scalability, and efficiency in such environments, particularly due to high false positive rates (up to 15%) and poor adaptability to rapid scaling and latency-sensitive operations. This paper proposes a new Self-Guided Quantum Generative Adversarial Network with Puma Optimizer (SGQGANet-PO) for cloud-native SFC anomaly detection. The model benefits from the FullSight Dataset, beginning with Min-Max Normalization (MMN) for uniform feature scaling and Fast Pure Transformer Network (FPTN) for fast text feature extraction. SGQGANet-PO is based on quantum-inspired methods and is optimized with the Puma Optimizer to improve the robustness and convergence of the model. The method proposed has an accuracy of 99.76%, a precision of 99.5%, recall of 97.8%, F1-score of 98.6%, and an extremely low 0.24% error rate. The outcome shows better performance than other methods, providing a safe method for anomaly detection in cloud-native systems.