Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dynamic Management of a Deep Learning-Based Anomaly Detection System for 5G Networks

Lorenzo Fernández Maimó, Alberto Huertas Celdrán, Manuel Gil Pérez, Félix J. García Clemente, Gregorio Martínez Pérez

TL;DR

The paper addresses the challenge of real-time, autonomic anomaly detection in 5G networks by proposing a MEC-oriented architecture that integrates deep learning detectors with policy-driven management of edge resources. It introduces a three-function Network Anomaly Detection framework (Flow Collection, Anomaly Symptom Detection, Network Anomaly Detection) and three policy families (Virtual Infrastructure, Anomaly Detection Function, Mobile Edge Application) to dynamically adapt compute resources, update models, and deploy auxiliary tools like DPI. Through a realistic 5G use case and extensive experiments, the work demonstrates how CPU- and GPU-accelerated detectors can operate at edge scales, with NAD improving generalization to unseen threats and policies enabling proactive resource provisioning. The approach has practical significance for securing decentralized 5G services by enabling low-latency, autonomous defense at the network edge, with future work targeting real 5G data and broader SELFNET integration.

Abstract

Fog and mobile edge computing (MEC) will play a key role in the upcoming fifth generation (5G) mobile networks to support decentralized applications, data analytics and management into the network itself by using a highly distributed compute model. Furthermore, increasing attention is paid to providing user-centric cybersecurity solutions, which particularly require collecting, processing and analyzing significantly large amount of data traffic and huge number of network connections in 5G networks. In this regard, this paper proposes a MEC-oriented solution in 5G mobile networks to detect network anomalies in real-time and in autonomic way. Our proposal uses deep learning techniques to analyze network flows and to detect network anomalies. Moreover, it uses policies in order to provide an efficient and dynamic management system of the computing resources used in the anomaly detection process. The paper presents relevant aspects of the deployment of the proposal and experimental results to show its performance.

Dynamic Management of a Deep Learning-Based Anomaly Detection System for 5G Networks

TL;DR

The paper addresses the challenge of real-time, autonomic anomaly detection in 5G networks by proposing a MEC-oriented architecture that integrates deep learning detectors with policy-driven management of edge resources. It introduces a three-function Network Anomaly Detection framework (Flow Collection, Anomaly Symptom Detection, Network Anomaly Detection) and three policy families (Virtual Infrastructure, Anomaly Detection Function, Mobile Edge Application) to dynamically adapt compute resources, update models, and deploy auxiliary tools like DPI. Through a realistic 5G use case and extensive experiments, the work demonstrates how CPU- and GPU-accelerated detectors can operate at edge scales, with NAD improving generalization to unseen threats and policies enabling proactive resource provisioning. The approach has practical significance for securing decentralized 5G services by enabling low-latency, autonomous defense at the network edge, with future work targeting real 5G data and broader SELFNET integration.

Abstract

Fog and mobile edge computing (MEC) will play a key role in the upcoming fifth generation (5G) mobile networks to support decentralized applications, data analytics and management into the network itself by using a highly distributed compute model. Furthermore, increasing attention is paid to providing user-centric cybersecurity solutions, which particularly require collecting, processing and analyzing significantly large amount of data traffic and huge number of network connections in 5G networks. In this regard, this paper proposes a MEC-oriented solution in 5G mobile networks to detect network anomalies in real-time and in autonomic way. Our proposal uses deep learning techniques to analyze network flows and to detect network anomalies. Moreover, it uses policies in order to provide an efficient and dynamic management system of the computing resources used in the anomaly detection process. The paper presents relevant aspects of the deployment of the proposal and experimental results to show its performance.
Paper Structure (15 sections, 4 equations, 8 figures, 3 tables)

This paper contains 15 sections, 4 equations, 8 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. MEC-oriented architecture for network anomaly detection
  4. A use case of 5G networks
  5. Management policies
  6. Virtual Infrastructure Policy
  7. Anomaly Detection Function Policy
  8. Mobile Edge Application Policy
  9. Use case orchestration
  10. Deployment and performance results
  11. ASD classification performance
  12. ASD model evaluation throughput
  13. Anomaly detection time estimation
  14. ASD system dynamics analysis.
  15. Conclusions and future work

Figures (8)

  • Figure 1: High-level management and orchestration plane.
  • Figure 2: Network Anomaly Detection System.
  • Figure 3: Use case showing different concerns and how the proposed solution is able to manage them. Concerns are highlighted in red and their mitigation in light green.
  • Figure 4: Sequence diagram of the MEApp policy enforcement.
  • Figure 5: Detail of the implementation of ASD and NAD modules.
  • ...and 3 more figures