Reasoning and Knowledge Acquisition Framework for 5G Network Analytics

Area of Science:

• Computer Science
• Network Engineering
• Artificial Intelligence

Background:

• Autonomic self-management is crucial for next-generation networks, presenting significant operational challenges.
• Understanding network status and predicting disruptions are key to maintaining network operability.

Purpose of the Study:

• To propose an automated analysis framework for inferring knowledge in 5G networks.
• To enhance network status understanding and predict potential disruptions.
• To support proactive or reactive mitigation strategies.

Main Methods:

• The framework is based on the Endsley situational awareness model.
• It integrates automated metrics discovery, pattern recognition, prediction techniques, and rule-based reasoning.
• A use case methodology allows customization of rules and parameters.

Main Results:

• The framework was instantiated using a reference network traffic dataset.
• It successfully identified suspicious patterns and predicted data volume behavior.
• Preliminary results show good accuracy in inferring anomalous traffic volumes with simple configurations.

Conclusions:

• The proposed framework effectively infers knowledge for autonomic self-management in 5G networks.
• It demonstrates adaptability and accuracy in identifying and predicting network anomalies.
• This approach aids in maintaining network operability through automated analysis.