|
| 1 | +# 5G Components management with LLM |
| 2 | + |
| 3 | +# 5G-GPT: AI-Driven Kubernetes-Based Management of Open5GS |
| 4 | + |
| 5 | +Open5GS-LLM is a Kubernetes-native application that leverages Large Language Models (LLMs) to monitor, manage, and explain Open5GS deployments. It provides a conversational interface—similar to ChatGPT—that enables users to query Open5GS status, update configurations, and gain insight into 4G/5G core functions, all through natural language. |
| 6 | + |
| 7 | +## Project Goals |
| 8 | + |
| 9 | +- Deploy and manage Open5GS on Kubernetes clusters |
| 10 | +- Use LLMs to interpret and edit Kubernetes manifests and Helm values |
| 11 | +- Explain Open5GS network functions (AMF, SMF, UPF, etc.) and configuration options |
| 12 | +- Provide a ChatGPT-like web UI for interactive management and learning |
| 13 | +- Enable dynamic configuration updates using natural language commands |
| 14 | + |
| 15 | +## Architecture |
| 16 | + |
| 17 | +Web Chat UI |
| 18 | +<--> LLM Backend (OpenAI, Aws Bedrock, Local LLM) |
| 19 | +<--> Kubernetes Cluster (Open5GS ) |
| 20 | + ↑ |
| 21 | + | |
| 22 | +Kubernetes Controller |
| 23 | + - Watches Open5GS Pods, CRDs, and Helm changes |
| 24 | + - Converts LLM instructions to safe Kubernetes actions |
| 25 | +Open5Gs Controller |
| 26 | + - Watches the logs for Open5G components |
| 27 | + - Configure and update Open5Gs components |
| 28 | + |
| 29 | + |
| 30 | +## Features |
| 31 | + |
| 32 | +- Real-time Monitoring |
| 33 | + - AMF, SMF, UPF, HSS status and logs |
| 34 | + - UE session tracking via Kubernetes events |
| 35 | + |
| 36 | +- Natural Language Configuration |
| 37 | + - Convert natural language to YAML or Helm changes |
| 38 | + - Apply changes to running clusters with safety checks |
| 39 | + |
| 40 | +- Conversational Explanations |
| 41 | + - Ask about config values and their impact |
| 42 | + - LLM provides context-aware descriptions |
| 43 | + |
| 44 | +## Components |
| 45 | + |
| 46 | +TODO: |
| 47 | + |
| 48 | +## Getting Started |
| 49 | + |
| 50 | +### Prerequisites |
| 51 | + |
| 52 | +TODO: |
0 commit comments