Integrating Large Language Models into network testbeds: A novel approach for automated experimentation and optimization

This manuscript introduces a novel approach to integrate Large Language Models (LLMs) into wireless network testbeds for automated experimentation and optimization. We propose a framework that leverages LLMs to define system configurations using a structured format called Blueprint and analyse network behaviour through user-prompted queries. Our methodology combines few-shot prompting and prompt engineering to enable automated analysis of network configurations, enhancing decision-making and experimentation efficiency. We validated our approach in both simulated and real-world wireless environments and demonstrated its efficacy in streamlining experiment processes and extracting actionable insights. • Novel LLM-based framework for system configurations using standard Blueprints format. • Domain-agnostic few-shot prompting for automated wireless network configurations. • Real-world validation in Wireless network environments (e.g. LoRaWAN and Wi-Fi). • Practical LLM-driven anomaly detection to enforce network corrective actions.