GraphCue for SDN Configuration Code Synthesis

2025-12-19

AI Summary

To address the challenge of automated SDN configuration code generation, this paper proposes a topology-aware retrieval-augmented and agent-in-the-loop framework. It models network topologies as JSON graphs and learns topology embeddings via a lightweight third-order GCN coupled with contrastive learning. Retrieved similar validation cases—based on these embeddings—are integrated with structured prompts to guide large language model (LLM) generation. Generated configurations undergo executable validation, with feedback driving iterative refinement. The core contributions are: (i) the first topology-driven graph retrieval mechanism for SDN configuration synthesis, and (ii) an agent-in-the-loop paradigm that unifies graph representation learning, structured prompt engineering, and executable feedback. Evaluated on 628 real-world cases, the framework achieves an 88.2% success rate within 20 iterations, with 95% of verification cycles completing in ≤9 seconds. Ablation studies confirm that topology-aware retrieval and structured prompting are critical to performance gains.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Algorithm Configuration
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Agentic search
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

We present GraphCue, a topology-grounded retrieval and agent-in-the-loop framework for automated SDN configuration. Each case is abstracted into a JSON graph and embedded using a lightweight three-layer GCN trained with contrastive learning. The nearest validated reference is injected into a structured prompt that constrains code generation, while a verifier closes the loop by executing the candidate configuration and feeding failures back to the agent. On 628 validation cases, GraphCue achieves an 88.2 percent pass rate within 20 iterations and completes 95 percent of verification loops within 9 seconds. Ablation studies without retrieval or structured prompting perform substantially worse, indicating that topology-aware retrieval and constraint-based conditioning are key drivers of performance.

Problem

• Automates SDN configuration code synthesis using topology
• Retrieves validated references to constrain code generation
• Verifies configurations via execution feedback to improve accuracy

Innovation

• Uses topology-grounded retrieval and agent-in-the-loop framework
• Embeds JSON graphs with lightweight GCN via contrastive learning
• Injects nearest reference into structured prompt to constrain generation