Multi-agent system builds verified network twins using language models
NDT Factory: Synthesizing Verified Network Digital Twins from Semantic Models via Multi-Agent LLM
Networking and Internet ArchitectureArtificial IntelligenceMultiagent SystemsSoftware Engineering
Summary
Managing complex networks automatically is hard because it usually needs lots of manual work to create analysis tools. The authors created a software system that uses several AI language models working together to build digital copies of networks, called network digital twins, from detailed descriptions. These twins can then be used to predict and analyze network behavior without manual coding. They tested this with a network decision process and found it very accurate and reliable.
What this means in practice
- •For network operators: Generate verified network simulation models on demand for decision making in network management without manual coding.
- •For telecom service planners: Automatically produce accurate behavioral models to test network service policies and admission controls before deployment.
Authors
Sudipta Acharya, Petar Djukic, Burak Kantarci
Abstract
Autonomous network management requires systems that can evaluate Network Service Intents (NSIs) under varying conditions without manual implementation of analysis logic, as envisioned in TM Forum Level~4 (L4) autonomy. Behavioral Network Digital Twins (NDTs) enable such evaluation, but existing NDTs rely on pre-defined analytical logic, limiting adaptability for evolving closed-loop control. This paper introduces the NDT factory, a multi-agent software system that synthesizes executable behavioral NDTs on demand from semantic models using Large Language Model (LLM). We validate the system using a Call Admission Control (CAC) case study, where deterministic what-if analysis serves as the admission decision process. The NDT factory generates a complete CAC NDT through parallel synthesis and orchestration, achieving 100% compilation and test pass rates across multiple runs. Simulation over 300 NSIs shows 99.3% decision agreement with a reference implementation, 90% admission rate, and correct attribution of all rejections, demonstrating reliable synthesis with deterministic, verifiable execution.