Skip to content
Review Open access

Network Configuration Translation: A Survey of Symbolic, Neural, Hybrid, and Agentic Approaches

2026 · IEEE Access · Vol 14, pp. 152424-152440 · 0 citations · 49 references

Abstract

Network configuration translation (NCT) is increasingly important in heterogeneous network environments, where operators must migrate configurations across vendors, platforms, software versions, and management paradigms while preserving intended operational semantics. However, the literature remains fragmented, and adjacent surveys do not treat NCT as their primary focus. This survey provides a structured and evidence-aware synthesis of the field. We clarify the scope of NCT, distinguish direct translation from related configuration generation and intent-realization tasks, and organize existing studies into symbolic, neural, hybrid, and agentic approaches. Studies are further classified as NCT-core, NCT-enabling, NCT-related, or NCT-evaluative to separate direct translation evidence from transferable methodological insights. Representative approaches are compared in terms of representation, grounding, validation, generalization, deployment assumptions, and design trade-offs. We also review datasets, metrics, benchmarking practices, and reproducibility limitations. Existing evidence is strongest for syntactic conversion, hierarchical restructuring, parameter mapping, and fragment-level properties. However, configuration-wide dependency preservation and network-wide behavioral equivalence remain insufficiently demonstrated. Future progress requires stable intermediate representations, authoritative knowledge grounding, behavior-level validation, reproducible benchmarks, and human-supervised deployment. Overall, this survey provides a unified foundation for developing reliable, explainable, and deployable NCT systems.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.