A Method to Automatically Extract a Network Device Configuration Model by Parsing Network Device Configurations

TL;DR

This work introduces an ANTLR-based method to automatically extract a network device configuration model from vendor configurations (e.g., Cisco, YAMAHA) to enable round-trip engineering between device configurations and a reusable configuration model. By parsing show running-config outputs, building a UML-based metamodel, and applying a specification-item mapping, the method generates both the model and device configuration commands, enabling verification without real devices. A case study on a campus-like network demonstrates substantial extraction (1102 specification-item values across 32 item types) and successful command generation when fed into a prior method, highlighting practical potential and current limitations (notably ACL handling). The approach aims to unify design- and operation-phase verification, improve multi-vendor support via inheritance-based extensions, and support future automation and integration with existing verification tools such as Batfish and OpenConfig-inspired workflows.

Abstract

When network engineers design a network, they need to verify the validity of their design in a test environment. Since testing on actual equipment is expensive and burdensome for engineers, we have proposed automatic verification methods using simulators and consistency verification methods for a network configuration model. Combining these methods with conventional verification methods for network device configurations will increase the number of verification options that do not require actual devices. However, the burden of writing existing networks into models has been a problem in our model-based verification. In this paper, we propose a method for automatically extracting a network device configuration model by parsing the contents obtained from network devices via show running-config commands and the like. In order to evaluate the effectiveness of the proposed method in realizing round-trip engineering between network device configurations and the network device configuration model, we extracted a model from existing network device configurations and generated device configuration commands. As a result, we obtained model and commands with high accuracy, indicating that the proposed method is effective.

Figures (10)

• Figure 1: Parsing flow using ANTLR
• Figure 2: Overview of the proposed method
• Figure 3: Network configuration metamodel
• Figure 4: Elements of the network configuration model
• Figure 5: Network configuration model