ML Parsing of Raw Configurations for Network Inventory
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual creation and maintenance of scripts for parsing network configuration data from various vendors and devices is time-consuming, prone to errors, and costly, and cannot adapt to changes in configuration formats due to software or firmware updates.
Innovation Solution
Utilizing a trained ML-based language model, such as a named entity recognition (NER) or large language model (LLM), to automate the extraction and mapping of unstructured network configuration data into a common information model, eliminating the need for individualized parser scripts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual parser scripts are created for each vendor or device type, then parsing accuracy is improved, but development time and cost increase significantly
Solution Approach 1:
The patent replaces manual mechanical script creation with an automated machine learning system. The ML model automatically learns parsing patterns from sample data and generates parser scripts without human intervention, eliminating the time-consuming manual coding process while maintaining accurate parsing across different vendor formats
Solution Approach 2:
The system enables self-service parsing by training the ML model on sample configuration data, allowing it to automatically adapt to new vendor formats without requiring manual script creation. The model serves itself by learning from data patterns and generating appropriate parsers autonomously
2Reliability
If manually created parsing scripts are used, then parsing reliability is improved, but maintenance complexity increases due to software/firmware updates
Solution Approach 1:
The patent implements dynamic adaptation where the ML model automatically adjusts to new configuration formats when software or firmware updates occur. Instead of maintaining static manual scripts that break with updates, the system dynamically learns new patterns from sample data and regenerates parsers accordingly
Solution Approach 2:
The system incorporates feedback mechanisms where parsing results are validated and used to refine the ML model. When configuration formats change, the model receives feedback through new sample data and automatically updates its parsing logic, ensuring continued reliability without manual intervention
3Loss of information
If individualized parser scripts are built for each device, then data extraction completeness is improved, but resource consumption and cost increase
Solution Approach 1:
The patent creates a universal ML-based parsing system that can handle multiple vendor and device types through a single trained model. Instead of building separate specialized scripts for each device, the model learns general parsing patterns and applies them across different configurations, reducing overall resource consumption while maintaining complete data extraction
Data Source
AI summary
An automated network inventory discovery method, including: at a network inventory discovery engine coupled to a network data source, receiving unstructured network configuration data associated with a network element (NE); and, using a trained machine learning (ML) model, parsing named entity attributes from text of the unstructured network configuration data and mapping the named entity attributes to a common information model having a predetermined data structure. The trained ML model includes a trained named entity recognition (NER) model and/or a trained large language model (LLM). Alternatively, the trained ML model includes a trained NER model that is used to pre-process the text of the unstructured network configuration data prior to feeding resulting data into a trained LLM. The trained ML model serves the function of a custom parser script that is specific to one or more of the unstructured network configuration data or the network data source.