Network Configuration Assets for Runtime IPU Reconfiguration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing network architectures face challenges in accurately establishing performance and information flows of network devices after deployment, leading to inefficiencies and service quality issues, necessitating dynamic reconfiguration mechanisms based on the behavior of deployed infrastructure processing units (IPUs).

Innovation Solution

Utilizing AI techniques for software-defined networking (SDN) to perform dynamic analysis of network configurations, generate functionality tags for grouped devices, determine fitness and error indices, and automatically generate configuration assets using large language models (LLMs) to optimize network device performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If network devices are deployed based on theoretical design, then initial configuration can be established, but the actual performance and information flows cannot be accurately established after deployment

Engineering Contradiction:
Improveperformance measurement accuracyVSAvoidconfiguration management difficulty
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent implements feedback mechanisms by collecting actual runtime data from deployed network devices and using AI models to compare theoretical configurations with observed behavior. This feedback loop enables continuous refinement of configuration assets based on real performance metrics and information flow patterns, resolving the contradiction between measurement accuracy and operational ease.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system enables self-service by automatically generating and refining configuration assets through AI-driven analysis of deployed device behavior. Rather than requiring manual configuration adjustments, the system autonomously learns from runtime data and generates optimized configurations, improving both measurement accuracy and operational simplicity.

Inventive Principle:
Principle #25Self-service

2Adaptability or versatility

If static configuration assets are used for network devices, then deployment is simplified, but dynamic adaptation to actual behavior is lost

Engineering Contradiction:
Improveconfiguration adaptabilityVSAvoidconfiguration management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent transforms static configuration assets into dynamic ones by implementing AI models that continuously learn from runtime behavior of network devices. The system adapts configurations based on observed information flows and performance metrics, enabling dynamic adjustment while managing complexity through automated machine learning processes rather than manual intervention.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system performs preliminary action by pre-training AI models on theoretical network designs before deployment. These pre-trained models serve as initial configuration assets that can be quickly adapted to actual deployed behavior, reducing the complexity of real-time adaptation while maintaining high adaptability to actual device performance.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If manual analysis and reconfiguration of network devices is performed, then service quality issues can be addressed, but time loss and inefficiency increase

Engineering Contradiction:
Improvenetwork management efficiencyVSAvoidreconfiguration time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent replaces manual mechanical analysis and reconfiguration processes with AI-driven automated systems. Machine learning models automatically analyze runtime data, identify performance issues, and generate reconfiguration assets without human intervention, dramatically improving productivity while minimizing time loss compared to manual network device management.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4672694A1Methods and apparatus to manage configuration assets for network devices
Publication Date: 2025.12.31 INTEL CORP
  • EP4672694A1 patent drawingFigure 1
  • EP4672694A1 patent drawingFigure 2
  • EP4672694A1 patent drawingFigure 3~4

AI summary

Methods, apparatus, systems, and articles of manufacture to manage configuration assets for network devices are disclosed. Example instructions cause at least one programmable circuit to generate network infrastructure instructions using a model, the network infrastructure instructions based on a configuration request and on a network infrastructure; and deploy a program corresponding to the network infrastructure instructions to at least one device in the network infrastructure.