Network Parameter Configuration via Prediction Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network parameter configuration methods are not universal and often require manual modification, as they do not consider application scenarios and rely on static parameters.
Innovation Solution
A method and apparatus for configuring network parameters using network running data from a first period, which is input into a prediction model or a network traffic type recognition model to obtain optimized parameter values for a subsequent period, eliminating the need for manual configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a static configuration formula is used to set network parameters, then the configuration process is simple, but the parameter optimization is insufficient for different application scenarios
Solution Approach 1:
The patent transforms the static configuration formula into a dynamic prediction model that adapts to different application scenarios. The model dynamically adjusts network parameters based on real-time traffic data and historical patterns, enabling the system to optimize parameters for diverse scenarios such as bulk transfer, interactive, and streaming applications without manual reconfiguration.
Solution Approach 2:
The patent changes the parameter optimization approach from fixed formula-based calculations to machine learning model predictions. The prediction model learns optimal parameter values through training on historical data and continuously adapts parameter settings based on current network conditions, achieving both automation and scenario-specific optimization.
2Manufacturing precision
If manual parameter modification is required for different scenarios, then parameter optimization can be achieved, but the operation complexity increases
Solution Approach 1:
The patent implements self-service through an automated prediction model that performs parameter optimization without human intervention. The model automatically collects network data, predicts optimal parameters, and configures them based on current conditions, eliminating the need for manual parameter modification while maintaining high optimization quality across different scenarios.
Solution Approach 2:
The patent incorporates feedback mechanisms where the prediction model continuously monitors network performance and adjusts parameters based on observed outcomes. The system learns from historical data and real-time feedback, automatically refining its predictions to achieve optimal parameter settings without requiring manual tuning or intervention.
3Productivity
If static parameters are used in configuration formulas, then the configuration is fast, but the parameters cannot adapt to changing network conditions
Solution Approach 1:
The patent applies preliminary action by pre-training the prediction model on extensive historical network data before deployment. This offline training phase prepares the model with learned patterns and knowledge, enabling it to rapidly predict optimal parameters in real-time without requiring complex calculations during actual network operation, thus achieving both speed and adaptability.
Solution Approach 2:
The patent transforms static parameters into dynamic predictions that automatically adapt to changing network conditions. The prediction model continuously updates its recommendations based on real-time traffic patterns, network state, and historical data, ensuring parameters remain optimized even as network conditions evolve, while maintaining fast configuration response times.
Data Source
AI summary
A network parameter configuration method, where in the method, network running data corresponding to a first period is input into a prediction model, so that the prediction model predicts, based on the input network running data, a value of a parameter of a network device in a second period, and the parameter of the network device in the second period is configured to the value predicted by the prediction model.