Parallel Data Analysis Modules for Mobile Access Network Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current data analysis methods for mobile access networks are inadequate in processing speed, particularly due to the short validity period of wireless air interface data, which poses a high requirement for rapid data analysis in these networks.

Innovation Solution

A data analysis apparatus comprising multiple modules that work in parallel to quickly analyze data from network nodes in mobile access networks, generating feature vectors, creating model instances, and adjusting configuration parameters based on prediction results, thereby optimizing network performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If data analysis is performed using traditional sequential methods, then processing accuracy can be maintained, but processing speed is insufficient to meet the short validity period of wireless air interface data

Engineering Contradiction:
Improvedata processing speedVSAvoidsystem complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent segments the data analysis system into multiple independent modules (feature extraction module, model instance creation module, prediction module, parameter adjustment module) that operate in parallel. Each module processes specific aspects of data analysis independently, allowing simultaneous execution of multiple analysis tasks which significantly improves processing speed while maintaining manageable complexity through modular design

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from sequential single-threaded processing to parallel multi-threaded processing by adding a temporal dimension to the execution model. Multiple model instances process different feature vectors simultaneously across different time slices, effectively increasing processing throughput without proportionally increasing physical hardware resources

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If model training is performed continuously to maintain accuracy, then prediction accuracy is improved, but processing time increases due to the overhead of retraining

Engineering Contradiction:
Improveprediction accuracyVSAvoidmodel training time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary model training offline to create a trained model, then uses this pre-trained model for rapid online predictions. Model updates are performed periodically or event-driven rather than continuously, allowing the system to maintain good prediction accuracy while avoiding the time overhead of constant retraining. Feature extraction and prediction can proceed using the pre-trained model structure

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates multiple copies (instances) of the trained model that can process different feature vectors simultaneously. Instead of retraining a single model continuously, multiple model instances are deployed to handle different data streams or feature sets in parallel, maintaining prediction accuracy across multiple concurrent operations without requiring repeated training cycles

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3720169B1Data analysis device, system, and method
Publication Date: 2024.09.11 HUAWEI TECH CO LTD
  • EP3720169B1 patent drawingFigure 1(a)~2
  • EP3720169B1 patent drawingFigure 3
  • EP3720169B1 patent drawingFigure 4

AI summary

This application provides a data analysis apparatus, system, and method. The data analysis apparatus may collect data of a network node in a mobile access network, and then perform data analysis and adjust a configuration parameter of the network node, to implement targeted data analysis and parameter adjustment for the network node, so that a speed of data analysis and configuration parameter adjustment can be improved. In addition, modules of the data analysis apparatus may run in parallel, so that the speed of data analysis and configuration parameter adjustment can be further improved.