Adaptive Traffic Processing via Machine Learning Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communications networks face challenges in adapting policy and charging control to rapidly changing traffic characteristics, particularly due to the quick adoption of new mobile applications, which leads to delayed and risky implementations that can affect user experience and service levels.
Innovation Solution
A system and method for adaptive traffic processing using machine-learning based clustering to detect and generate new traffic rules, simulate their effects, and apply them if performance criteria are met, enabling dynamic adaptation to changes in traffic patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual policy and charging control is used for new traffic, then network reliability is maintained, but implementation time is delayed and technical complexity increases
Solution Approach 1:
The system enables self-service by automatically detecting new traffic patterns, clustering them using machine learning, generating policy rules, and simulating their effects without requiring manual intervention from network operators. This automation eliminates the time-consuming manual configuration process while maintaining network reliability through automated evaluation and validation.
Solution Approach 2:
The patent replaces the mechanical manual configuration process with an automated machine learning-based system. The machine learning algorithms automatically analyze traffic data, identify patterns, and generate policy rules, substituting the manual mechanical process of network configuration with an intelligent automated system that operates continuously.
2Stability of the object's composition
If manual policy and charging control is used for new traffic, then network stability is maintained, but device complexity increases
Solution Approach 1:
The system performs self-service by automatically detecting traffic patterns, clustering them using machine learning algorithms, generating policy rules, and simulating their effects without requiring manual intervention from network operators. This automation eliminates the time-consuming manual configuration process while maintaining network reliability through automated evaluation and validation.
Solution Approach 2:
The patent changes the approach from static manual configuration to dynamic automated parameter generation. The machine learning system continuously analyzes traffic characteristics and automatically adjusts policy parameters, transforming the network from a static manually-configured state to a dynamic self-adjusting state that adapts to changing traffic conditions.
3Adaptability or versatility
If automated machine-learning based clustering is used, then adaptability to new traffic is improved, but measurement precision requirements increase
Solution Approach 1:
The patent applies segmentation by dividing traffic into distinct clusters using machine learning algorithms. This segmentation allows the system to identify and process different traffic types independently, improving adaptability to new traffic patterns while managing measurement complexity through organized categorization rather than requiring perfect precision for all traffic types.
Solution Approach 2:
The system incorporates feedback mechanisms where simulated policy effects are evaluated against performance criteria, and the results feed back into the rule generation process. This feedback loop allows the system to iteratively improve its traffic detection and classification accuracy, adapting to new traffic patterns while continuously refining measurement precision through practical validation.
Data Source
AI summary
There is provided adaptive traffic processing in a communications network. Traffic flows from real-life traffic are detected, rules are applied to detected traffic flows, machine-learning based clustering of the detected traffic flows is performed, at least one new traffic rule is generated on the basis of one or more generated traffic clusters, the generated rules are simulated for obtaining a result indicating behavioural effect in the traffic of the communications network, the result is evaluated against one or more performance criteria of the communications network and the generated rules are applied if the performance criteria of the communications network have been met.