Network Congestion Detection via Baseline Traffic Demand Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current congestion detection methods in radio access networks rely on instantaneous traffic characteristics, leading to inaccuracy and overcompensation, failing to differentiate between medium- to long-term congestion and short-term traffic bursts, and often misinterpret non-congested conditions as congested.
Innovation Solution
A method to calculate a scalar congestion value based on the comparison between measured and/or estimated traffic and traffic demand within a network, capturing medium- to long-term congestion levels, allowing for differentiated congestion mitigation and flexible configuration to distinguish between congestion scenarios and low connection performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If conventional congestion detection methods use instantaneous traffic characteristics, then congestion can be detected quickly, but the detection accuracy deteriorates leading to false positives and overcompensation
Solution Approach 1:
The patent applies preliminary action by establishing a baseline traffic demand profile before congestion occurs. The system pre-calculates expected traffic patterns under normal conditions, then compares real-time traffic against this pre-established baseline to detect deviations indicating congestion. This allows accurate congestion detection without relying solely on instantaneous traffic spikes, reducing false positives while maintaining quick response.
2Quantity of substance
If network operators upgrade network capacities to satisfy exponential traffic demand, then network capacity increases, but congestion management complexity increases
Solution Approach 1:
The patent implements feedback by continuously monitoring actual traffic demand and comparing it against baseline expectations, then using this feedback to dynamically adjust congestion detection thresholds and mitigation actions. The system learns from historical traffic patterns and refines its congestion detection accuracy over time, enabling effective congestion management even as network capacity and traffic complexity increase exponentially.
3Reliability
If multiple measurement parameters are used to detect congestion, then detection coverage improves, but the ability to differentiate true congestion from non-congested conditions deteriorates
Solution Approach 1:
The patent applies parameter changes by transforming multiple traffic measurement parameters into a normalized congestion indicator that accounts for baseline traffic demand. Instead of relying on absolute thresholds for individual parameters like throughput or resource utilization, the system calculates relative deviations from expected baseline patterns, enabling precise differentiation between true congestion and normal high-traffic conditions while maintaining comprehensive detection coverage.
Data Source
AI summary
A method for providing congestion information in a network is performed in a memory available to a computing entity. A traffic demand is obtained within a certain part of the network by evaluating an amount of traffic in the part of the network per time. A congestion value representing a congestion level of a bottleneck connection link in the network is calculated. The congestion value is a scalar and calculated based on a comparison between measured and/or estimated traffic and traffic demand within a certain part of said network.