Network Congestion Detection via Historical Traffic Boundary Trends
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network congestion detection in network devices is challenging due to the inability to accurately identify chronic congestion, leading to inefficient network resource utilization and potential packet loss.
Innovation Solution
A system comprising a data logging device and a congestion analyzer that collects and analyzes traffic measurements over time intervals to determine upper and lower boundaries, identifying trends to detect congestion by identifying a substantially constant upper boundary trend concurrent with a rising lower boundary trend.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If network administrators monitor traffic to identify congestion, then network performance can be improved, but the complexity of detection and measurement increases
Solution Approach 1:
The patent transforms the congestion detection problem by changing the parameters being monitored from raw traffic volume to statistical boundaries (upper and lower boundaries) derived from historical data. This allows the system to identify chronic congestion patterns without requiring complex real-time analysis of every traffic metric, resolving the contradiction between detection accuracy and complexity
Solution Approach 2:
The system performs preliminary analysis by collecting historical traffic data and establishing upper and lower boundaries before congestion occurs. This preliminary action creates a baseline that simplifies future congestion detection, as the system only needs to compare current traffic against pre-established boundaries rather than performing complex real-time analysis
2Productivity
If network administrators upgrade infrastructure to address congestion, then network performance improves, but cost and resource allocation complexity increase
Solution Approach 1:
The patent implements a feedback mechanism where congestion detection triggers notifications to network administrators. This feedback loop provides evidence-based insights into where congestion occurs, enabling administrators to make informed infrastructure upgrade decisions rather than guessing, thus improving productivity while managing complexity through data-driven guidance
3Productivity
If packet loss occurs due to congestion, then network efficiency decreases, but implementing congestion control protocols increases system complexity
Solution Approach 1:
The system enables network devices to self-monitor and self-report congestion conditions by automatically comparing traffic against established boundaries. This self-service approach allows the network to identify and report congestion issues without requiring complex external control protocols, maintaining network efficiency while minimizing additional complexity
Data Source
AI summary
Detecting network congestion at a network device by collecting, for various time intervals, measurements of a quantifiable characteristic of traffic transiting the network device at a plurality of respective measurement times during each of the time intervals. Described systems and methods select two or more sets of measurements, each set corresponding to a respective time interval; determine, for each of the selected sets of measurements, respective upper and lower boundaries for the quantifiable characteristic during the respective time interval; and analyze trends for a time window inclusive of the time intervals corresponding to the selected sets of measurements, the trends including an upper boundary trend and a lower boundary trend. Congestion is then identified responsive to the analysis detecting a substantially constant upper boundary trend concurrent with a rising lower boundary trend.