Asynchronous Network Traffic Rate Calculation via Random Packet Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network monitoring systems face delays in providing reliable traffic rate measurements due to the need for regularly spaced intervals, which reduces responsiveness and accuracy, and are resource-intensive, especially with methods like Cisco's Netflow™.
Innovation Solution
The method involves asynchronously calculating traffic rates based on randomly sampled packets using a single-pole recursive low-pass filter, updating state variables as samples arrive, and calculating rates at irregular intervals, reducing the need for continuous updates and minimizing hardware costs by performing analysis remotely.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If regularly spaced intervals are used for calculating traffic rates, then measurement reliability is improved, but responsiveness and speed are worsened due to delays of up to two minutes
Solution Approach 1:
The patent transitions from static, fixed-interval rate calculations to dynamic, event-driven calculations. Rate values are calculated asynchronously based on when samples are actually received rather than adhering to a rigid time schedule, allowing the system to adapt to varying traffic conditions and provide timely updates without sacrificing reliability
Solution Approach 2:
The patent replaces continuous periodic interval-based calculations with periodic sample-based calculations. Instead of calculating rates at fixed time intervals regardless of sample arrival, the system calculates rates periodically whenever sufficient samples are accumulated, eliminating unnecessary delays while maintaining measurement integrity
2Speed
If measurement interval is reduced to improve responsiveness, then speed is improved, but accuracy is worsened due to fewer samples contributing to smaller intervals
Solution Approach 1:
The patent changes the parameter of measurement interval from a fixed time value to a flexible value based on sample accumulation. By adjusting the effective measurement interval dynamically based on when samples arrive and the desired confidence level, the system can use shorter intervals when samples arrive frequently (maintaining accuracy) and longer intervals when samples are sparse (maintaining responsiveness)
3Measurement precision
If traditional packet sampling analysis methods are used, then traffic rate estimation is achieved, but resource consumption and equipment costs are increased
Solution Approach 1:
The patent extracts the computationally intensive rate calculation function from the network nodes themselves and relocates it to external analysis systems. Nodes only perform lightweight sampling and forwarding of sample data, while the heavy lifting of rate calculation, filtering, and analysis is performed remotely, significantly reducing the resource burden on network infrastructure
Solution Approach 2:
The patent introduces an intermediary analysis system that acts as a mediator between packet sampling and rate calculation. This intermediary layer processes sample data using efficient algorithms (such as exponential weighted moving averages) to derive traffic rates without requiring expensive hardware or intensive processing at the network node level
Data Source
AI summary
A method for calculating rates based on random sampling includes updating at least one state variable when sampled information is received. The method further includes calculating a rate value based on the state variable and the elapsed time since the state variable was previously updated.