Interest Metric for Selective Path Measurement Reporting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current approaches to reporting path measurements for Quality of Experience (QoE) prediction in computer networks are inefficient, as they transmit and process excessive data, without considering the impact on prediction model performance, leading to wastage of network resources and inaccurate user experience assessments.
Innovation Solution
Implementing an interest metric that determines the relevance of path measurements by calculating differences between current and prior measurements, and using heuristics such as variation, sampling, and utility to selectively report only data that improves the effectiveness of QoE prediction models, thereby reducing unnecessary data transmission and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If path measurements are reported at small timescales to improve QoE prediction accuracy, then prediction model performance is improved, but network data transmission volume increases excessively
Solution Approach 1:
The patent extracts only the essential and relevant path measurement data that contributes to QoE prediction accuracy, filtering out redundant information. This is achieved by identifying and reporting only measurements that provide new information or significant changes, thereby reducing data transmission volume while maintaining prediction effectiveness.
Solution Approach 2:
The patent applies different reporting strategies to different types of path measurements based on their relevance to QoE prediction. Critical measurements are reported at higher frequencies with greater detail, while less relevant measurements are reported at lower frequencies or with reduced precision, optimizing the balance between prediction accuracy and network resource consumption.
2Reliability
If all path measurements are transmitted to improve prediction model training, then model performance is improved, but network resources are wasted
Solution Approach 1:
The patent implements partial action by transmitting only the necessary subset of path measurements required for effective model training, rather than all available measurements. This selective transmission approach ensures sufficient training data quality while avoiding the excessive consumption of network resources associated with transmitting complete measurement sets.
Solution Approach 2:
The patent dynamically adjusts reporting parameters such as measurement frequency, data granularity, and selection criteria based on current network conditions, device state, and prediction model needs. This allows the system to optimize the balance between model training quality and network resource consumption under varying operational conditions.
3Measurement precision
If comprehensive path measurements are reported to ensure accurate QoE assessment, then prediction accuracy is improved, but data processing complexity increases
Solution Approach 1:
The patent segments path measurements into distinct categories based on their relevance and importance to QoE prediction. This segmentation enables selective processing and transmission of different measurement types, reducing overall processing complexity while ensuring that critical measurements are handled with appropriate detail and frequency.
Data Source
AI summary
In one embodiment, a device determines a first difference between current path measurements and prior path measurements. The device determines a second difference between current predictions and prior predictions made by a prediction model based on path measurements. The device computes, based on the first difference and the second difference, an interest metric for the current path measurements. The device sends at least a portion of the current path measurements for input to the prediction model, when the interest metric exceeds a predefined threshold.