Network Path State Latent Space for QoE Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network path monitoring relies on SLA thresholds, which are inadequate for capturing complex network impairments, leading to unnoticed quality of experience (QoE) degradations and reactive routing decisions that can negatively impact user experience.
Innovation Solution
A device forms a latent space by applying dimensionality reduction to timeseries snippets of path metrics, extracts path states, and associates transitions between these states with degraded application experiences, enabling predictive routing to prevent QoE degradations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SLA threshold-based monitoring is used, then network path metrics can be tracked, but complex network impairments affecting QoE cannot be captured
Solution Approach 1:
The patent introduces an intermediary machine learning model that acts as a mediator between raw network path metrics and QoE degradation detection. The model transforms complex multi-dimensional metric data into actionable QoE assessments, enabling precise detection without requiring direct complex monitoring of each individual impairment type.
Solution Approach 2:
The system changes the parameter representation by transforming traditional SLA threshold parameters into machine learning model inputs. Instead of monitoring fixed thresholds, the system uses dynamic parameter transformations through dimensionality reduction and clustering algorithms to capture complex network states that affect QoE.
2Reliability
If traditional aggregate statistics are used, then long-standing network phenomena can be captured, but real-life network path states affecting QoE cannot be fully detected
Solution Approach 1:
The patent applies dimensionality reduction techniques to transform high-dimensional network path metric data into a lower-dimensional latent space while preserving essential information. This dimensional transformation enables the system to capture complex network path states that affect QoE without being overwhelmed by the complexity of raw multi-dimensional data.
Solution Approach 2:
The system performs preliminary dimensionality reduction and state extraction before actual QoE assessment. By pre-processing the network metrics through clustering algorithms to identify representative path states, the system prepares the data in advance for more accurate and efficient QoE degradation detection.
3Productivity
If proactive rerouting based on predicted SLA violations is implemented, then application performance can be improved, but needlessly rerouting traffic can negatively impact user experience
Solution Approach 1:
The patent implements a feedback mechanism where the machine learning model continuously learns from actual QoE outcomes of routing decisions. By monitoring whether predicted SLA violations actually resulted in QoE degradation, the system refines its predictions over time, reducing false positives and avoiding unnecessary rerouting that would negatively impact user experience.
Solution Approach 2:
The system performs preliminary prediction of SLA violations using the trained machine learning model before making routing decisions. By assessing multiple potential path states and their predicted QoE impacts in advance, the system can make informed decisions about whether rerouting is truly necessary, preventing premature or unnecessary traffic redirection.
Data Source
AI summary
In one embodiment, a device forms a latent space by applying dimensionality reduction to timeseries snippets of path metrics for a network path via which traffic for an online application is conveyed. The device extracts a plurality of path states from the latent space. The device presents the plurality of path states for display by a user interface. The device associates a set of transitions between the plurality of path states with the online application providing degraded application experience.