Browser Waterfall Data QoE Prediction Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Assessing the quality of experience (QoE) of web-based applications is challenging due to their complex workflows and varied resource interactions, as traditional methods rely on service level agreements (SLAs) that may not capture nuanced user experiences, and existing approaches are reactive and prone to unnecessary rerouting, which can degrade application performance.
Innovation Solution
A system that collects browser waterfall data and user feedback to train prediction models, allowing for proactive adjustments to network routing to optimize QoE by predicting potential SLA violations and rerouting traffic before they occur, using machine learning to correlate detailed network activity with user satisfaction metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If SLA thresholds are used as a proxy for QoE, then network impairment detection is simplified, but complex types of impairments go unnoticed and QoE assessment is inaccurate
Solution Approach 1:
The patent introduces browser waterfall data as an intermediary measurement tool that bridges the gap between simple SLA thresholds and true QoE. Waterfall data captures detailed resource loading sequences, timing, and status information, serving as a mediator that translates complex user experience into measurable network performance indicators without requiring direct user feedback for every interaction.
Solution Approach 2:
The patent replaces the traditional mechanical SLA threshold checking system with a data-driven approach using browser waterfall analysis. Instead of relying on predetermined network performance thresholds, the system substitutes a dynamic measurement mechanism that captures actual resource loading behavior, replacing static SLA checks with adaptive waterfall-based QoE assessment.
2Reliability
If proactive rerouting is performed based on SLA predictions, then SLA violations are prevented, but unnecessary rerouting occurs which degrades application performance
Solution Approach 1:
The patent applies preliminary action by using browser waterfall data to identify early signs of QoE degradation before SLA violations occur. The system analyzes resource loading patterns, timing anomalies, and error sequences in advance, enabling proactive network adjustments only when genuine QoE issues are detected, rather than reacting to every potential SLA threshold breach.
Solution Approach 2:
The patent implements a feedback mechanism where browser waterfall data provides real-time information about actual user experience quality. This feedback loop allows the system to continuously monitor resource loading performance and trigger rerouting decisions only when waterfall analysis confirms genuine QoE degradation, preventing unnecessary rerouting while maintaining SLA compliance.
3Measurement precision
If detailed browser waterfall data is collected, then QoE assessment accuracy is improved, but data collection complexity and processing overhead increase
Solution Approach 1:
The patent applies self-service by leveraging the browser's existing waterfall data collection capabilities that are already built into modern web browsers for debugging and performance monitoring purposes. The system repurposes this self-collected data without requiring additional client-side instrumentation or complex data gathering mechanisms, reducing overall system complexity while maintaining high measurement precision.
Data Source
AI summary
In one embodiment, a device obtains browser waterfall data from a web browser of a client that is used to access an online application via a network. The device obtains user feedback from the client indicative of whether a user of the client is satisfied with their experience with the online application. The device trains, using the browser waterfall data and user feedback as training data, a prediction model to predict a quality of experience metric for the online application. The device causes an adjustment to the network based on a prediction by the prediction model.