Beacon Cluster Visualization for Web Performance Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional systems face challenges in efficiently visualizing and analyzing real user measurement (RUM) data due to the large volume of data collected, which exceeds memory, network, and CPU limits, making it difficult to identify performance issues in real-time user experiences on websites.
Innovation Solution
An algorithmic approach that reduces the number of nodes and links in visualizations, using force-directed charts and convex hulls to efficiently represent large datasets, allowing for real-time visualization of RUM data with reduced memory, network, and CPU usage, enabling identification of performance patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all RUM data beacons are rendered on the display screen, then complete data visualization is achieved, but memory usage and processing time exceed system limits
Solution Approach 1:
The patent extracts and removes redundant data points from the visualization. By identifying and eliminating duplicate or highly similar beacons, the system reduces the total number of nodes rendered while preserving the essential patterns and insights. This extraction approach maintains data completeness for decision-making purposes while significantly reducing memory consumption and processing requirements.
Solution Approach 2:
The patent applies partial action by rendering only a representative subset of beacons rather than all collected data. The system determines an optimal subset that captures the essential performance patterns without requiring complete data visualization. This partial rendering approach achieves sufficient analytical value while operating within system resource constraints.
2Loss of information
If all RUM data beacons are rendered on the display screen, then complete data visualization is achieved, but CPU resources and processing time are overwhelmed
Solution Approach 1:
The patent extracts redundant beacons from the dataset before rendering. By removing duplicate or highly similar data points, the system reduces the computational burden of rendering and interaction operations. This extraction process maintains the essential performance patterns while significantly improving processing speed and reducing CPU resource consumption.
Solution Approach 2:
The patent implements partial action by rendering only a representative subset of beacons rather than the complete dataset. This approach achieves sufficient analytical value for performance monitoring while dramatically reducing the processing time and computational resources required for visualization and interaction calculations.
3Loss of information
If a large number of data points are displayed, then comprehensive data coverage is achieved, but user ability to identify patterns is reduced due to screen clutter
Solution Approach 1:
The patent extracts and removes redundant beacons that would contribute to screen clutter. By eliminating duplicate or highly similar data points, the system reduces visual noise while preserving the essential performance patterns. This results in a cleaner display where meaningful patterns are more easily identifiable by users.
Solution Approach 2:
The patent applies partial action by displaying only a representative subset of beacons that captures comprehensive data coverage without creating visual clutter. This selective rendering approach maintains adequate data representation while significantly improving the user's ability to identify performance patterns and anomalies.
4Loss of information
If interaction calculations are performed between all nodes, then complete interaction analysis is achieved, but calculation time increases with the square of the number of nodes
Solution Approach 1:
The patent extracts redundant beacons from the dataset, which directly reduces the number of nodes requiring interaction calculations. By removing duplicate or highly similar data points, the system dramatically reduces the computational complexity from O(n²) to a much lower order, while preserving the essential interaction patterns needed for performance analysis.
Solution Approach 2:
The patent implements partial action by performing interaction calculations only on a representative subset of beacons rather than all collected data. This approach achieves sufficient interaction analysis for performance monitoring while reducing calculation time from quadratic complexity to a manageable level, enabling real-time or near-real-time analysis.
Data Source
AI summary
A computer-implemented method for creating a visualization of beacons collected over a specified time period from users on a website. Beacons are rendered as nodes grouped into clusters, with relatedness between beacons being represented as a link. The number of nodes rendered is reduced along with the number of links that connect pairs of nodes. The resulting data structure is rendered as a force-directed chart by assigning force unit values to each of the nodes and links, each representative node of a cluster of nodes being assigned a relatively high negative charge and all remaining nodes in the cluster being assigned a relatively low negative charge. Link distances and strengths between unrelated/related clusters are also assigned values. A set of physical laws is applied to all of the nodes and links to determine their relative position in the visualization based on their assigned force unit values.