Concurrent Network Metric Visualization for Correlation Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network telemetry visualization tools obscure certain network events, particularly those occurring at specific frequencies or intervals, making it difficult to diagnose network anomalies and failures effectively.
Innovation Solution
A system for concurrent visualization of time-series network metrics from multiple disparate sources, allowing for synchronized display and comparative analytics through a single-pane-of-glass view, using adaptive resolution time-bucketing and dynamic scaling to accommodate vastly different frequency scales.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If network telemetry is visualized using traditional tools focusing on distinct data sources, then the visualization is simplified and easier to interpret, but certain network events occurring at specific frequencies or intervals are obfuscated and become invisible
Solution Approach 1:
The patent combines multiple time-series visualizations from different data sources into a single synchronized view. The system overlays events from disparate telemetry sources on a common time axis, allowing correlated events to be visible simultaneously. This merging approach prevents information loss by displaying multiple event types together rather than in separate isolated views.
Solution Approach 2:
The patent adds temporal synchronization as a new dimension to the visualization. By aligning events from different sources according to their timestamps and displaying them on a shared time axis, the system creates an additional temporal dimension that reveals correlations and patterns invisible in traditional separate views. This dimensional approach allows events at different frequencies to coexist visibly in the same space.
2Measurement precision
If multiple time-series from disparate data sources are displayed concurrently, then correlation inference is improved, but the visualization complexity and difficulty of interpretation increase
Solution Approach 1:
The patent segments the visualization into distinct time-series layers, each representing a different data source or event type. Each time-series is visually separated through layering while maintaining temporal alignment, allowing analysts to distinguish between different event sources. This segmentation enables precise correlation inference by keeping data sources distinct yet synchronized, preventing visual confusion while maintaining analytical accuracy.
Solution Approach 2:
The patent introduces a common time axis as an intermediary element that mediates between multiple disparate data sources. This shared temporal reference frame acts as a mediator that aligns events from different sources without requiring direct comparison of their native formats. The time axis serves as an intermediary structure that simplifies interpretation by providing a unified temporal context for all events.
3Loss of information
If adaptive resolution time-bucketing is used to accommodate vastly different frequency scales, then the visualization can display events at different frequencies clearly, but the computational complexity increases
Solution Approach 1:
The patent implements dynamic time-bucketing where the resolution of time bins adapts based on the frequency characteristics of each data source. High-frequency events are aggregated into coarser time buckets, while low-frequency events use finer resolution buckets. This dynamic adaptation allows the system to represent events at vastly different frequencies clearly without overwhelming the visualization with excessive data points, managing computational complexity through adaptive aggregation.
Data Source
AI summary
In one embodiment, a device obtains telemetry data indicative of a plurality of different types of events that occurred in a network. The device computes event counts for each of the plurality of different types of events within each of a sequence of predefined timespans based on the telemetry data and generates timeseries for each of the plurality of different types of events using the event counts. The device provides display data that causes a user interface to display a selected two or more of the timeseries concurrently.