Network Motif Identification from High Frequency Telemetry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network systems rely on threshold-based mechanisms that fail to capture complex and dynamic network impairments, leading to unnoticed Quality of Experience (QoE) degradation in online applications, as they rely on aggregate statistics and lack visibility into repetitive patterns of network issues.
Innovation Solution
The implementation of a device that extracts motifs from high-frequency network telemetry by applying a sliding time window, groups similar patterns, and receives user feedback to identify motifs associated with degraded application experiences, enabling predictive application aware routing to proactively optimize network paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If threshold-based mechanisms are used to monitor network SLA, then the system is simple to operate, but it fails to capture complex and dynamic network impairments leading to unnoticed QoE degradation
Solution Approach 1:
The patent segments network telemetry data into discrete time windows and extracts individual metric values (delay, jitter, packet loss) from continuous streams. This segmentation enables the system to analyze specific time-bound network conditions rather than relying on aggregate thresholds, improving detection of transient impairments while maintaining manageable data processing through structured segmentation.
Solution Approach 2:
The system transitions from static threshold-based monitoring to dynamic motif-based detection. By continuously extracting telemetry metrics at regular intervals and comparing them against learned motifs (recurring patterns), the system adapts to changing network conditions and detects complex impairments that static thresholds cannot capture, thereby improving measurement precision without proportionally increasing complexity.
2Measurement precision
If aggregate statistics are computed for network path metrics, then the system is easy to operate, but it cannot capture all network issues that affect application QoE in real-life
Solution Approach 1:
The system performs preliminary extraction of telemetry metrics at regular time intervals before actual analysis occurs. By pre-processing and storing metric values in structured formats (extracting delay, jitter, packet loss at each time window), the system prepares data for subsequent motif matching and pattern recognition, enabling comprehensive QoE analysis without requiring complex real-time processing during actual monitoring events.
Solution Approach 2:
The system creates copies of network telemetry data by extracting and storing metric values at multiple time points. These copied data points form the basis for motif extraction and pattern recognition, allowing the system to analyze recurring network conditions without needing to process the entire continuous telemetry stream, thereby improving visibility while managing complexity through selective data replication.
3Reliability
If traditional SLA thresholds are used as proxy for QoE, then the system is simple to implement, but it cannot detect complex types of impairments that go unnoticed
Solution Approach 1:
The system implements feedback by continuously comparing extracted telemetry metrics against stored motifs and using this comparison to generate alerts or routing decisions. The motif-based approach provides feedback on recurring network patterns, enabling the system to learn from historical data and improve QoE assessment accuracy over time without requiring complex real-time analysis of every network event.
Solution Approach 2:
The system changes from monitoring single threshold parameters to analyzing multiple telemetry parameters (delay, jitter, packet loss) simultaneously. By extracting and analyzing combinations of these parameters together rather than individually, the system can detect complex impairment patterns that affect QoE, improving reliability of QoE assessment while managing complexity through multi-parameter correlation analysis.
Data Source
AI summary
In one embodiment, a device extracts portions of a timeseries of a network path metric by applying a sliding time window to the timeseries. The device groups a subset of the portions of the timeseries into a motif based on their similarities. The device provides data regarding the motif for display to a user via a user interface. The device receives, from the user interface, a label for the motif indicative of whether the motif is associated with degraded application experience for a particular online application.