CDN Server Streaming KPI Prediction for Root Cause Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content delivery networks (CDNs) lack effective methods for predicting and identifying the root causes of streaming-media quality issues, leading to delayed detection of performance anomalies that can impact end-user experience.
Innovation Solution
A method and system for quantifying streaming-media KPIs, establishing thresholds, and using machine learning to predict anomalies and identify root causes in CDN performance, enabling proactive issue resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional CDN monitoring methods are used, then system complexity is reduced, but detection speed and anomaly prediction capability deteriorate
Solution Approach 1:
The system performs preliminary actions by pre-calculating KPI thresholds and training machine learning models in advance. The anomaly detection model is trained offline with historical data, and KPI thresholds are predetermined based on service level agreements, enabling fast real-time detection without complex runtime calculations
Solution Approach 2:
The patent introduces intermediary elements including a machine learning anomaly detection model that acts as a mediator between raw KPI data and anomaly identification, and a root cause analysis module that mediates between multiple KPI sources and the final diagnosis, simplifying the detection process
2Reliability
If reactive issue resolution is used, then system simplicity is maintained, but end-user experience deteriorates due to delayed detection
Solution Approach 1:
The system performs preliminary anomaly detection by comparing real-time KPIs against predetermined thresholds and trained machine learning models before actual service degradation occurs. This proactive approach enables early warning and preventive action, improving reliability while minimizing detection delay
Solution Approach 2:
The patent implements feedback mechanisms where anomaly detection results trigger root cause analysis, which in turn provides actionable insights back to the CDN system. This closed-loop feedback enables continuous improvement and proactive resolution, enhancing end-user experience
3Extent of automation
If manual root cause analysis is performed, then automation level is reduced, but analysis accuracy may improve due to expert judgment
Solution Approach 1:
The system enables self-service automation through machine learning models that automatically detect anomalies and perform root cause analysis without human intervention. The anomaly detection model and root cause analysis module work autonomously to identify issues and suggest resolutions, significantly reducing analysis time while maintaining high accuracy through algorithmic precision
Data Source
AI summary
Aspects of the subject disclosure may include, for example, identifying a threshold value of a streaming-media key performance indicator (KPI) based on a predetermined target portion of end-user devices that provide an acceptable performance. Performance records are obtained for a content delivery network (CDN) adapted to cache and serve media content requested by the end-user devices. Predicted values of the streaming-media KPI are generated according to the performance records of the CDN and compared to the threshold value of the streaming-media KPI to obtain a comparison. An anomaly is detected according to the comparison, to indicate that a predetermined number of the predicted values of the streaming-media KPI fail to satisfy the threshold value. Other embodiments are disclosed.