Featureless Telemetry Image Matching for Time-Series Outage Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems fail to accurately and efficiently predict system failures in technology infrastructure by neglecting the analysis of a series of previous events in a time series, leading to reduced accuracy and computational inefficiency in failure detection.
Innovation Solution
A computing platform trains an image comparison model using telemetry data to generate telemetry state images, which are classified using parallel processing to identify the likelihood of failure, and sends preemptive resolution commands to prevent predicted failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a time series of previous events is analyzed for failure prediction, then the accuracy of failure detection is improved, but the computational efficiency deteriorates due to slow and unparallelizable processing
Solution Approach 1:
The patent segments the time series data into multiple independent telemetry state images, each representing a specific time point. This segmentation allows parallel processing of different time periods, transforming the previously sequential analysis into concurrent operations that maintain accuracy while improving computational efficiency.
Solution Approach 2:
The patent transforms the time series data into a visual dimension by generating telemetry state images that plot telemetry parameters over time. This dimensional transformation enables parallel processing techniques to analyze failure patterns across different time points simultaneously, resolving the contradiction between accuracy and computational efficiency.
2Measurement precision
If a time series of previous events is analyzed for failure prediction, then the accuracy of failure detection is improved, but the processing time increases due to sequential analysis requirements
Solution Approach 1:
The patent divides the time series into discrete telemetry state images that can be processed independently. This segmentation enables parallel processing across different time points, significantly reducing the total processing time while maintaining the ability to detect failure patterns that require temporal context.
Solution Approach 2:
The patent performs preliminary processing by pre-computing and storing telemetry state images from historical data. This allows the system to quickly compare current state images against pre-processed historical patterns, reducing real-time processing time while maintaining high accuracy in failure detection.
3Device complexity
If traditional image comparison methods are used without parallel processing, then the simplicity of the approach is maintained, but the scalability deteriorates
Solution Approach 1:
The patent segments the image comparison process into independent operations that can be executed in parallel. Each telemetry state image can be processed simultaneously using the same comparison logic, enabling the system to scale horizontally by adding more processing units without increasing overall complexity.
Solution Approach 2:
The patent creates a universal image comparison model that can process any telemetry state image using the same parallel processing framework. This universal approach allows the system to scale to handle larger datasets and more complex analysis requirements while maintaining the same operational simplicity through standardized parallel operations.
Data Source
AI summary
A computing platform may train an image comparison model to predict system failure for technology infrastructure based on telemetry state images. The computing platform may receive telemetry data for the plurality of computing systems over a period of time. The computing platform may generate, based on the telemetry data and for each parameter represented in the telemetry data, a telemetry state image, where each telemetry state image plots the period of time on an x axis, plots the plurality of computing systems on a y axis, and is specific to a respective parameter represented in the telemetry data. The computing platform may classify, using the image comparison model and using parallel processing, the telemetry state images. The computing platform may identify a likelihood of failure for the technology infrastructure, and may cause modification of operations at one or more of the plurality of computing systems to prevent a predicted failure.