Time-Series Search for Proactive Computer Performance Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computer monitoring systems are reactive and fail to predict performance issues proactively, relying on detecting problems after they occur rather than anticipating them.
Innovation Solution
A computer monitoring system that predicts near-term and long-term performance by comparing current time series data with historical data using cosine similarity and dimensionality reduction techniques, allowing for proactive identification of events and anomalies through a searchable database and visualization of application behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional monitoring systems detect problems after they occur, then the system is simple and reactive, but the system cannot predict performance issues proactively
Solution Approach 1:
The system performs preliminary analysis by comparing current performance metrics with historical data patterns before problems actually occur. By identifying similar historical patterns and predicting their outcomes, the system takes proactive action to prevent performance issues rather than reacting after they happen, thus improving reliability while managing complexity through structured pattern matching
Solution Approach 2:
The system adds a temporal dimension to performance monitoring by incorporating historical time-series data and pattern recognition. This transforms the monitoring approach from checking current state only to analyzing trends and patterns across time, enabling prediction capabilities while organizing complexity through structured historical comparisons
2Measurement precision
If the system stores and searches large amounts of historical time series data, then prediction accuracy improves, but data storage and processing requirements increase
Solution Approach 1:
The system extracts only the essential pattern information from historical data rather than storing and processing all raw time series data. By identifying and storing characteristic patterns and their outcomes, the system maintains prediction accuracy while significantly reducing data storage requirements, as it only keeps the distilled pattern knowledge needed for future comparisons
3Speed
If the system performs real-time comparison with historical data, then prediction timeliness improves, but computational load increases
Solution Approach 1:
The system performs preliminary organization of historical data into structured patterns during off-peak times, creating a ready-to-use pattern library. During real-time operation, it only needs to compare current metrics against these pre-organized patterns rather than searching through all historical data, thus achieving fast prediction response with reduced computational load
Solution Approach 2:
The system segments historical data into distinct, identifiable patterns with specific characteristics and outcomes. This segmentation allows for efficient indexing and rapid matching during real-time comparisons, reducing computational requirements while maintaining prediction timeliness through structured pattern categories
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer monitoring system may predict near term and long term performance by comparing a segment of current time series data with previously observed time series to find matching segments. From a matching segment, a prediction of performance may be made by examining later observations in the time series. Each time series element may include a large number of parameters, and one mechanism for comparing segments may be treating the elements as multi-dimensional vectors and using cosine similarity for finding significant matches. A deployment mechanism may store time series segments in a searchable database, and search the database with a newly observed time series segment for matches.