Interpolating Performance Data Using System State Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current interpolation techniques in computing systems, such as arithmetic mean and linear regression, are not accurate due to their reliance on limited sampled performance data, failing to account for various system states and factors affecting performance metrics like network IO bandwidth.
Innovation Solution
A method that determines the system state of a computing system using clustering techniques and neural networks to generate performance robustness values, allowing for accurate interpolation of performance metrics by adjusting interpolated values based on data clusters associated with specific system states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If arithmetic mean or linear regression interpolation techniques are used, then the interpolation process is simple, but the accuracy of performance metric interpolation is poor
Solution Approach 1:
The patent transforms the interpolation approach by changing from simple arithmetic parameters (mean, linear regression) to parameters that incorporate system state information. It introduces performance robustness values and system state clustering as new parameters to enhance interpolation accuracy while maintaining computational feasibility through structured data organization and state-based categorization.
2Productivity
If limited sampled performance data is used, then the data collection process is efficient, but the reliability of interpolation results is poor
Solution Approach 1:
The patent introduces system state information as an intermediary element between the limited sampled performance data and the interpolation results. By clustering system states and using them as mediators to guide the interpolation process, the system enhances the reliability of results without requiring additional performance data collection, thus maintaining data collection efficiency while improving result reliability.
3Device complexity
If traditional interpolation methods are used, then the computational complexity is low, but the adaptability to different system states is poor
Solution Approach 1:
The patent segments the system operation space into distinct system states through clustering. By dividing the continuous system behavior into discrete state clusters, the interpolation method can adapt to different system conditions by selecting appropriate clusters. This segmentation approach maintains computational efficiency by using pre-computed cluster information while significantly improving adaptability to various system states.
Data Source
AI summary
Aspects of the invention include determining an event associated with a computing system, the event occurring at a first time, obtaining system data associated with the computing system, determining a system state of the computing system at the first time based on the system data, determining, based on the system state, two or more system data clusters comprising clustered system data associated with the system state of the computing system, determining, via an interpolation algorithm, an interpolated data value for the first time based on the system data, and adjusting the interpolated data value based on a determination that the interpolate data value is outside the two or more system data clusters.