Time Series Forecasting for Virtualized Resource Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for estimating disk space utilization in multi-user computing environments fail to accurately account for individual usage patterns, leading to over-provisioning or under-provisioning errors due to aggregated data analysis and the dynamic nature of data, resulting in less accurate forecasting of resource utilization.
Innovation Solution
A method utilizing time series analysis to determine individual seasonally-adjusted predictions for each user, which are then summed to provide accurate future resource usage demands, allowing for precise provisioning decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional aggregated data analysis is used for forecasting resource utilization, then the forecasting process is simple, but the prediction accuracy deteriorates due to over-provisioning and under-provisioning errors
Solution Approach 1:
The patent segments the forecasting process by creating separate time series analysis models for individual users or workloads rather than using a single aggregated model. Each model captures unique usage patterns, seasonality, and trends of specific users, thereby improving prediction accuracy while managing complexity through modular model development and selection
Solution Approach 2:
The patent changes the parameter of analysis from aggregate-level to individual-user-level data. By transforming the forecasting approach to analyze parameters at the individual workload level and then combining results, the system achieves higher accuracy in predicting resource utilization while accounting for diverse usage patterns across different users
2Measurement precision
If a single time series analysis model is used for all users, then the model selection process is simple, but the prediction accuracy deteriorates because different users have different usage patterns
Solution Approach 1:
The patent divides the user base into separate segments, creating individual time series analysis models for each user or workload. This segmentation allows each model to be optimized for its specific usage patterns, seasonality, and trends, significantly improving prediction accuracy compared to a single generic model
Solution Approach 2:
The patent implements a dynamic model selection process where the appropriate time series analysis model is selected based on the characteristics of each user's historical data. This dynamic approach adapts to different usage patterns by choosing the most suitable model for each segment, balancing accuracy with manageable complexity
3Measurement precision
If thickly provisioned disks and reservations are captured at the storage pool level, then the data collection is simple, but the prediction accuracy deteriorates due to over-provisioning error
Solution Approach 1:
The patent segments the data collection process to capture actual usage metrics at the individual user or workload level rather than relying on aggregated storage pool-level data. This segmentation enables accurate measurement of real consumption patterns, eliminating over-provisioning errors that occur when thickly provisioned disks are counted as fully utilized
Solution Approach 2:
The patent extracts actual usage data from the aggregate storage pool metrics by implementing individual monitoring for each user or workload. This extraction process isolates the true consumption patterns from the inflated provisions data, enabling accurate forecasting based on actual usage rather than allocated capacity
Data Source
AI summary
A method for time series analysis of time-oriented usage data pertaining to computing resources of a computing system. A method embodiment commences upon collecting time series datasets, individual ones of the time series datasets comprising time-oriented usage data of a respective individual computing resource. A plurality of prediction models are trained using portions of time-oriented data. The trained models are evaluated to determine quantitative measures pertaining to predictive accuracy. One of the trained models is selected and then applied over another time series dataset of the individual resource to generate a plurality of individual resource usage predictions. The individual resource usage predictions are used to calculate seasonally-adjusted resource usage demand amounts over a future time period. The resource usage demand amounts are compared to availability of the resource to form a runway that refers to a future time period when the resource is predicted to be demanded to its capacity.