Virtual Storage Capacity Risk Forecasting With Probabilistic Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage capacity forecasting methods in virtualized systems are inaccurate due to unpredictable trends in data usage, making it challenging to quantify changes in utilization patterns, leading to reduced confidence in estimates.
Innovation Solution
A probabilistic model based on stochastic processes, utilizing a moving average variance to estimate changes in behavior, calculates the probability of exceeding storage capacity within different time frames, adapting to discontinuities in storage utilization and allowing for different risk profiles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If time series forecasting methods are used to predict full capacity events, then storage capacity forecasting is performed, but accuracy is reduced due to unpredictable trends in capacity
Solution Approach 1:
The patent transitions from static time series forecasting to a dynamic probabilistic model that continuously adapts to changing storage utilization patterns. The model incorporates moving average variance calculations and stochastic processes that evolve with system behavior, allowing it to respond to discontinuities and unpredictable trends in capacity usage while maintaining forecasting accuracy.
Solution Approach 2:
The invention changes the fundamental parameters of the forecasting approach by moving from deterministic point predictions to probabilistic distributions. It introduces multiple parameters including mean utilization, variance, confidence levels, and time-based probability thresholds, enabling the system to capture uncertainty and adaptability in storage capacity forecasting.
2Reliability
If deterministic time series models are used, then a specific time of capacity exhaustion is predicted, but confidence in estimates is reduced
Solution Approach 1:
The patent introduces probabilistic distributions and statistical intermediaries between the raw time series data and the capacity exhaustion prediction. By using moving average variance, confidence intervals, and probability density functions as intermediaries, the model preserves uncertainty information rather than losing it, thereby increasing confidence in the estimates through quantified reliability metrics.
3Measurement precision
If storage capacity forecasting tools are built to avoid operational issues, then capacity management is performed, but accuracy is compromised due to highly unpredictable trends
Solution Approach 1:
The patent replaces traditional mechanical time series forecasting mechanisms with a probabilistic statistical framework. Instead of relying on fixed mathematical models that assume predictable patterns, it uses stochastic processes, variance analysis, and probability distributions that naturally accommodate unpredictable trends and make the difficulty of measuring utilization changes a feature rather than a bug of the system.
Data Source
AI summary
Examples of systems are described for calculating a probability of exceeding storage capacity of a virtualized system in a particular time period using probabilistic models. The probabilistic models may advantageously take variances of storage capacity into consideration.