Hybrid Storage Capacity Prediction Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current information handling systems face challenges in predicting storage capacity effectively, often resulting in last-minute alerts that do not provide sufficient warning time to prevent storage space shortages, and complex models require significant data and processing capacity, limiting real-time predictions.
Innovation Solution
A hybrid prediction model is implemented, combining a lightweight predictive model on the storage system for real-time short-term predictions using algorithms like linear regression and a resource-intensive model at a remote system for long-term predictions using LSTM or XGBoost, allowing for accurate capacity forecasting without overwhelming processing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a complex prediction model (LSTM/XGBoost) is used for storage capacity forecasting, then prediction accuracy is improved, but processing resource requirements increase significantly
Solution Approach 1:
The patent divides the prediction model into two separate segments: a lightweight model (linear regression) deployed locally on the storage system for real-time predictions, and a complex model (LSTM/XGBoost) deployed remotely on a separate processing system for accurate long-term forecasts. This segmentation allows each model to operate in its optimal environment without resource conflicts.
Solution Approach 2:
A remote processing system acts as an intermediary between the storage system and the complex prediction model. The remote system receives usage information from the storage array, processes it through the resource-intensive LSTM or XGBoost model, and returns predictions to the storage system, thereby isolating the computational burden from the storage system itself.
2Speed
If real-time predictions are implemented on the storage system, then immediate capacity awareness is achieved, but system processing capacity is overwhelmed
Solution Approach 1:
The patent separates prediction functions into two segments based on time horizon and resource requirements: short-term real-time predictions using a lightweight linear regression model that runs locally on the storage system with minimal resource consumption, and long-term accurate predictions using a complex model that runs remotely.
Solution Approach 2:
The patent employs a simple, computationally inexpensive linear regression model for real-time predictions on the storage system. This lightweight model consumes minimal processing power and can be executed frequently without burdening the system, providing adequate short-term forecasting capability.
3Loss of time
If sufficient warning time is provided for storage capacity planning, then proactive storage management is enabled, but complex models require too much data and processing capacity
Solution Approach 1:
The patent segments the forecasting task into short-term and long-term components, with the remote processing system handling long-term predictions using complex models to provide advance warning (e.g., 30-60 day forecasts), while the local system handles immediate operational needs.
Solution Approach 2:
The remote processing system performs preliminary long-term capacity forecasting using complex models to provide advance warning before storage exhaustion occurs. This preliminary action enables proactive storage planning and capacity expansion before the storage system actually runs out of space.
Data Source
AI summary
An information handling system includes a storage system and a remote processing system. The storage system includes a storage array and a local storage usage predictor. The local storage usage predictor receives usage information from the storage array, and predicts a first usage prediction for the storage array based upon the usage information. The remote processing system includes a remote storage usage predictor remote from the storage system. The remote storage usage predictor receives the usage information and to predicts a second usage prediction for the storage array based upon the usage information.