Storage Space Prediction Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage systems lack the ability to predict future usage states, leading to unforeseen shortages of storage space, as users and administrators can only access current and historical data, not future states, making it difficult to manage storage effectively.
Innovation Solution
A method that collects state data at predetermined intervals, compares it to threshold conditions, and uses machine learning to generate an association relationship between historical usage states and time points, allowing for the estimation of future storage space usage and predicting when storage space will be exhausted.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If storage space is continuously expanded to meet user demands, then storage capacity increases, but the ability to predict future storage exhaustion is lost
Solution Approach 1:
The system performs preliminary actions by collecting historical state data and training a machine learning model in advance to predict future storage exhaustion time. This allows the system to proactively identify potential storage issues before they occur, enabling timely intervention and resource allocation planning.
Solution Approach 2:
The system establishes a feedback loop by continuously monitoring current storage state data, comparing it with historical data, and using the difference to update predictions. The prediction results are fed back to users and administrators, creating a closed-loop system that improves storage management decision-making.
2Ease of operation
If only current and historical storage data is monitored, then data collection simplicity is maintained, but future storage state prediction capability is lost
Solution Approach 1:
The system introduces a machine learning model as an intermediary between historical data and future predictions. This intermediary processes historical state data and current differences to generate reliable future storage state predictions, bridging the gap between simple data collection and reliable forecasting.
Solution Approach 2:
The system replaces traditional mechanical monitoring approaches with machine learning-based prediction. Instead of relying solely on direct observation of current and historical data, the system uses ML algorithms to infer future states, enhancing reliability while maintaining operational simplicity.
3Device complexity
If storage space is managed reactively based on current state, then system complexity is reduced, but storage shortages occur unexpectedly
Solution Approach 1:
The system performs preliminary prediction of storage exhaustion time using machine learning models trained on historical data. This allows administrators to take proactive measures before storage shortages occur, improving storage availability without requiring complex real-time intervention systems.
Solution Approach 2:
The system provides self-service storage management by automatically predicting future storage states and alerting users. This reduces the need for complex manual monitoring and intervention systems while maintaining high storage availability through automated insights.
Data Source
AI summary
A group of state data of storage space in a storage system is collected according to a predetermined time interval, the group of state data being collected at a group of time points, respectively, and the group of time points being divided according to the predetermined time interval. The group of state data is compared with a threshold condition of the storage system, the threshold condition representing that storage space in the storage system is to be exhausted. An association relationship between a state of storage space in the storage system and a future time point is generated based on the group of state data in accordance with determining at least one state data in the group of state data satisfies the threshold condition. A state estimate of storage space in the storage system at a specified future time point is obtained based on the generated association relationship.