Database Object Growth Forecasting for Abnormal Space Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increased size and complexity of storage spaces in big data environments make it difficult to identify and address issues related to wasted resources, performance problems, and outages, particularly due to abnormal growth of database objects.
Innovation Solution
Implementing a non-linear regression model to forecast object growth and a classification model to identify abnormal growth, followed by automated operations such as index creation and query tuning to maintain query performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If storage spaces become larger and more complex to accommodate big data, then data capacity increases, but it becomes difficult to identify and address problems related to wasted resources, performance issues and outages
Solution Approach 1:
The system performs preliminary actions by continuously monitoring database object growth patterns and predicting future size requirements before actual problems occur. The machine learning models analyze historical data to forecast when storage capacity will be exceeded, allowing proactive capacity planning and prevention of performance issues before they manifest.
Solution Approach 2:
The system implements feedback mechanisms where actual object growth is compared against predicted growth patterns. The machine learning models continuously learn from actual performance data and adjust their predictions accordingly, creating a closed-loop system that improves accuracy over time and enables real-time detection of anomalous growth patterns that indicate potential problems.
2Reliability
If conventional capacity management techniques are used, then storage space is monitored, but query performance deteriorates due to inability to detect and address abnormal growth
Solution Approach 1:
The patent replaces conventional mechanical monitoring approaches with machine learning-based prediction systems. Instead of simply monitoring storage usage metrics, the system uses neural networks and time-series analysis to predict future growth patterns, enabling more intelligent capacity management that maintains query performance while improving management efficiency.
Solution Approach 2:
The system changes the parameters of capacity management by transitioning from static threshold-based monitoring to dynamic prediction-based management. The machine learning models continuously adapt to changing data access patterns and growth rates, allowing the system to optimize storage allocation and maintain performance under varying workloads.
Data Source
AI summary
Methods, apparatus and processor-readable storage media for object level space forecasting are provided herein. An example computer-implemented method includes determining a predicted growth of an object in a database using a non-linear regression model, wherein the non-linear regression model is trained using a set of data records indicative of a size of the object over time; identifying a threshold indicative of abnormal growth of the object using a classification model, wherein the classification model is trained at least in part on one or more of the data records that are labeled as anomalous based at least in part on one or more prior values of the predicted growth and actual growth of the object; and performing one or more automated operations in response to detecting that the predicted growth of the object exceeds the threshold.