Database Defragmentation Scheduling via ML Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Database fragmentation leads to reduced performance, and existing methods for detection and repair often require offline defragmentation, causing performance degradation and inefficiencies due to the need for frequent database queries, which can further degrade performance.
Innovation Solution
A system utilizing machine learning techniques, including a classification model for detecting fragmentation and a time series forecasting model to predict low database utilization periods, schedules defragmentation during these times, minimizing the need for offline procedures and reducing performance impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional defragmentation methods are used, then database fragmentation is repaired, but database performance degrades due to offline procedures and frequent queries
Solution Approach 1:
The system performs preliminary detection of database fragmentation using machine learning classification models on log data, and predicts future low-utilization periods using time series forecasting, before scheduling defragmentation operations. This allows defragmentation to be prepared and executed at optimal times without impacting current performance.
Solution Approach 2:
The system automatically monitors database log data, detects fragmentation conditions, predicts suitable maintenance windows, and schedules defragmentation operations without requiring manual intervention or offline procedures. The database manages its own maintenance needs while remaining online and operational.
2Measurement precision
If frequent database queries are performed for detection, then fragmentation is detected early, but performance further degrades
Solution Approach 1:
The system uses machine learning classification models as intermediaries to analyze database log data and predict fragmentation conditions. Instead of querying the actual database structure directly, the ML model processes log data to detect fragmentation, eliminating the need for performance-degrading direct database queries.
Solution Approach 2:
The system replaces traditional mechanical database queries with machine learning-based analysis of log data. The ML classification model substitutes direct database access mechanisms, allowing fragmentation detection without the performance overhead of frequent structural queries.
3Reliability
If offline defragmentation is scheduled, then defragmentation can be performed without impacting current operations, but database downtime increases
Solution Approach 1:
The system dynamically schedules defragmentation operations based on predicted low-utilization periods using time series forecasting models. Instead of fixed offline maintenance windows, the system adapts to actual database usage patterns, performing defragmentation during naturally low-activity periods while the database remains online.
Solution Approach 2:
The database system performs defragmentation operations while remaining online and operational, eliminating the need for offline procedures. The system manages its own maintenance needs without requiring downtime, allowing continuous availability while performing repair operations.
Data Source
AI summary
Methods, information handling systems and computer readable media are disclosed for detection and repair of fragmentation in databases. In one embodiment, a method includes obtaining log data reflecting transactions in a database, where the log data is generated during operation of the database. The method continues with applying a machine learning classification model to at least a portion of the log data to obtain a first prediction, where the first prediction indicates whether defragmentation of the database should be scheduled. In this embodiment the method also includes using a machine learning time series forecasting model to obtain a second prediction, where the second prediction identifies a future time interval of low relative database utilization, and scheduling a defragmentation procedure for performance during the future time interval of low relative database utilization.