Partial Database Reorganization for Workload Performance Declines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database storage systems require periodic reorganizations that are often performed at the table or partition level, which can be disruptive and do not address granular workload performance issues, leading to inefficiencies and performance degradation.
Innovation Solution
A system and method for performing online partial reorganizations of databases based on identified performance-sensitive workloads, analyzing performance information to determine and address issues such as clustering, hot spots, overflow, and space usage at a more granular level, allowing for localized reorganizations without taking the database offline.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If periodic full-table or partition-level reorganization is performed, then database performance is improved, but database downtime increases and availability decreases
Solution Approach 1:
The patent divides the database into multiple workloads and performs reorganization on individual workloads or portions of workloads rather than the entire database. This segmentation allows reorganization to proceed in smaller, isolated units while the rest of the database remains accessible, thereby reducing overall downtime and maintaining availability.
Solution Approach 2:
The patent implements partial reorganization by selecting only specific workloads or portions of workloads that require reorganization based on performance metrics, rather than reorganizing the entire database. This partial action approach reduces the time required for reorganization while still achieving performance improvement for the affected portions.
2Productivity
If full-table reorganization is performed, then overall database performance is improved, but the reorganization process becomes inefficient and time-consuming
Solution Approach 1:
The patent applies reorganization selectively to specific workloads or portions of workloads that exhibit performance degradation, rather than uniformly reorganizing the entire database. This localized approach focuses resources on areas that need improvement, increasing reorganization efficiency and reducing overall duration.
Solution Approach 2:
The patent performs partial reorganization on selected workloads based on performance analysis, avoiding the time-consuming process of reorganizing the entire database. This partial action maintains productivity by concentrating efforts where they are most needed while reducing total reorganization time.
3Reliability
If offline reorganization is performed, then performance issues are addressed, but continuous database operation is disrupted
Solution Approach 1:
The patent segments the database into multiple workloads and performs reorganization on individual workloads while others remain operational. This segmentation enables performance maintenance through targeted reorganization without disrupting continuous database operation, as other workloads continue to serve requests.
Solution Approach 2:
The patent enables continuous database operation during reorganization by allowing non-affected workloads to continue processing requests. The reorganization process is performed on selected workloads without taking the entire database offline, maintaining the continuity of useful action across the system.
4Measurement precision
If traditional reorganization methods are used, then performance declines are addressed, but granular workload performance issues are not effectively targeted
Solution Approach 1:
The patent divides the database into multiple workloads and analyzes performance metrics at the workload level, enabling precise identification of performance issues in specific workloads. This segmentation allows targeted reorganization of only those workloads exhibiting performance declines, improving both measurement precision and adaptability to granular workload needs.
Solution Approach 2:
The patent applies performance analysis and reorganization selectively to specific workloads based on their individual performance metrics, rather than treating the entire database uniformly. This localized approach enables precise identification and targeted remediation of granular workload performance issues.
Data Source
AI summary
A system, computer program product, and method are provided for driving database reorganizations subject to determined workloads, and, more specifically, to providing an online partial reorganization of a database to improve workload performance. The method includes collecting performance information for one or more performance sensitive workloads associated with one or more portions of a database and analyzing the collected performance information. For each of the one or more performance sensitive workloads, indications of performance declines thereof are determined. Subject to the determined performance declines, at least one of the one or more database portions affected by the determined performance declines are determined. Subject to the database portions determination, the determined performance declines are classified. Subject to the classification of the performance declines, one or more partial reorganizations of the one or more database portions is executed.