Database Tablespace Space Management via Local Global Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Database space management is inefficient due to skewed space distribution across tablespaces, leading to outages when one tablespace becomes full despite other tablespaces having available space, often caused by high space consumption rates and monitoring failures, which can result in business disruptions.
Innovation Solution
The system splits tablespaces into local and global partitions, allowing dynamic borrowing and lending of space between them, with algorithms to identify and utilize available space from global partitions to alleviate capacity threats and optimize space utilization, thereby preventing outages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If tablespaces are managed independently with fixed space allocation, then space allocation is simple and predictable, but space distribution becomes skewed and outages occur when one tablespace becomes full despite other tablespaces having available space
Solution Approach 1:
The patent segments each tablespace into two distinct partitions: a local partition for storing objects local to that tablespace, and a global partition for storing objects from other tablespaces. This segmentation allows independent management of local space while enabling shared access to global space, preventing outages caused by skewed space distribution in traditional fixed allocation models.
Solution Approach 2:
The global partition serves multiple functions: it acts as a shared storage pool accessible by multiple tablespaces, provides overflow storage when local partitions are full, and enables dynamic space redistribution. This multi-functionality eliminates the need for complex inter-tablespace space transfer mechanisms while maintaining database availability.
2Reliability
If manual space monitoring and intervention is performed, then space allocation can be optimized, but operational complexity increases and human errors can cause disruptions
Solution Approach 1:
The database system automatically monitors space utilization in local and global partitions and dynamically redirects object storage based on available space thresholds. When a local partition approaches capacity, the system autonomously redirects overflow objects to global partitions without requiring manual intervention, eliminating human errors while maintaining optimized space allocation.
Solution Approach 2:
The system continuously monitors space utilization metrics in local and global partitions and uses this feedback to dynamically adjust object storage locations. This closed-loop feedback mechanism ensures real-time space optimization and prevents outages before they occur, replacing manual monitoring with automated responsive control.
3Adaptability or versatility
If high space consumption rates are allowed, then data storage flexibility increases, but monitoring failures can lead to outages without timely detection
Solution Approach 1:
The system implements continuous monitoring of space utilization thresholds in local and global partitions, providing real-time feedback on space consumption rates. This feedback mechanism enables early detection of high space consumption patterns before they cause outages, maintaining both flexibility and detectability.
Solution Approach 2:
The system pre-establishes space thresholds and automated redirect rules before space consumption issues arise. When monitoring detects approaching thresholds, the system proactively redirects objects to global partitions, preventing outages before they occur and eliminating the need for reactive manual intervention.
Data Source
AI summary
A database is identified, wherein the database has two or more tablespaces. A local partition and a global partition for each tablespace of the two or more tablespaces is created, wherein the created two or more global partitions are included in a global storage pool. A request to move an object to a first local partition of a first tablespace of the two or more tablespaces is received. That an amount of used space of the first local partition is above a first threshold is determined. Responsive to determining that the amount of used space of the first local partition is above the first threshold, at least a portion of the object is stored in the global storage pool.