Cloud Database Reorganization Through Pseudo-Device IO Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud-based data base management systems face challenges in reorganizing databases without stopping business operations, as existing replication methods like Amazon EBS snapshots result in performance degradation due to slow data transfer and require business interruption for setup.
Innovation Solution
A cloud data base system utilizing a virtual computing environment with first and second pseudo devices to manage IO operations, allowing seamless switching between volumes without registering new volumes in the DBMS, thus enabling continuous business operations during reorganization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If EBS snapshot method is used for volume replication in cloud, then volume replication is achieved, but data transfer speed is slow and business interruption occurs
Solution Approach 1:
The patent introduces a pseudo-device as an intermediary layer between the DBMS and the actual storage volume. This pseudo-device manages IO operations and enables seamless switching between volumes without requiring slow snapshot-based replication. The pseudo-device acts as a mediator that handles volume switching and IO routing, eliminating the need for traditional snapshot methods and their associated slow data transfer issues.
Solution Approach 2:
The patent replaces the mechanical snapshot-based replication system with a virtualized IO management system using pseudo-devices. Instead of relying on physical snapshot creation and data transfer, the system uses virtual IO redirection and volume switching mechanisms, achieving fast and non-intrusive volume replication without business interruption.
2Reliability
If EBS snapshot method is used for volume replication, then volume replication is achieved, but business interruption is required for setup
Solution Approach 1:
The patent implements preliminary action by pre-creating multiple volumes and configuring the pseudo-device to manage IO operations before actual reorganization is needed. The system prepares volume switching capabilities in advance, allowing reorganization to proceed without business interruption. The pseudo-device is pre-configured to handle volume switching, eliminating the need for disruptive setup procedures.
Solution Approach 2:
The patent ensures continuity of useful action by enabling volume switching and reorganization operations to proceed continuously without business interruption. The pseudo-device maintains active IO handling throughout the reorganization process, allowing database reorganization to occur while business operations continue uninterrupted, thus eliminating the time loss associated with traditional snapshot methods.
3Productivity
If traditional DB reorganization method is used in cloud, then DB reorganization is achieved, but performance degradation occurs
Solution Approach 1:
The patent applies segmentation by separating DB reorganization operations from business operations through the pseudo-device layer. The pseudo-device manages IO operations independently, allowing reorganization to proceed on one volume while business operations continue on another volume without performance degradation. This segmentation enables parallel execution of reorganization and business operations, maintaining high performance throughout the process.
Solution Approach 2:
The pseudo-device serves as an intermediary that protects business operations from reorganization-related performance degradation. It manages IO routing and volume switching in a way that isolates reorganization operations from active business transactions, ensuring that DB reorganization can proceed without affecting the performance and reliability of ongoing business operations.
Data Source
AI summary
The invention provides a cloud data base system and a data management method that can reorganize a DB without stopping and without affecting business in a cloud. The cloud data base system includes a virtual machine provided by a cloud service, an object storage, and a volume configured to store a DB. When the DB is reorganized, the virtual machine switches an issue destination of an IO from a first pseudo device to a second pseudo device, and causes a replication function to create a replicated volume and causes the first pseudo device to reorganize the DB stored in the replicated volume, while causing the second pseudo device to continue to issue an IO to a replication source volume that is a volume storing the DB.