Pointer-Based Database Backup Mounting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database backup solutions require provisioning storage resources for point-in-time copies and waiting for large data sets to traverse networks, making them inefficient for seamless database recovery and access.
Innovation Solution
The method generates self-contained database backups that include both data and metadata, allowing for seamless live-mounting by retaining physical point-in-time data and metadata on backup storage, and using pointer-based snapshots to mount a file system on the target device without needing to transfer large data sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional database backup solutions are used to create point-in-time copies, then database data can be recovered, but storage resources must be provisioned and large data sets must traverse networks, causing delays
Solution Approach 1:
The patent uses pointer-based snapshots that copy only metadata (pointers) rather than actual data blocks. The backup storage system creates a snapshot directory with pointers to data blocks stored in a data store, enabling rapid access without transferring large data sets across the network. This resolves the contradiction by maintaining recovery capability while eliminating time-consuming data traversal.
Solution Approach 2:
The patent segments backup data into separate components: control files stored in a backup storage system and data blocks stored in a data store. The snapshot mechanism separates metadata (pointers to data blocks) from actual data, allowing the metadata to be copied and accessed rapidly while data blocks remain in their original locations. This segmentation enables fast recovery without provisioning new storage or transferring large data sets.
2Reliability
If traditional backup solutions create point-in-time copies, then database data can be restored, but storage resources must be provisioned for the copies
Solution Approach 1:
The patent creates pointer-based snapshots that copy only metadata (pointers) rather than actual data blocks. The snapshot directory contains pointers to data blocks that already exist in the data store, eliminating the need to provision new storage resources for copies. This resolves the contradiction by maintaining recovery capability while avoiding the complexity of storage provisioning.
Solution Approach 2:
The patent implements a nested structure where the snapshot directory is contained within the backup storage system, and the pointers within the snapshot directory reference data blocks stored externally in the data store. This nested architecture allows the backup system to reference existing data without duplicating it, eliminating the need for additional storage provisioning while maintaining full recovery capability.
3Reliability
If large data sets are transferred for backup recovery, then complete data can be restored, but the process becomes inefficient and slow
Solution Approach 1:
The patent copies only pointer metadata rather than large data blocks during backup operations. The snapshot mechanism creates a lightweight copy of the directory structure with pointers to actual data blocks, enabling rapid restoration without transferring large data sets. This resolves the contradiction by maintaining data completeness through pointer references while dramatically improving recovery efficiency.
Solution Approach 2:
The patent introduces pointer-based snapshots as an intermediary between the backup storage system and the actual data blocks. These snapshots act as mediators that provide metadata about data locations without containing the data itself, enabling efficient access to complete data information without transferring large data sets during backup and recovery operations.
Data Source
AI summary
In general, in one aspect, the invention relates to a method for accessing backup data includes receiving a live mount connection request from a client device, in response to the live mount connection request: generating a directory, populating the directory with a backup control file image and a plurality of pointers to backup data file images in a second directory, wherein the backup control file image is obtained from the second directory, restoring at least a portion of the backup data file images in the directory using the backup control file image and at least a portion of the backup data file images in the second directory, and providing after the restoring, a handle to the directory to the client device.