Fully Hydrated Backup Metadata Tree for Instant Database Restore

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing methods for database snapshots, particularly incremental snapshots, require significant storage and computational resources for recovery, as they involve daisy-chaining full and incremental backups, leading to lengthy restoration times and resource consumption.

Innovation Solution

The implementation of a fully hydrated backup system using a file system metadata snapshot tree that allows incremental backups to be stored and accessed directly, providing a complete view of the storage volume at a specific point in time, reducing the need for rehydration and enabling almost instantaneous database restore and cloning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If incremental snapshots are used to reduce storage requirements, then storage space is reduced, but recovery time and computational resources increase

Engineering Contradiction:
Improvestorage spaceVSAvoidrecovery time
Core Design Contradiction:
Quantity of substanceVSLoss of time

Solution Approach 1:

The system pre-computes and stores metadata trees during the backup process that contain information about data locations and relationships. This preliminary action enables direct navigation to the correct data blocks during recovery without needing to sequentially process all incremental snapshots, thus reducing recovery time while maintaining storage efficiency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces a metadata tree structure as an intermediary between the incremental snapshots and the recovery process. This metadata tree contains pre-computed information about data block locations, parent-child relationships, and version information, allowing the recovery system to directly access needed data without sequentially applying all incremental changes.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Speed

If full snapshots are performed frequently to enable fast recovery, then recovery speed improves, but storage requirements increase significantly

Engineering Contradiction:
Improverecovery speedVSAvoidstorage space
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The system segments the backup data into incremental snapshots while maintaining a metadata tree that organizes these segments. The metadata tree divides the recovery process into direct lookups of data block locations rather than sequential processing, enabling fast recovery from incremental backups without storing complete full snapshots at all time points.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent adds a metadata dimension to the traditional incremental backup approach. Instead of only storing data blocks, the system creates a parallel metadata tree structure that provides direct access paths through the incremental snapshots. This additional dimensional organization enables O(1) or O(log n) access to recovery data rather than O(n) sequential processing.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Quantity of substance

If incremental snapshots are used to minimize storage, then storage efficiency improves, but computational resources for recovery increase

Engineering Contradiction:
Improvestorage efficiencyVSAvoidcomputational resources
Core Design Contradiction:
Quantity of substanceVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary computational work during the backup phase to build and store metadata trees that encode data block relationships, locations, and version information. This shifts computational burden from the recovery phase to the backup phase, reducing the processing power needed during recovery while maintaining storage efficiency through incremental snapshots.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates and stores copies of metadata information (data block locations, parent-child relationships, version tags) in the metadata tree structure. These metadata copies enable the recovery system to quickly determine which data blocks need to be retrieved and in what order, without needing to perform complex computations during recovery.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP3796174B1Restoring a database using a fully hydrated backup
Publication Date: 2022.10.19 COHESITY INC
  • EP3796174B1 patent drawingFigure 1
  • EP3796174B1 patent drawingFigure 2
  • EP3796174B1 patent drawingFigure 3A

AI summary

A backup of at least a portion of contents of a storage volume that includes data of the database is performed. The backup includes a step of creating a new metadata tree root node. Creating the new metadata tree root node includes copying from another metadata tree root node of a tree data structure corresponding to a previous backup instance, one or more references to one or more lower tier metadata tree nodes associated with the tree data structure corresponding to the previous backup instance.