Backup Data Object Consolidation for Restore Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deduplication in backup systems leads to fragmented data storage, reducing restore throughput as the number of backups increases, making routine restore operations inefficient.
Innovation Solution
Implementing pseudo full backups by merging data objects from multiple containers into fewer, more accessible containers, and using doubly sorted queuing to optimize disk access patterns during restore operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If deduplication is used to reduce storage space and bandwidth consumption, then storage efficiency is improved, but restore throughput deteriorates as the number of backups increases
Solution Approach 1:
The patent segments backup data into containers grouped by backup ID, where each container holds data objects from the same backup. This segmentation allows the system to efficiently locate and retrieve data from specific backups without scanning all deduplicated data, thereby maintaining high restore throughput even as the number of backups increases.
Solution Approach 2:
The patent introduces a new organizational dimension by creating containers that group data objects according to their backup ID, adding a hierarchical layer above the traditional deduplication structure. This dimensional change enables efficient restore operations by allowing direct access to containers associated with a specific backup, avoiding the performance degradation caused by linear search through all deduplicated data.
2Quantity of substance
If multiple backups are performed with deduplication, then storage efficiency is improved, but data becomes fragmented across multiple containers, reducing restore efficiency
Solution Approach 1:
The patent merges data objects from the same backup into unified containers, combining scattered data objects into organized groups. This merging process maintains storage efficiency through deduplication while improving restore efficiency by consolidating related data objects into accessible containers that can be retrieved as a unit.
3Quantity of substance
If data objects are distributed across many containers due to deduplication, then storage capacity is optimized, but access time during restore increases
Solution Approach 1:
The patent performs preliminary organization of data objects into containers grouped by backup ID during the backup process. This preliminary action ensures that when restore operations are needed, data objects are already organized and can be quickly located and retrieved, significantly reducing restore access time while maintaining optimized storage capacity utilization.
Data Source
AI summary
Containers that store data objects that were written to those containers during a particular backup are accessed. Then, a subset of the containers is identified; the containers in the subset have less than a threshold number of data objects associated with the particular backup. Data objects that are in containers in that subset and that are associated with the backup are copied to one or more other containers. Those other containers are subsequently used to restore data objects associated with the backup.