Synthetic Backup Re-synthesis for Restore Speed and Metadata Scalability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Synthetic full backups in data processing lead to slower restores and tape copies due to disk fragmentation and increased processing time, along with challenges in scaling metadata for billions of indexed entries across incremental and synthetic full backup images, impacting the efficiency of backup operations.
Innovation Solution
A computer-implemented method that performs partial and complete re-synthesis of synthetic backups, using virtual fragments with instructions to combine physical data segments, and prevents deduplication of specific fragments, optimizing the management of synthetic full backups within a deduplication backup storage system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If synthetic full backups are used to reduce full backup time, then backup speed is improved, but restore speed and tape copy speed deteriorate due to disk fragmentation
Solution Approach 1:
The patent performs preliminary re-synthesis of synthetic full backups into traditional full backup formats during off-peak hours. This preliminary action converts fragmented synthetic backups into contiguous traditional backups before restore operations, thereby maintaining fast backup speeds while ensuring fast restore speeds when needed.
Solution Approach 2:
The patent segments the backup management process into two distinct modes: synthetic full backup creation for speed, and selective re-synthesis for restore preparation. By segmenting when re-synthesis occurs (only when needed for restore or tape copy), the system maintains the speed benefits of synthetic backups while eliminating performance degradation during critical operations.
2Productivity
If synthetic full backups are used to reduce full backup time, then backup speed is improved, but processing time to resolve indirect pointer chunks increases
Solution Approach 1:
The system performs preliminary re-synthesis that pre-resolves all indirect pointer references into direct file offsets during off-peak hours. This preliminary resolution of pointers eliminates the need for complex pointer chasing during restore operations, reducing processing time significantly.
Solution Approach 2:
The patent creates a copied version of the synthetic full backup in traditional full backup format with directly accessible file pointers. Instead of resolving pointers on-demand during restores, the system maintains a copied traditional format version that has pre-resolved pointers, eliminating processing delays.
3Productivity
If synthetic full backups are used, then full backup time is reduced, but metadata scalability becomes difficult due to billions of indexed entries
Solution Approach 1:
The patent extracts and removes the complex metadata indexing structure from synthetic full backups during re-synthesis. By converting to traditional backup format, the system eliminates the need to manage billions of indexed entries and their complex relationships, simplifying metadata management while maintaining backup speed benefits.
Solution Approach 2:
The system performs preliminary conversion of synthetic backups to traditional format, which pre-processes and simplifies the metadata structure before storage. This preliminary metadata simplification reduces the complexity burden for future operations by eliminating the need to maintain complex indexed entry relationships.
Data Source
AI summary
A method, a system, and a computer program product for executing synthetic backup processes and deduplication backup storage with landing zone. A synthetic backup of a data file is received. A partial re-synthesis of the synthetic backup of the data file is performed. A total size of the partial re-synthesized backup of the data file and the received synthetic backup is determined. A size of a complete re-synthesis of the synthetic backup of the data file is computed. The complete re-synthesis of the synthetic backup of the data file is performed when the determined total size exceeds the computed size of the complete re-synthesis of the synthetic backup of the data file.