Selective Entity Backup for Distributed Storage Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database backup systems face inefficiencies in storage space usage and consistency in creating point-in-time backups, especially in distributed storage systems where coordinating backup versions across multiple datacenters is challenging.
Innovation Solution
A system and method that selectively store backup versions of modified entities, generating database snapshots by merging backup versions from a backup portion with current versions from the production portion, thereby avoiding duplication and ensuring consistency without excessive storage usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a backup of an entire database is created by copying all entities, then a complete backup is obtained, but storage space is wasted on entities that have not been modified
Solution Approach 1:
The patent extracts only the modified entities from the database and stores them as backup versions, rather than copying the entire database. This is achieved by tracking changes to entities and selectively preserving only those that have been modified since the last backup, thereby eliminating redundant storage of unchanged data while maintaining complete backup capability.
Solution Approach 2:
The backup system dynamically adapts its behavior based on entity modification status. Instead of a static full database copy, the system continuously monitors which entities have changed and adjusts the backup scope accordingly, storing only the delta of modified entities rather than the entire dataset.
2Reliability
If a backup of all data in a distributed database is created, then a consistent point-in-time backup is obtained, but the backup process takes a non-trivial amount of time
Solution Approach 1:
The patent extracts only modified entities across distributed datacenters for backup, rather than synchronizing the entire database state. This selective extraction approach maintains consistency by focusing coordination only on entities that changed, significantly reducing the time required to assemble the backup while ensuring all modifications are captured.
Solution Approach 2:
Instead of performing a complete database copy across all datacenters, the system applies partial action by backup only the necessary modified entities. This partial backup approach is sufficient to achieve consistent point-in-time recovery without the time cost of synchronizing the entire distributed database state.
3Reliability
If the database is placed into read-only mode for backup, then a consistent backup is created, but database usability is substantially degraded
Solution Approach 1:
The patent segments the backup operation from the production database workload. By separating the backup process into an independent phase that processes only modified entities, the system eliminates the need to place the entire database into read-only mode, allowing continuous read/write operations during backup while maintaining consistency through change-tracking mechanisms.
Solution Approach 2:
The system performs preliminary tracking of entity modifications before the actual backup operation. By maintaining a record of which entities have changed since the last backup, the system can execute the backup quickly without blocking production operations, as the backup process only needs to retrieve and store the pre-identified modified entities.
4Quantity of substance
If backup versions of entities are generated as they are modified, then storage space is saved, but coordinating timing across datacenters is difficult
Solution Approach 1:
The patent implements a feedback mechanism where each datacenter tracks and reports the timestamp of its last backup operation. This feedback information is used to determine which entities have been modified since the last coordinated backup point, enabling automatic identification of changes without complex real-time coordination protocols across distributed systems.
Solution Approach 2:
Each datacenter preliminarily tracks the last backup timestamp and maintains a list of modified entities before coordination is needed. When backup coordination is required, these pre-collected timestamps and change lists are exchanged, allowing efficient determination of what needs to be backed up without complex real-time synchronization protocols.
Data Source
AI summary
A method and a server system for creating consistent snapshots in a distributed storage system are disclosed. The server system stores entities in databases that includes a production portion for storing current versions of entities and a backup portion for storing backup versions of modified entities, receives a request for a snapshot of a portion of the databases at a predefined snapshot time. In response to the request, the server system generates a snapshot of the databases, the snapshot including: a backup version of an entity retrieved from the backup portion and a current version of another entity retrieved from the databases and produces a response to the request based on the snapshot. When there are several backup portions, a backup portion may be used separately or in conjunction with other backup portions to provide required backup versions to generate a snapshot.