Deduped Cloud Object Defragmentation for Retention-Locked Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing deduplication systems fail to efficiently clean up and defragment deduplicated data in cloud storage due to the presence of mixed retention locked (RL) segments, leading to increased storage costs and inefficiencies as expired segments remain locked with active segments, preventing garbage collection.
Innovation Solution
Implement methods to defragment deduplicated objects based on retention time, separating segments into new objects based on lock status and expiration criteria, allowing garbage collection to delete expired segments and reclaim storage space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deduped objects contain mixed RL expired and RL active segments, then storage space is retained for compliance, but garbage collection cannot delete the objects leading to storage fragmentation and increased costs
Solution Approach 1:
The patent segments deduped objects into separate components based on retention lock status. It identifies and separates RL expired segments from RL active segments within the same object, allowing independent management of each segment type. This segmentation enables the system to release expired segments while preserving active ones, resolving the contradiction between compliance retention and storage efficiency.
Solution Approach 2:
The patent extracts RL expired segments from mixed deduped objects and places them into a separate collection or new object. This extraction process removes the blocking factor (expired segments) that prevents garbage collection, while the remaining RL active segments continue to satisfy compliance requirements. The extracted expired segments can then be deleted to reclaim storage space.
2Reliability
If cloud space is fragmented with mixed RL expired and RL active segments, then compliance requirements are met, but cleanup processes cannot remove expired data leading to increased storage costs
Solution Approach 1:
The system segments storage space by retention lock status, creating separate logical groupings for RL expired and RL active segments. This segmentation allows the storage system to track and manage compliant data separately from expendable data, enabling cost-effective cleanup operations that don't compromise compliance adherence.
Solution Approach 2:
The patent implements a mechanism to discard RL expired segments after verifying they are no longer needed for compliance, while recovering the storage space they occupied. The system recovers space by deleting expired segments that are no longer referenced by any active backup files, thereby improving storage capacity utilization without violating compliance requirements.
3Reliability
If objects contain both expired and active segments, then data protection is maintained, but defragmentation and cleanup operations are blocked
Solution Approach 1:
The patent applies segmentation to enable cleanup operations by dividing objects into processable units based on segment expiration status. It creates separate processing streams for RL expired and RL active segments, allowing cleanup operations to proceed on expired segments independently without blocking the entire object, thus improving productivity while maintaining data protection for active segments.
Solution Approach 2:
The system extracts the blocking factor (RL expired segments) from the object before attempting cleanup operations. By removing or isolating expired segments first, the patent enables defragmentation and cleanup processes to execute efficiently on the remaining structure, while the extracted expired segments are handled separately through deletion or archival processes.
Data Source
AI summary
One example method includes identifying a cloud object as a potential candidate for defragmentation, evaluating the cloud object to determine what portion of segments of the cloud object are expired, when the portion of expired segments meets or exceeds a threshold, segregating the expired segments and unexpired segments of the cloud object, creating a first new cloud object that includes only unexpired segments, creating a second new cloud object that includes only expired segments, and deleting the cloud object from storage.