Differential Cloud Snapshot Restoration Mechanism
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing snapshot restoration solutions require significant time and processing power due to their serial restoration approach, which involves restoring snapshots from the first snapshot to the last, leading to inefficiencies and resource wastage.
Innovation Solution
Implementing a differential cloud snapshot mechanism that restores snapshots in reverse chronological order, comparing metadata files to identify and write only the necessary data, thereby reducing redundant data transfers and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If snapshots are restored in serial order from first to last, then all snapshots can be restored, but restoration time and processing power increase significantly
Solution Approach 1:
The patent inverts the traditional serial restoration approach by restoring snapshots in reverse chronological order (from last to first). This allows the system to start with the most recent snapshot and work backwards, enabling early termination when the desired restoration point is reached, thus significantly reducing restoration time while maintaining completeness.
Solution Approach 2:
The patent performs preliminary actions by first downloading and storing metadata files for all snapshots before the actual restoration process. This preliminary metadata preparation enables efficient comparison and identification of data blocks that need to be restored, reducing processing time during the actual restoration operation.
2Reliability
If all snapshot data is transferred from cloud storage, then complete restoration is achieved, but data transfer volume and processing power are wasted
Solution Approach 1:
The patent extracts and compares metadata files from different snapshots to identify only the specific data blocks that need to be restored. By extracting the necessary information from metadata and comparing offsets, the system transfers only the required data portions rather than all snapshot data, eliminating wasted processing power and energy.
Solution Approach 2:
The patent applies local quality by treating different data blocks differently based on their restoration necessity. Instead of uniformly processing all snapshot data, the system identifies and processes only the specific local regions (data blocks) that contain changes or are required for the restoration point, optimizing resource utilization.
3Productivity
If metadata files are compared to identify necessary data, then redundant data transfers are reduced, but additional processing steps are required
Solution Approach 1:
The patent introduces metadata files as intermediary structures that facilitate the restoration process. These metadata files contain offset information that acts as a mediator between the cloud storage system and the local storage, enabling efficient identification of necessary data blocks without requiring complex direct comparisons of actual data content.
Data Source
AI summary
A method comprises receiving a request to restore a given snapshot of a plurality of snapshots from a cloud storage platform to a storage array. In the method, a first metadata file corresponding to the given snapshot is downloaded from the cloud storage platform, and data corresponding to the given snapshot is written to the storage array. An additional metadata file corresponding to a previous snapshot of the plurality of snapshots with respect to the given snapshot is downloaded. The first metadata file and the additional metadata file are compared to determine whether one or more offsets in the additional metadata file are present in the first metadata file. Data corresponding to any of the one or more offsets determined to be absent from the first metadata file is written to the storage array, and the one or more absent offsets are added to the first metadata file.