Storage Restore System Parallel Directory Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional storage restore systems face limitations in utilizing parallel processing for fast recovery of directories and files due to the sequential recovery requirement from highest to lowest directory layers, restricting the full utilization of parallel processing benefits even with multiple nodes.
Innovation Solution
A storage restore system and method that generates a directory indicated by parent directory information in object data, allowing files or directories to be restored under the newly created directory, thereby bypassing the sequential recovery constraint and enabling parallel processing across multiple nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sequential recovery from highest to lowest directory layers is used, then the file system structure is properly reconstructed, but the recovery speed is limited and parallel processing cannot be fully utilized
Solution Approach 1:
The patent segments the recovery process into independent operations by introducing a work queue that holds multiple directory recovery tasks. Each directory recovery operation can be processed independently once its parent directory is ready, allowing multiple nodes to work in parallel on different directories simultaneously rather than strictly sequentially through the hierarchy.
Solution Approach 2:
The patent performs preliminary actions by pre-processing the object data to extract and organize directory hierarchy information before recovery begins. The work queue is pre-populated with recovery tasks, and parent directory structures are prepared in advance, enabling child directories to be recovered in parallel once their dependencies are satisfied.
2Productivity
If multiple nodes are used for parallel processing, then processing capacity is increased, but the sequential dependency from highest to lowest layers prevents full utilization of parallel processing benefits
Solution Approach 1:
The patent introduces dynamic task allocation and management through the work queue mechanism. Multiple nodes can dynamically pick up and execute recovery tasks from the queue based on their availability and capacity. The system dynamically balances the workload across nodes, allowing flexible parallel processing that adapts to the actual execution state rather than following a rigid sequential schedule.
Solution Approach 2:
The work queue serves as an intermediary between the object data storage and the multiple recovery nodes. It decouples the sequential nature of directory hierarchy from the parallel processing capability of multiple nodes, allowing nodes to independently consume tasks from the queue while maintaining proper parent-child directory relationships through the queued task structure.
3Adaptability or versatility
If object data is stored without file system structure, then storage flexibility is improved, but recovery requires sequential processing from highest to lowest layers
Solution Approach 1:
The patent performs preliminary processing of object data to extract directory hierarchy information and reconstruct the file system structure in the work queue before recovery begins. This pre-processing step transforms the flexible but structureless object data into an organized recovery task queue that guides the recovery process, eliminating the need for complex sequential processing during actual recovery execution.
Data Source
AI summary
A storage restore system includes a processor, a memory, and a restore part configured to generate, in a recovery-destination storage area to which to recover a file or a directory stored in a designated storage area and indicated by object data including the file or the directory and parent directory information which is information on a directory to which the file or the directory belongs, a directory indicated by the parent directory information in the object data, and configured to generate the file or the directory indicated by the object data under the directory thus generated.