Granular Self-Healing for Distributed File Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional self-healing solutions in distributed file systems lock entire files during the self-healing process, leading to prolonged downtime and potential timeouts, especially in cloud environments where large files may cause client devices to wait excessively for access.
Innovation Solution
A granular self-healing module that acquires a full-file lock to inspect change logs, iteratively locks and releases regions of the file, allowing write access to unlocked regions and redirecting requests to fresh copies, enabling incremental self-healing without locking the entire file.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional self-healing solutions lock an entire file during the self-healing process, then data consistency is ensured, but client devices experience prolonged wait times and potential timeouts
Solution Approach 1:
The patent divides the file into multiple regions and applies locks granularly at the region level rather than locking the entire file. This allows different regions to be accessed simultaneously by different operations, reducing the overall wait time for clients while maintaining data consistency within each locked region during self-healing operations.
Solution Approach 2:
Instead of locking the entire file (excessive action), the patent locks only the specific regions that need self-healing (partial action). This minimizes the impact on client access to other regions of the file that do not require self-healing, thereby reducing client wait time while ensuring data consistency in the regions being healed.
2Reliability
If traditional self-healing solutions lock an entire file for large files, then complete data recovery is ensured, but the self-healing process takes a long period causing virtual machine instances to hang
Solution Approach 1:
The file is segmented into multiple regions that can be self-healed independently and in parallel. This reduces the overall self-healing duration by allowing multiple regions to be processed simultaneously, while still ensuring complete data recovery across the entire file through systematic region-by-region healing.
Solution Approach 2:
The patent enables continuous self-healing operations across different regions without interruption. By processing regions in parallel and allowing overlapping operations where safe, the system maintains continuous progress on self-healing the entire file, reducing total duration while ensuring complete recovery.
3Reliability
If traditional self-healing solutions lock an entire file, then data integrity during healing is maintained, but productivity of the file system decreases due to blocked access
Solution Approach 1:
By segmenting the file into regions and locking only the necessary regions during self-healing, the system maintains data integrity in healed regions while allowing concurrent access to unhealed regions. This increases file system throughput and productivity by enabling parallel operations that were previously blocked by full-file locks.
Solution Approach 2:
The patent applies locking only to the extent necessary for maintaining data integrity during self-healing operations, rather than locking the entire file. This partial locking approach preserves data integrity for regions being healed while allowing productive access to other regions, thereby maintaining higher overall file system productivity.
Data Source
AI summary
A server processing device acquires a lock on a region of a file which has stale data. The server processing device self-heals the stale data in the locked region using current data and receives a write request to modify another region of the file during the self-healing of the locked region of the file. The server processing device provides write access to the requested region during the self-healing of the locked region of the file.