SSD Fault Domain Segmentation for Rapid Data Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In RAID-based storage systems, especially those including solid state disks (SSDs), the reconstruction time for faulty SSDs is excessively long, requiring the reconstruction of data across all logical addresses when only a portion is faulty, which is inefficient and time-consuming.
Innovation Solution
Implementing a data recovery method that uses an erasure code algorithm to map specific logical addresses to physical addresses within faulty fault domains, allowing for partial reconstruction and reducing the need to rebuild data across all SSD addresses, thereby increasing reconstruction speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data reconstruction is performed across all logical addresses of a faulty SSD, then data reliability is ensured, but reconstruction time becomes excessively long
Solution Approach 1:
The patent segments the SSD into multiple fault domains (e.g., different dies, channels, or planes) and further divides data into chunk groups distributed across these domains. When a fault occurs, only the affected fault domain needs reconstruction rather than the entire SSD. This segmentation enables parallel reconstruction across multiple domains simultaneously, dramatically reducing total reconstruction time while maintaining data reliability through the erasure code algorithm.
2Reliability
If full SSD reconstruction is performed, then complete data recovery is achieved, but system productivity decreases due to long reconstruction duration
Solution Approach 1:
The patent implements partial action by reconstructing only the necessary portion of data affected by the fault domain rather than performing excessive full SSD reconstruction. The erasure code algorithm enables recovery of lost chunks from affected fault domains using data from other healthy domains, achieving sufficient data recovery without the overhead of complete system reconstruction, thus maintaining productivity.
3Loss of time
If chunk groups are distributed across multiple fault domains, then reconstruction scope is limited to affected domains, but system complexity increases
Solution Approach 1:
The patent introduces a controller as an intermediary that manages the complex mapping relationships between logical addresses, chunk groups, and fault domains. The controller maintains metadata structures that track which chunks belong to which chunk groups and which domains they reside in. This intermediary layer abstracts the complexity from the reconstruction process, enabling efficient domain-specific recovery without requiring complex distributed algorithms across all nodes.
Data Source
AI summary
Embodiments of the present disclosure provide a data recovery method in a storage system. A solid state disk is divided into a plurality of fault domains, and each fault domain is used to provide a physical address for a specific range of logical addresses of an SSD, so that when a fault domain of the solid state disk is faulty, it is unnecessary to reconstruct data in the entire SSD.