Erasure-Coded Object Rebuild Prioritization by Fragment Risk
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large-scale distributed storage systems, maintaining synchronization and accessing databases for low-cost, low-reliability storage devices becomes a bottleneck due to the immense scale of data, exceeding the protective capabilities of RAID 5 and RAID 6, especially when using hierarchical erasure coding techniques.
Innovation Solution
Implementing a distributed storage system that assesses the risk of data object loss by evaluating the health of fragments across the system without relying on a centralized or synchronized database, using rebuild nodes to collect and assess object fragment information, and incorporating additional risk information from S.M.A.R.T. systems to prioritize and proactively rebuild data objects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hierarchical erasure coding techniques are used to protect data on low-cost storage devices, then data protection capability is improved, but database synchronization and access becomes a bottleneck due to immense scale
Solution Approach 1:
The patent divides the storage system into independent storage nodes, each maintaining its own local metadata about data fragments. This segmentation eliminates the need for a centralized database, allowing each node to operate independently and avoid synchronization bottlenecks while maintaining erasure coding protection across the distributed system
Solution Approach 2:
The patent extracts the centralized database component from the storage system architecture. Instead of relying on a central authority to track and synchronize metadata, the system distributes metadata management across individual storage nodes, removing the synchronization bottleneck entirely
2Loss of information
If a centralized database is used to track data objects in a distributed storage system, then data location tracking is improved, but system performance deteriorates due to synchronization requirements
Solution Approach 1:
The patent segments the centralized metadata database into distributed metadata stored locally at each storage node. Each node independently tracks its own data fragments and makes this information available to rebuild nodes, eliminating synchronization overhead while maintaining complete data location tracking capability
Solution Approach 2:
Storage nodes autonomously maintain and provide their own metadata information without requiring centralized coordination. When a rebuild node needs fragment information, storage nodes self-serve by providing their local metadata directly, eliminating the performance penalty of centralized database access
3Quantity of substance
If low-cost storage devices are used to reduce storage costs, then storage cost is improved, but device reliability deteriorates with higher failure rates
Solution Approach 1:
The patent implements proactive risk assessment that identifies data objects at risk of loss before actual failures occur. By using S.M.A.R.T. data and fragment distribution analysis, the system cushions against future failures by triggering preemptive rebuilds, ensuring data protection even with unreliable low-cost devices
Solution Approach 2:
The system performs preliminary risk assessment and triggers rebuild operations before storage devices actually fail. By monitoring device health metrics and fragment availability, the system takes preliminary action to reconstruct data objects while still using reliable devices, preventing data loss from impending failures
Data Source
AI summary
A rebuild node of a storage system can assess risk of the storage system not being able to provide a data object. The rebuild node(s) uses information about data object fragments to determine health of a data object, which relates to the risk assessment. The rebuild node obtains object fragment information from nodes throughout the storage system. With the object fragment information, the rebuild node(s) can assess object risk based, at least in part, on the object fragments indicated as existing by the nodes. To assess object risk, the rebuild node(s) treats absent object fragments (i.e., those for which an indication was not received) as lost. When too many object fragments are lost, an object cannot be rebuilt. The erasure coding technique dictates the threshold number of fragments for rebuilding an object. The risk assessment per object influences rebuild of the objects.