Storage Control Grid Metadata Management for Fault Tolerance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Fault-tolerant data storage systems face challenges in ensuring continuous data availability and integrity, particularly in scenarios involving multiple hardware failures, where existing RAID schemes and grid storage architectures may not adequately protect against component failures and service disruptions.
Innovation Solution
A method and system for a storage control grid with multiple data servers, where each server has primary or secondary responsibility for handling requests, with continuous metadata saving and updating across disk units, using cache memory and non-volatile memory to ensure data integrity and redundancy, and employing a predefined criterion for metadata protection and retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RAID schemes are used to protect data from internal component failures, then data protection is improved, but the system cannot protect against damage or failure of other components within the subsystem
Solution Approach 1:
The system segments protection responsibilities by introducing multiple types of metadata (configuration metadata, data portion metadata, parity portion metadata) that are stored separately and managed independently, allowing different protection strategies for different components
Solution Approach 2:
The patent introduces metadata as an intermediary layer between the physical storage components and the data protection mechanism. This metadata layer tracks the state of data portions, parity portions, and disk drives, enabling coordinated recovery across multiple failure scenarios
2Reliability
If fault tolerant storage systems are implemented with multiple copies of data across nodes, then data availability is improved, but system complexity increases
Solution Approach 1:
The system dynamically assigns primary and secondary responsibility for handling requests to different servers based on operational status. When a server fails, the secondary server automatically takes over, providing adaptive fault tolerance without requiring complex manual reconfiguration
Solution Approach 2:
The patent introduces a temporal dimension to fault tolerance by continuously updating metadata to reflect current system state. This allows the system to adapt to changing conditions and maintain data availability through time-based recovery strategies
3Reliability
If continuous metadata saving is implemented across disk units, then data integrity is improved, but write performance may deteriorate
Solution Approach 1:
The system performs preliminary actions by pre-calculating and storing metadata about data portions and parity portions before actual write operations complete. This allows the system to have recovery information ready in advance, reducing the impact on write performance
Solution Approach 2:
The patent creates copies of metadata information across multiple disk units and servers. By maintaining redundant metadata copies, the system ensures data integrity without requiring synchronous updates to all copies, thus improving write performance
Data Source
AI summary
There is provided a storage system comprising a plurality of disk units adapted to store data at respective ranges of logical block addresses (LBAs), said addresses constituting an entire address space, and a storage control grid operatively connected to the plurality of disk units and comprising a plurality of data servers, each server comprising operatively coupled cache memory and non-volatile memory. The method of operating the storage system comprises: a) configuring a first server among said plurality of data servers to have a primary responsibility for handling requests directed to a certain range of LBAs; b) continuously obtaining by the first server, information indicative of configuration and/or changes thereof related to said certain data range, thus giving rise to configuration metadata; c) saving said configuration metadata and/or derivatives thereof at one or more disk units among said plurality of disk units in accordance with a predefined criterion; d) continuously saving in cache memory of the first server said configuration metadata obtained between said savings at disk units, thus giving rise to recent configuration changes metadata; e) destaging the recent configuration changes metadata to non-volatile memory of the first server if the first server fails.