Replicating Metadata in Scale-Out Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scale-out data storage systems, replicating data items across multiple nodes is inefficient and unnecessary, especially when data redundancy is provided by storage area networks like RAID systems.
Innovation Solution
The system generates metadata items for each data item stored and replicates these metadata items across multiple storage nodes, while the actual data items are stored only on a storage area network, optimizing storage efficiency and maintaining data redundancy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data items are replicated across multiple storage nodes, then data redundancy and reliability are improved, but storage efficiency deteriorates due to unnecessary duplication
Solution Approach 1:
The patent extracts only the metadata from the data item and replicates it across multiple storage nodes, while the actual data item is stored only once in the storage area network. This separation of metadata replication from data replication resolves the contradiction by providing redundancy where needed (in metadata for access efficiency) while avoiding unnecessary data duplication.
Solution Approach 2:
The patent creates a copy of the metadata item and stores it on a second storage node, while the original data item remains stored only in the storage area network. This selective copying of metadata rather than full data provides the needed redundancy without the storage overhead of replicating entire data items.
2Speed
If metadata items are replicated across multiple storage nodes, then data retrieval speed is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the essential metadata from data items and replicates it across storage nodes, simplifying the replication process compared to full data replication while still enabling fast data retrieval through metadata availability at multiple nodes.
Solution Approach 2:
The patent changes the parameter being replicated from full data items to condensed metadata items, reducing the volume and complexity of replicated information while maintaining the ability to quickly locate and retrieve original data from the storage area network.
Data Source
AI summary
Computer-implemented methods for storing data with replicated metadata in a scale-out data storage system are provided. Aspects include receiving, by a first data storage node of the scale-out data storage system, a data item to be stored and storing, by the first data storage node, the data item on a storage area network in communication with the scale-out data storage system. Aspects also include generating, by the first data storage node, a metadata item corresponding to the data item and replicating the metadata item on a second data storage node of the scale-out data storage system.