Logical Mirroring to Multi-Tier Target Node
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mirroring solutions are inefficient for memory-aware file systems and hybrid storage systems, as they require a software layer to simulate a block interface, are designed for HDD and SSD speeds, and assume symmetry between storage types, failing to provide effective redundancy in modern clustered systems with mixed storage tiers.
Innovation Solution
A system and method for logical mirroring that utilizes a multi-tiered target node with both fast and slower access tiers, allowing efficient mirroring of data and metadata from a memory-based initiator node, reducing the size of the expensive fast access tier and providing cost-effective redundancy by selectively copying data and metadata based on access probability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current mirroring solutions (DRBD, RAIN) are used for memory-aware file systems, then data redundancy is achieved, but efficiency deteriorates due to software layer simulation of block interface and mismatch with memory-speed requirements
Solution Approach 1:
The patent implements logical mirroring by creating a copy of the initiator node's data and metadata structure at the target node, maintaining identical file system semantics. This allows memory-speed mirroring without block interface simulation, as the copy operates natively at the file system level rather than requiring translation layers.
Solution Approach 2:
The patent changes the operational parameters of mirroring from block-based to file-system-based operations, allowing memory-aware file systems to mirror data at memory speeds. By operating at the file system level rather than block level, the system eliminates the performance penalty of interface simulation while maintaining data redundancy.
2Device complexity
If symmetric storage configuration is used (same storage type on both nodes), then simplicity is maintained, but adaptability deteriorates when mixing fast and slow storage tiers
Solution Approach 1:
The patent embraces asymmetric storage configurations where the target node can have different storage tier compositions than the initiator node. The logical mirroring mechanism abstracts away the underlying storage differences, allowing fast storage on the initiator to mirror to a target with mixed fast/slow tiers without requiring symmetric configuration. This enables adaptability to heterogeneous storage environments while maintaining file system semantics.
3Reliability
If full fast access tier is allocated on target node, then data availability is ensured, but cost increases due to expensive fast storage resources
Solution Approach 1:
The patent applies local quality by differentiating which data resides on fast versus slow storage tiers at the target node based on access patterns and importance. Critical data that requires fast access is placed on fast storage, while less time-sensitive data is placed on slower, cheaper storage. This selective placement maintains data availability for critical operations while reducing the total capacity of expensive fast storage required.
Solution Approach 2:
The patent implements partial mirroring where not all data needs to be replicated to the same extent on the target node. By using logical mirroring with file system semantics, the system can prioritize mirroring critical data structures and metadata while using slower storage for less critical data, thereby reducing the amount of fast storage capacity needed while maintaining sufficient data availability.
Data Source
AI summary
Logical mirroring of an initiator server running a memory aware file system to a multi-tiered target server by receiving at a first tier of the target server data that was modified at the initiator server and retaining at the first tier of the target server a first subset of the data and moving to a second tier of the target server a second subset of the data to efficiently utilize the multi-tiered target server.