Unified Storage Architecture Deduplication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Block-based and file-based data storage systems often operate independently, making it difficult to transfer advancements between the two, leading to inefficiencies such as stranded storage where resources go unused due to type mismatches.
Innovation Solution
A unified data path architecture that expresses both block-based and file-based objects as files within an underlying file system, allowing for common services like replication, deduplication, and storage resource reallocation across types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If block-based and file-based storage systems are co-located in a combined system, then both storage types can be processed in a single system, but the system becomes more difficult to support and maintain
Solution Approach 1:
The patent merges block-based and file-based storage systems into a single unified system where both storage types share common infrastructure. The block-based system is represented as files within the file-based system's file system, allowing both types of IO requests to be processed through a single data path and storage architecture, thereby reducing overall system complexity while maintaining versatility
Solution Approach 2:
The unified file system serves multiple functions by accommodating both block-based objects (represented as files) and file-based objects. This universal approach allows a single system to handle diverse storage workloads, replication operations, deduplication, and resource allocation for both storage types without requiring separate specialized systems
2Reliability
If block-based and file-based storage systems operate independently, then each system can be optimized for its specific type, but advancements in one system cannot be transferred to the other
Solution Approach 1:
By representing block-based objects as files within the file-based system, the patent enables advancements in file-based systems (such as deduplication algorithms, replication mechanisms, or resource management techniques) to be directly applied to block-based storage operations. The unified architecture allows feature transfer between storage types while preserving the optimization benefits of each system
3Quantity of substance
If storage resources are allocated to block-based or file-based objects, then storage capacity is provided, but stranded storage accumulates when storage is freed and cannot be reused
Solution Approach 1:
The unified file system enables universal storage resource allocation where freed storage space can be reallocated to any object type regardless of whether it was previously block-based or file-based. The system tracks storage allocation at the file level, allowing flexible reallocation of storage units to meet diverse storage demands and eliminating stranded storage inefficiencies
Solution Approach 2:
The system implements dynamic storage allocation and reallocation mechanisms that adapt to changing storage demands. When storage is freed from one object type, the system can dynamically reallocate it to another object type through the unified file system, optimizing storage utilization and preventing accumulation of stranded resources
Data Source
AI summary
A technique forms contents of a deduplication container from a standalone container file system stored on the same data storage array on which the deduplication container is stored. Such a standalone container includes a container file which is formed by standalone blocks that are not shared with any other file within the standalone container outside of snapshots of that file. The standalone blocks store host data and which are hierarchically related to each other. Upon receiving a deduplication command, the storage processor carries out a set of deduplication operations which include creating and storing the deduplication container in the data storage array. The deduplication container includes container files which are formed by deduplication blocks storing the host data and which are hierarchically related to each other.