Unified Data Path for Active-Active Storage Pools
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional active-active data storage systems are limited to fixed-sized LUNs and do not support unified data paths for both block-based and file-based objects, preventing customers from benefiting from pooled LUNs and requiring complex management.
Innovation Solution
Implementing a clustered storage pool and file system across multiple storage processors to provide active-active access to both block-based and file-based objects, ensuring consistency and allowing parallel processing of IO requests, thereby enabling access to pooled data objects in a unified namespace.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional active-active arrangements are used, then multiple storage processors can access LUNs simultaneously, but the system is limited to fixed-sized LUNs and does not support pooled LUNs
Solution Approach 1:
The patent implements a unified data path architecture where storage processors can simultaneously handle both block-based objects (LUNs) and file-based objects through a common file system interface. This universal approach allows the same storage processor to service different types of storage requests without requiring separate active-active pairs for different object types, thereby supporting pooled LUNs with dynamic sizing while avoiding increased system complexity.
2Productivity
If separate active-active arrangements are implemented for block-based and file-based objects, then both object types can be accessed simultaneously, but the system complexity increases and management becomes more difficult
Solution Approach 1:
The patent merges the data paths for block-based and file-based objects into a unified architecture. Storage processors participate in a single active-active arrangement that handles both object types through a common clustered file system and storage pool infrastructure. This consolidation maintains parallel processing capability while significantly reducing system complexity compared to implementing separate active-active pairs for each object type.
Solution Approach 2:
Each storage processor is designed with universal capability to service both block-based LUN requests and file-based requests through the unified data path. The same storage processor can dynamically handle different object types based on incoming requests, eliminating the need for dedicated processors for specific object types and thereby reducing overall system complexity.
3Reliability
If storage processors are differentiated as active or passive, then data consistency can be maintained, but management complexity increases and ease of operation decreases
Solution Approach 1:
The patent implements homogeneity among all storage processors by allowing each processor to operate in the same manner with identical capabilities. All storage processors participate equally in the active-active arrangement and can independently service requests for both block-based and file-based objects. This homogeneous design maintains data consistency through coordinated access to the shared storage pool while greatly simplifying management, as administrators no longer need to track or configure different roles for different processors.
Data Source
AI summary
A technique for processing IO requests in a data storage system provides active-active access to pooled data objects from multiple storage processors. Data objects, which may include both block-based objects and file-based objects, are realized in the form of respective files stored in a set of clustered internal file systems of the data storage system. For providing active-active access to a data object, each of multiple storage processors operates such a clustered file system, which coordinates with a clustered file system on each of the other storage processors to present a consistent file system image having a single namespace across all such storage processors. On each storage processor, the clustered file system is built upon a clustered storage pool, which maintains consistency with each of the clustered storage pools running on each of the other storage processors, to present a consistent image of storage allocation across all storage processors.