Clustered Array Controller for SAN Global Redundancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional storage arrays with two controllers for redundancy and load sharing are limited in the number of drives they can access and only allow one controller for redundancy and I/O operations, restricting scalability and availability in a Storage Area Network (SAN).
Innovation Solution
Implementing a clustered array controller with multiple array modules, each with primary and secondary controllers, allowing for N-way active and passive controller configurations, direct attachment of drive enclosures to a storage network switch, and using fiber-channel interposers to present SAS drives as FC drives, enabling global redundancy and load sharing across the SAN.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional storage arrays are implemented with two controllers for redundancy and load sharing, then controller redundancy is provided, but the number of drives that can be accessed is limited and scalability is restricted
Solution Approach 1:
The storage system is segmented into multiple independent array modules, each with its own controllers and drives. This segmentation allows each module to operate independently while contributing to the overall system capacity, thereby breaking the limitation of conventional two-controller architectures and enabling scalable growth by adding more array modules to the fabric.
Solution Approach 2:
The array modules are designed with universal interfaces and protocols (Fibre Channel, SAS, SATA) that allow them to function in multiple roles within the storage network. Each module can serve as both a standalone storage unit and part of a distributed storage system, providing multi-functionality that enhances both reliability through redundancy and scalability through flexible deployment options.
2Device complexity
If only one controller is used for redundancy and I/O operations, then device complexity is reduced, but availability and load sharing capability are limited
Solution Approach 1:
Multiple controllers within array modules are merged into a unified logical unit that presents a single interface to hosts. The controllers work in concert through the fabric, sharing I/O operations and providing redundancy. This merging approach maintains simplicity from the host perspective while enabling advanced features like load sharing and failover internally.
Solution Approach 2:
The fabric (storage network infrastructure) acts as an intermediary between multiple controllers and hosts, enabling controllers to communicate and coordinate without direct peer-to-peer connections. This intermediary approach allows multiple controllers to provide redundancy and load sharing while maintaining manageable complexity through standardized fabric protocols and routing.
3Ease of operation
If drive enclosures are directly connected to controllers, then connection simplicity is maintained, but the number of accessible drives is limited
Solution Approach 1:
The architecture transitions from a two-dimensional direct connection model (host-controller-drive) to a three-dimensional fabric-based model where array modules can be distributed throughout the storage network. This dimensional change allows drives to be accessed not only through direct controller connections but also through the fabric, dramatically increasing the number of accessible drives while maintaining connection simplicity through standardized fabric interfaces.
Data Source
AI summary
An apparatus comprising a first of a plurality of array modules and a second of a plurality of array modules. The first of the plurality of array modules may be configured to connect a first one or more of a plurality of host devices to a first one or more of a plurality of storage arrays. The first array module may comprise a primary controller and a secondary controller. The second of the plurality of array modules may be configured to connect a second one or more of the plurality of host devices to a second one or more of the plurality of storage arrays. The second array module may comprise a primary controller and a secondary controller. Control of access to the storage arrays may be transferred to the secondary controller of the first array module during a failure of the primary controller of the first array module.