Redundant Communication Fabric for Network Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional communication fabrics in network storage systems, such as Fibre Channel, are costly and complex, and direct-attach storage technologies like SAS have limitations in providing redundancy and scalability.
Innovation Solution
A redundant communication fabric is implemented using multiple paths between storage controllers and storage devices, with horizontal and vertical expanders configured to provide redundancy beyond the number of ports, allowing data access even if up to m−1 paths fail, and supporting a large number of storage devices across multiple isolated subsystems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Fibre Channel technology is used for communication fabric, then reliability and communication speed are improved, but cost and device complexity increase
Solution Approach 1:
The system segments the communication fabric into multiple independent paths between storage controllers and storage devices. Each path operates independently, allowing the system to maintain communication functionality even when some paths fail. This segmentation provides redundancy without requiring a single complex high-performance fabric, thereby reducing overall system complexity while maintaining reliability.
Solution Approach 2:
The system changes the parameter of path quantity from single-path to multi-path architecture. By configuring multiple communication paths (m paths) between storage controllers and storage devices (with m > n, where n is the number of storage device ports), the system achieves redundancy that tolerates up to m-1 path failures, improving reliability without requiring expensive Fibre Channel infrastructure.
2Device complexity
If direct-attach storage devices are used, then cost and simplicity are improved, but redundancy and scalability are limited
Solution Approach 1:
The communication fabric is designed with universal multi-functionality to serve multiple purposes: it provides basic direct-attach connectivity while simultaneously offering redundant path capabilities and scalability to connect a large number of storage devices across multiple isolated subsystems. This multi-functional fabric uses simple expanders rather than complex Fibre Channel switches, achieving both simplicity and enhanced reliability.
Solution Approach 2:
The system adds the dimension of multiple communication paths between storage controllers and storage devices. Instead of relying on a single direct-attach path, the fabric creates multiple routing dimensions, allowing data to traverse different paths. This dimensional expansion provides redundancy and scalability while maintaining the simplicity of direct-attach storage technology.
3Reliability
If multiple paths are configured for redundancy, then reliability is improved, but fabric complexity increases
Solution Approach 1:
The communication fabric is designed to self-manage the complexity of multiple paths through automated routing and path selection mechanisms. The expander devices and storage controllers work together to automatically establish and maintain redundant paths without requiring manual configuration or intervention. This self-service capability provides path redundancy while minimizing the operational complexity burden on users.
Data Source
AI summary
The techniques introduced here include storage systems including a storage controller configured to access data and a storage subsystem including a storage device having n ports, where n is an integer greater than one, and where the storage device is configured to store the data and to make the data available to the storage controller via each of the n ports. The storage systems also include a communication fabric configured to couple the storage controller to each of the n ports of the storage device via m paths, where m is an integer greater than n, so that the storage system is configured to tolerate failure in up to m−1 paths through the communication fabric, such that the data in the storage device remains accessible to the storage controller even in the presence of failure in up to m−1 paths of the m paths.