1:N Redundancy in FCoE Switch Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current high availability (HA) networks using Fiber Channels (FC) and Fiber Channels over Ethernet (FCoE) connections face increasing costs with the number of switches in a cluster, as they require synchronization of connection states across all switches, which is complex and limits scalability.
Innovation Solution
Implementing a 1:N redundancy model that establishes alternate paths through redundant Fiber Channel over Ethernet (FCoE) forwarders without synchronizing connection states across switches, using new A_Ports and VA_Port virtual links to create failure domains within fabrics, reducing the need for extensive synchronization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If connection state synchronization is implemented across all switches in an HA cluster, then network reliability is improved, but device complexity and cost increase linearly with the number of switches
Solution Approach 1:
The patent segments the HA cluster into N independent switches that each maintain their own connection state locally, rather than synchronizing across all switches. Each switch operates autonomously with its own state information, eliminating the need for complex inter-switch synchronization while maintaining reliability through redundant paths.
Solution Approach 2:
The patent introduces FCoE forwarders as intermediary devices that enable failover without requiring direct synchronization between switches. The forwarders act as mediators that can detect failures and redirect traffic through alternate paths, allowing switches to maintain independent state while still achieving high availability.
2Reliability
If redundant components are added to reduce service outage, then network reliability is improved, but cost increases linearly with the number of switches
Solution Approach 1:
The patent makes each switch in the N-switch cluster multi-functional by enabling them to handle multiple roles: primary service provision, standby capability, and alternate path routing. Each switch can serve as both an active service provider and a backup for other switches, eliminating the need for dedicated redundant switches and reducing the total quantity required.
Solution Approach 2:
The patent implements dynamic role assignment where switches can transition between active and standby states based on real-time operational needs. This dynamic flexibility allows the system to optimize resource utilization and reduce the number of physical switches needed compared to static 1:1 redundancy configurations.
3Quantity of substance
If 1:N redundancy model is implemented with alternate paths, then cost is reduced, but connection state synchronization becomes more complex
Solution Approach 1:
The patent enables switches to self-manage their connection states and failure domains autonomously without requiring centralized coordination or synchronization with other switches. Each switch independently monitors its own connections and can independently initiate failover actions, simplifying the overall system architecture despite the 1:N redundancy model.
Data Source
AI summary
Network devices, systems, and methods, including executable instructions and/or logic thereon to achieve fiber channel one for N (1:N) redundancy. A network device includes a processing resource coupled to a memory. The memory includes program instructions executed by the processing resource to group a number of switches in a 1:N cluster and provide each switch with a (virtual) A_Port link to all members of the 1:N cluster. If a failure of a fiber channel over ethernet forwarder (FCF) occurs, the program instructions execute to re-establish or redirect a connection over an alternate path through a redundant FCF without having to synchronize a connection state across all switches in the cluster.