FCoE Gateway System Reducing SAN Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Fibre Channel gateway systems for connecting server IHSs to Fibre Channel storage area networks (SANs) are complex and costly due to the use of proprietary converged Ethernet switches and FCoE transit switches, which introduce issues like head of line blocking, nondeterministic paths, and increased complexity.
Innovation Solution
A Fibre Channel gateway system that establishes direct communication between server IHSs and FC switch IHSs via an FC gateway, using an Ethernet switch with a fabric services engine and FCoE/FC converter to handle FCoE initialization protocol data traffic and provide native FC NPIV gateway uplinks, reducing complexity and cost by eliminating the need for proprietary switches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If proprietary converged Ethernet switches and FCoE transit switches are used to connect server IHSs to Fibre Channel SANs, then communication capability is provided, but system complexity and cost increase
Solution Approach 1:
The patent extracts the Fibre Channel switching functionality from the complex proprietary converged Ethernet switch architecture and places it directly in the chassis enclosure as a native FC switch. This removes the need for FCoE transit switches and proprietary converged Ethernet switches, eliminating head of line blocking and nondeterministic paths while reducing overall system complexity.
Solution Approach 2:
The native Fibre Channel switch in the chassis enclosure serves multiple functions: it provides direct FC connectivity to SANs, handles FCoE initialization protocol traffic, and enables NPIV gateway uplinks. This multi-functional approach eliminates the need for separate proprietary switches while maintaining full communication capability.
2Adaptability or versatility
If proprietary converged Ethernet switches with multiple ASICs and FPGAs are used, then Fibre Channel connectivity is achieved, but cost increases
Solution Approach 1:
The patent merges the Fibre Channel switching functionality directly into the chassis enclosure, combining what were previously separate components (converged Ethernet switch, FCoE transit switch, and FC switch) into a single integrated native FC switch. This consolidation eliminates the need for multiple expensive ASICs and FPGAs while maintaining full FC connectivity.
Solution Approach 2:
The patent replaces expensive proprietary converged Ethernet switches with standard, more affordable Fibre Channel switching hardware that can be easily manufactured and deployed. This substitution significantly reduces cost while maintaining the required connectivity capabilities.
3Adaptability or versatility
If FCoE transit switches with FIP snooping bridges are used, then FCoE initialization is supported, but communication efficiency decreases due to head of line blocking and nondeterministic paths
Solution Approach 1:
The patent extracts the FCoE initialization protocol handling from the FCoE transit switch and implements it directly in the native FC switch within the chassis enclosure. This eliminates the multi-hop FCoE path and associated head of line blocking, providing both FCoE initialization support and deterministic, efficient communication paths simultaneously.
Data Source
AI summary
An Ethernet switch includes an Ethernet processing system. Layer two (L2) forwarding tables are provided with the Ethernet processing system, and a memory system is coupled to the Ethernet processing system and includes Ethernet processing system software. A control subsystem establishes communication with the Ethernet processing system and modifies the L2 forwarding tables by removing default Ethernet entries, adding Fibre Channel over Ethernet (FCoE) formatted FC well know addresses (WKAs), and setting the one or more L2 forwarding tables to remain persistent. The control subsystem also instructs the Ethernet processing system to forward FC control frames to the control subsystem, and replaces first link state instructions in the Ethernet processing system software with second link state instructions to change Ethernet behavior to FC behavior. The control subsystem may then process FC information in received FCoE format Ethernet frames.