Distributed FCoE Switching for Scalable Data Center Networks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing Fibre Channel over Ethernet (FCoE) technology faces limitations in scalability due to the use of low-cost large scale integrated circuits, which restricts the storage capacity and network performance as the number of end nodes increases, particularly in data centers with many servers and virtual servers, as the data plane has not been adequately considered for decentralization.

Innovation Solution

The proposed solution involves a network architecture with multiple switches, including Spine and Leaf switches, that use relay information to efficiently route packets between end nodes by sharing relay information and decentralizing the data plane, allowing for improved scalability and load distribution even with low-cost hardware, by modifying packet headers and using FCoE relay tables to manage packet transmission.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If the number of end nodes is increased to improve network scale, then the network capacity is improved, but the storage capacity of low-cost LSI is exceeded and system scalability is limited

Engineering Contradiction:
Improvenumber of end nodesVSAvoidsystem scalability
Core Design Contradiction:
Quantity of substanceVSAdaptability or versatility

Solution Approach 1:

The patent segments the centralized FCoE forwarder functionality into distributed network switches. Each switch maintains a portion of the relay information table and independently performs FCoE frame forwarding. This segmentation allows the system to handle a larger number of end nodes by distributing the storage burden across multiple switches, thereby overcoming the storage capacity limitation of individual low-cost LSI components while maintaining system scalability.

Inventive Principle:
Principle #1Segmentation

2Ease of manufacture

If low-cost LSI is used to reduce device cost, then device cost is reduced, but storage capacity is limited and cannot accommodate increased end nodes

Engineering Contradiction:
Improvedevice costVSAvoidstorage capacity
Core Design Contradiction:
Ease of manufactureVSQuantity of substance

Solution Approach 1:

The patent merges the storage resources of multiple low-cost LSI components across different network switches into a distributed storage system. By combining the relay information tables of multiple switches, the system achieves a total storage capacity that exceeds what any single low-cost LSI could provide, while maintaining the cost benefits of using inexpensive components throughout the network infrastructure.

Inventive Principle:
Principle #5Merging (Combining)

3Ease of operation

If centralized FCoE forwarder is used to simplify management, then network management is simplified, but the data plane cannot handle increased end nodes with limited storage

Engineering Contradiction:
Improvenetwork managementVSAvoidstorage capacity
Core Design Contradiction:
Ease of operationVSQuantity of substance

Solution Approach 1:

The patent segments the centralized FCoE forwarder into multiple distributed forwarding entities across different network switches. Each switch maintains local relay information and performs independent forwarding decisions. This segmentation distributes the storage burden while maintaining simplified management through standardized protocols and uniform switch configurations, allowing the system to scale to accommodate increased end nodes without compromising management simplicity.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS10404586B2Information processing system and information processing method
Publication Date: 2019.09.03 FUJITSU LTD
  • US10404586B2 patent drawing
  • US10404586B2 patent drawing
  • US10404586B2 patent drawing

AI summary

An information processing system includes a plurality of nodes; a plurality of first switches respectively including information on a node to which the first switch is coupled; and a plurality of second switches respectively including information on an aggregation which the second switch is in charge of, wherein a first node transmits a first packet including destination information indicating that a destination is a second node, to a first switch coupled to the first node, and the first switch coupled to the first node transmits the first packet to a second switch that is in charge of an aggregation to which the second node belongs, when the second node is not coupled to the first switch, and the second switch transmits the first packet to a first switch coupled to the second node, and the first switch coupled to the second node transmits the first packet to the second node.