Distributed Fabric Protocol Switching Network Architecture

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional network switching architectures require independent configuration and management of each switch, leading to complexity in deployment, configuration, and management of network fabric, especially in large networks where switches operate independently at the management, control, and data planes.

Innovation Solution

A distributed fabric protocol (DFP) switching network architecture that centralizes management, control, and data planes by implementing a master switch tier and a follower switch tier, where master switches manage data traffic and routing, and follower switches operate in a pass-through mode, allowing unified virtualized switch management and load balancing across multiple master switches.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If each switch operates independently at management, control, and data planes, then each switch can be configured and managed individually, but the overall network deployment, configuration, and management becomes complex

Engineering Contradiction:
ImproveIndividual switch configuration and managementVSAvoidNetwork fabric deployment and management complexity
Core Design Contradiction:
Ease of operationVSDevice complexity

Solution Approach 1:

The patent merges multiple independent switches into a unified virtualized switch entity. The fabric broker consolidates management, control, and data planes across multiple physical switches, creating a single logical switch that simplifies deployment and management while maintaining individual switch operational capabilities through virtual port representations.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The virtualized switch entity provides universal management capabilities across heterogeneous switch infrastructure. The fabric broker implements multi-functional roles including management plane consolidation, control plane coordination, and data plane optimization, allowing a single management interface to control multiple physical switches with different capabilities.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Productivity

If multiple master switches are implemented for load balancing, then traffic distribution improves, but system complexity increases

Engineering Contradiction:
ImproveTraffic handling capacity and load balancingVSAvoidMulti-master switch coordination complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The fabric broker acts as an intermediary that coordinates multiple master switches. It manages the virtual port mappings and traffic distribution across master switches, abstracting the complexity of multi-master coordination from individual switches and providing unified load balancing control through centralized fabric broker logic.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent introduces a hierarchical dimension with fabric brokers at the coordination layer and switches at the execution layer. This dimensional separation allows load balancing to be managed at the fabric broker level through virtual port mappings, while individual switches focus on data plane forwarding, reducing operational complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Ease of operation

If follower switches operate in pass-through mode, then management is simplified, but control capability over data traffic is reduced

Engineering Contradiction:
ImproveSwitch management simplicityVSAvoidData traffic control capability
Core Design Contradiction:
Ease of operationVSAdaptability or versatility

Solution Approach 1:

The patent combines the control capabilities of follower switches with the fabric broker. While follower switches operate in simplified pass-through mode for easy management, the fabric broker consolidates control plane functions, providing centralized adaptability and versatility for data traffic control across the entire fabric without requiring complex local control at each follower switch.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS8797843B2High availability distributed fabric protocol (DFP) switching network architecture
Publication Date: 2014.08.05 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US8797843B2 patent drawing
  • US8797843B2 patent drawing
  • US8797843B2 patent drawing

AI summary

In a switching network, each of a plurality of lower tier entities is coupled to each of multiple master switches at an upper tier by a respective one of multiple links. At each of the multiple master switches, a plurality of virtual ports each corresponding to a respective one of a plurality of remote physical interfaces (RPIs) at the lower tier are implemented on each of a plurality of ports. Each of the plurality of lower tier entities implements a respective egress port mapping indicating which of its plurality of RPIs transmits egress data traffic through each of its multiple links to the multiple master switches. In response to failure of one of the multiple links coupling a particular lower tier entity to a particular master switch, the particular lower tier entity updates its egress port mapping to redirect egress data traffic to another of the multiple master switches without packet dropping.