TOR Switch Automation for Data Center Network Fabric Provisioning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data centers face challenges in rapidly deploying IT infrastructure and accurately forecasting bandwidth and IT capacity requirements, leading to prolonged provisioning times and over-provisioning.

Innovation Solution

A method involving a switch configuration manager that automates the configuration of top-of-rack (TOR) switches by creating virtual networks and associating them with TOR switches in response to customer requests, thereby streamlining network setup and access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If manual configuration methods are used for TOR switches, then configuration accuracy can be maintained, but provisioning time is prolonged

Engineering Contradiction:
Improveprovisioning timeVSAvoidconfiguration automation
Core Design Contradiction:
Loss of timeVSExtent of automation

Solution Approach 1:

The system enables self-service automation where the TOR switches automatically configure themselves by receiving configuration parameters from the network controller and autonomously establishing virtual network connections, eliminating manual configuration steps while maintaining accuracy through standardized protocols

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The network controller pre-configures virtual network parameters and configuration templates in advance, so that when a TOR switch needs to be provisioned, the configuration is already prepared and can be rapidly deployed without ad-hoc manual setup

Inventive Principle:
Principle #10Preliminary action

2Reliability

If over-provisioning is used to ensure capacity availability, then service reliability is improved, but resource utilization efficiency deteriorates

Engineering Contradiction:
Improveservice reliabilityVSAvoidnetwork capacity resources
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The system implements dynamic bandwidth allocation where network capacity is initially provisioned at a baseline level and automatically scaled up or down based on real-time traffic monitoring and demand detection, ensuring service reliability while optimizing resource utilization

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The network controller continuously monitors traffic patterns and bandwidth utilization on TOR switches, using this feedback to dynamically adjust capacity allocation and prevent both over-provisioning and under-provisioning scenarios

Inventive Principle:
Principle #23Feedback

3Adaptability or versatility

If traditional network configuration methods are used, then system stability is maintained, but adaptability to new services deteriorates

Engineering Contradiction:
Improveservice integration capabilityVSAvoidconfiguration complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The network controller implements a universal configuration interface that can provision multiple service types (virtual networks, VLANs, bandwidth allocations) through a single standardized protocol, enabling the system to adapt to new services without increasing operational complexity

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

Data Source

PatentUS20250193131A1Network service integration into a network fabric of a data center
Publication Date: 2025.06.12 CYXTERA DATA CENTERS INC
  • US20250193131A1 patent drawing
  • US20250193131A1 patent drawing
  • US20250193131A1 patent drawing

AI summary

Top-of-rack (TOR) switches are connected to a network fabric of a data center. Each TOR switch corresponds to a rack of the data center, and is configured to provide access to the network fabric for computing devices mounted in the rack. In one method, a client device of a user is used to select various network service options. The service options correspond to services that can be provided to computing equipment of the user that is mounted in various racks of the data center. In response to receiving the selection of one or more service options, the network fabric of the data center is configured to connect the computing equipment to the selected services. In one approach, the network fabric is configured by creating and/or configuring one or more virtual networks to provide the connection to the services.