Intra-Node Network Boot for Virtual Machines

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing network boot methods for virtual machines are dependent on external infrastructure, such as physical networks and servers, which can lead to management challenges and performance issues during large-scale updates of boot images.

Innovation Solution

An integrated network boot and installation protocol that leverages virtual networks and local server instances, using DHCP and PXE protocols, to enable OS installation on virtual machines without communicating over a physical network, by hosting a DHCP server and network boot server on a controller VM and utilizing OpenFlow rules and VXLANs to route traffic internally.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If network boot images are stored at a central repository for easier updates, then ease of operation is improved, but dependence on external infrastructure increases

Engineering Contradiction:

Improveboot image update managementVSAvoiddependence on external infrastructure

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system segments the boot infrastructure by separating the DHCP server and network boot server into distinct virtual machine instances that can be independently managed and deployed within the existing infrastructure, reducing dependence on external centralized repositories while maintaining ease of update management

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces virtual network infrastructure (virtual switches, VXLANs) as intermediaries between the physical network and boot image delivery, enabling centralized management capabilities while isolating the system from external infrastructure dependencies through virtualization layers

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of operation

If traditional network boot protocols are used over physical networks, then ease of operation is improved, but performance deteriorates due to network traffic

Engineering Contradiction:

Improveboot process simplicityVSAvoidinstallation traffic speed

Core Design Contradiction:

Ease of operationVSSpeed

Solution Approach 1:

The patent extracts the network boot protocol processing (DHCP, PXE) from the physical network domain and relocates it to virtual machine instances operating over virtual network infrastructure, separating the control plane from the data plane to improve installation traffic performance

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system transitions from a single physical network dimension to multiple virtual network dimensions (VXLANs, virtual switches), allowing boot protocol traffic to operate in a virtualized dimension that is logically separated from physical network performance constraints

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS11635970B2Integrated network boot operating system installation leveraging hyperconverged storage

Publication Date: 2023.04.25 NUTANIX INC

AI summary

A computing node is configure to implement an intra-node network boot and installation protocol (protocol) for booting and installing an operating system (OS) on a virtual machine hosted on the computing node without communicating over a physical network. The protocol includes hosting a dynamic host configuration protocol (DHCP) server instance and/or a network boot server instance on a controller virtual machine of the computing node to emulate DHCP protocol and network boot server protocol communications. In some examples, the protocol further utilizes one or more virtual extensible local area networks (LANs)(VXLANs) and a virtual switch hosted at a hypervisor running on the computing node.