VM Migration Network Notification Loss Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Network notification loss during virtual machine migration leads to communication downtime, which is difficult to detect and debug due to its similarity with other downtime sources and inefficiencies in updating network forwarding tables.

Innovation Solution

Implementing monitoring nodes within the network to detect incoming packets with the virtual machine's address as the source address field, and a reporting node to determine if the network has updated forwarding tables, with actions taken to resolve lost notification packets and reduce downtime.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If network components update forwarding tables during virtual machine migration, then network communication is maintained, but notification packets may be lost causing communication downtime

Engineering Contradiction:

Improvenetwork communication reliabilityVSAvoidnotification packet loss

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The source host monitors for incoming packets with the virtual machine's address as the source address field before the migration is fully complete. This preliminary monitoring action allows the system to detect notification packet loss and take corrective actions to ensure network forwarding tables are properly updated, preventing communication downtime.

Inventive Principle:

Principle #10Preliminary action

2Difficulty of detecting and measuring

If monitoring nodes detect incoming packets to verify network updates, then notification loss is detected, but device complexity increases

Engineering Contradiction:

Improvenotification loss detection capabilityVSAvoidnetwork monitoring infrastructure

Core Design Contradiction:

Difficulty of detecting and measuringVSDevice complexity

Solution Approach 1:

The source host performs self-monitoring by inspecting incoming packets for the virtual machine's address as the source address field. This self-service approach allows the source host to detect notification packet loss without requiring separate monitoring infrastructure, thereby reducing device complexity while maintaining detection capability.

Inventive Principle:

Principle #25Self-service

3Loss of time

If the source host monitors incoming packets after migration, then communication downtime is reduced, but processing overhead increases

Engineering Contradiction:

Improvecommunication downtimeVSAvoidhost processing resources

Core Design Contradiction:

Loss of timeVSUse of energy by moving object

Solution Approach 1:

The source host monitors for incoming packets with the virtual machine's address as the source address field for a limited threshold period after migration. This partial monitoring action is sufficient to detect notification packet loss and enable corrective actions, while avoiding excessive continuous monitoring that would consume unnecessary processing resources.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10838752B2Network notification loss detection for virtual machine migration

Publication Date: 2020.11.17 RED HAT ISRAEL

AI summary

A method includes receiving, by a processing device of a monitoring node, an indication over a network that a virtual machine successfully migrated from a first host to a second host. The indication includes a virtual machine address of the virtual machine executing on the second host. The method also includes, responsive to the indication that the virtual machine successfully migrated from the first host to the second host, starting to monitor incoming packets of the monitoring node for an incoming packet that includes a source address field having the virtual machine address, and, upon determining, after a threshold period of time, that none of the incoming packets include the source address field having the virtual machine address, notifying a reporting node that the incoming packet was not received to facilitate performance of an action to reduce downtime of communication with the virtual machine over the network.