Live Migration Notification for VM Timeout Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

During live migration of virtual machines (VMs), downtime occurs when the VM is moved to new underlying hardware, leading to potential timeout events that can erroneously indicate a connection is inactive, causing network recovery, availability, and performance issues.

Innovation Solution

The system employs proactive and reactive notification methods to inform associated nodes about the status of VMs undergoing live migration, allowing nodes to adjust connection parameters and reduce the likelihood of timeout events.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If timeout values are increased or number of acceptable retries is increased, then the likelihood of false timeout events during live migration is reduced, but the ability to accurately detect inactive connections is delayed

Engineering Contradiction:
Improveconnection status accuracyVSAvoidnetwork recovery time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary action by sending a notification to the client before the actual migration downtime occurs. This allows the client to proactively adjust its timeout parameters in advance, so when the VM becomes temporarily unavailable during migration, the client already has extended timeout values configured, preventing false timeout events while maintaining accurate connection detection.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback by monitoring the migration status of the VM and dynamically adjusting timeout parameters based on real-time information. The notification mechanism provides feedback to the client about the upcoming downtime, enabling the client to modify its behavior accordingly and reduce false timeout events while maintaining network recovery efficiency.

Inventive Principle:
Principle #23Feedback

2Reliability

If notification is sent to all nodes in the network, then the coverage of status information is maximized, but the network traffic overhead increases

Engineering Contradiction:
Improveconnection status notification coverageVSAvoidnetwork bandwidth consumption
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The system applies local quality by selectively sending notifications only to nodes that have active connections to the migrating VM, rather than broadcasting to all network nodes. This is achieved by maintaining a list of associated nodes (clients with active connections) and sending notifications only to those nodes, thereby reducing unnecessary network traffic while ensuring reliable delivery to affected clients.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system uses an intermediary approach by having the migration manager or hypervisor act as a mediator that tracks which nodes are associated with the VM and manages the notification distribution. This intermediary mechanism efficiently identifies and contacts only the relevant nodes without requiring network-wide broadcasting, thus reducing bandwidth consumption while maintaining comprehensive notification coverage.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If the VM is frozen during live migration to enable data transfer, then the data transfer completeness is ensured, but the connection responsiveness is lost

Engineering Contradiction:
Improvedata transfer completenessVSAvoidconnection responsiveness
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system performs preliminary action by notifying clients before the VM freeze occurs during migration. This allows clients to prepare for the temporary unresponsiveness by adjusting their timeout parameters and connection handling logic in advance, so the frozen state during data transfer does not cause false timeout events or connection failures.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system converts the harmful effect of VM freeze (temporary unresponsiveness) into a beneficial situation by using the notification mechanism to inform clients of the upcoming freeze. This allows clients to proactively adjust their behavior during the freeze period, transforming the harmful temporary unresponsiveness into a controlled and expected state that does not affect connection reliability.

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Data Source

PatentEP4550138A1Traffic monitoring and control during live migration
Publication Date: 2025.05.07 MELLANOX TECHNOLOGIES LTD(IL)
  • EP4550138A1 patent drawingFigure 1
  • EP4550138A1 patent drawingFigure 2A~2B
  • EP4550138A1 patent drawingFigure 3

AI summary

Systems and methods are directed toward proactive and/or reactive notification and connection parameter adjustments responsive to live migration events. When an application or program is migrated to new underlying hardware, messages may be provided to associated senders to dynamically and/or temporarily adjust different connection parameters to reduce a likelihood of determining the connection is inactive. A proactive notification may broadcast status information during the live migration to each node in a network and/or to particular nodes associated with the migrating receiver. A reactive notification may provide a message regarding receiver status responsive to receiving incoming traffic.