Virtual Machine Undo via Differential Memory Updates

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing virtual machine migration techniques face challenges in alleviating overload at the destination computer, leading to prolonged service quality reduction due to the time-consuming process of transferring virtual machines, especially when the destination computer becomes overloaded.

Innovation Solution

A virtual machine handling system that includes a first computer and a second computer connected through a network, with a transfer unit for transmitting the virtual machine's memory image, a memory retaining unit for retaining the memory image, an update recording unit for recording partial updates, and a return unit for transmitting partial data updates back to the first computer upon an undo request, allowing the first computer to continue executing the virtual machine using the retained memory image and received data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a virtual machine is transferred from the first computer to the second computer through live migration, then the load balance and resource utilization are improved, but the destination computer may become overloaded and service quality deteriorates due to the time-consuming transfer process

Engineering Contradiction:
Improveresource utilizationVSAvoidservice quality
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The source computer retains the memory image of the virtual machine before transfer, preparing it in advance for potential undo operations. This preliminary action enables rapid recovery if the destination computer becomes overloaded, preventing service quality deterioration while maintaining resource utilization benefits.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system allows the virtual machine to be discarded at the destination computer and recovered at the source computer through the undo function. When the destination computer becomes overloaded, the virtual machine can be quickly returned to the source computer using the retained memory image, thus recovering service quality without losing the migration benefits.

Inventive Principle:
Principle #34Discarding and recovering

2Reliability

If the entire memory image is transferred during undo operation, then the virtual machine can be returned to the source computer, but the process takes a long time and consumes excessive network and CPU resources

Engineering Contradiction:
Improveundo capabilityVSAvoidtransfer time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

Instead of transferring the entire memory image during undo operation, the system segments the transfer by only sending the differential data (changes) that occurred since the last snapshot. This segmentation dramatically reduces the amount of data to be transferred, enabling fast undo operations without consuming excessive network and CPU resources.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses differential copying where only the changed portions of the memory image are copied from the destination computer back to the source computer. This copying approach is much more efficient than transferring the complete memory image, as it leverages the snapshot to avoid redundant data transmission.

Inventive Principle:
Principle #26Copying

3Reliability

If the source computer retains the memory image for potential undo operations, then rapid undo is enabled, but the memory resources of the source computer are consumed

Engineering Contradiction:
Improverapid undo capabilityVSAvoidmemory resource consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system extracts only the essential memory image data needed for undo operations, storing it in a compressed or optimized format. By taking out only the necessary portions rather than retaining complete high-resolution copies, the system enables rapid undo capability while minimizing memory resource consumption on the source computer.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The retained memory image is nested within the virtualization infrastructure, allowing efficient storage and retrieval. The memory image data is nested in a way that enables quick access for undo operations without requiring dedicated large-scale storage resources, thus balancing rapid undo capability with memory resource constraints.

Inventive Principle:
Principle #7Nested doll (Nesting)

Data Source

PatentUS9047110B2Virtual machine handling system, virtual machine handling method, computer, and storage medium
Publication Date: 2015.06.02 NEC CORP
  • US9047110B2 patent drawing
  • US9047110B2 patent drawing
  • US9047110B2 patent drawing

AI summary

A memory retaining unit (106) of a virtual machine handling system (10) retains a memory image (IMG) of a first computer (100). An update recording unit (203) records update information (UD) indicating a partial region of a memory image (IMG1) which is updated by the execution of a transferred virtual machine (VM) by a virtual machine execution unit (202) of a second computer (200). When receiving an undo request, a return unit (160) transmits partial data (PD) indicated by the update information (UD) from the second computer (200) to the first computer (100). A virtual machine execution unit (102) of the first computer (100) executes a continuing process of the virtual machine (VM) using the memory image (IMG) retained in the memory retaining unit (106) and the partial data (PD) received from the second computer (200).