Virtual Machine Migration Demand Paging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for migrating virtual machines between physical machines are inefficient due to the need to copy large amounts of data, which can result in wasteful duplication of data transfers and resource consumption, especially when modified data is copied before it is needed on the destination machine.
Innovation Solution
The virtual machine is migrated by stopping its execution on the origin machine, copying state information, and then starting it on the destination machine, with data transferred on demand as needed by the virtual machine, using techniques like demand paging to avoid unnecessary data transfers and duplicate copies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is copied from the first physical machine to the second physical machine before migration, then the virtual machine can be migrated, but the migration time increases and network resources are consumed
Solution Approach 1:
The patent applies preliminary action by copying only the necessary state information and essential data portions from the first physical machine to the second physical machine before migration, while leaving out data that can be retrieved on-demand later. This reduces the preliminary data transfer time while ensuring critical information is available for immediate VM execution.
Solution Approach 2:
The patent extracts and separates data into different categories: state information that must be copied beforehand, and other data that can be retrieved on-demand after migration. This extraction principle allows the system to migrate the virtual machine quickly by taking out only the essential components needed for immediate operation.
2Reliability
If all data is copied before migration, then the virtual machine can execute on the destination machine, but network resources are wasted copying data that may not be needed
Solution Approach 1:
The patent extracts only the essential state information and frequently accessed data portions for initial copying to the destination machine. Non-essential or rarely accessed data is left on the source machine or in storage, to be retrieved on-demand only when actually needed by the virtual machine, thereby avoiding waste of network resources.
Solution Approach 2:
The patent applies partial action by copying only the necessary portion of data (state information and essential data portions) rather than all data. This partial copying approach provides just enough information for the virtual machine to execute, avoiding the excessive action of copying unnecessary data that would consume network resources.
3Reliability
If data is copied before migration, then the destination machine has data available, but modified data may be duplicated if copied again after changes
Solution Approach 1:
The patent implements feedback mechanisms where the destination machine monitors which data portions are actually accessed and needed by the virtual machine. This feedback information is used to determine what additional data should be retrieved on-demand, avoiding duplication of effort by only copying data that has been proven necessary through actual usage patterns.
Data Source
AI summary
To migrate a virtual machine from a first physical machine to a second physical machine over a network, execution of the virtual machine in the first physical machine is stopped, where the first physical machine contains data associated with the virtual machine. State information of the virtual machine is copied from the first physical machine to the second physical machine, and execution of the virtual machine is started on the second physical machine based on the copied state information.