Virtual Machine Hibernation via Compressed Memory Page Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtual machine hibernation methods based on Hypervisor result in long hibernation delays and potential failure, especially in cloud environments where resource utilization is non-consecutive.
Innovation Solution
A hibernation method that involves compressing and transferring memory pages of a virtual machine to a storage object, allowing the virtual machine to inherit its previous state upon wakeup, while optimizing resource usage and reducing delays through server rescheduling and parallel data transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hibernation is based on virtual machine monitor Hypervisor, then the virtual machine can be stopped and resources released, but the hibernation delay becomes long
Solution Approach 1:
The system performs preliminary actions by pre-compressing memory pages during normal operation and preparing storage structures before hibernation is actually needed. When hibernation is triggered, the compressed data is already ready for rapid transfer, significantly reducing the actual hibernation delay while ensuring reliability through pre-validated compression and storage mechanisms.
2Reliability
If hibernation is based on operating system persistence capability, then the state can be preserved, but the hibernation may fail in virtual machine scenarios
Solution Approach 1:
The patent introduces an intermediary layer between the operating system and the virtual machine monitor. This intermediary captures memory pages at the virtual machine level, compresses them, and stores them independently of OS-specific hibernation mechanisms. This approach ensures state preservation while achieving broad virtual machine compatibility by bypassing OS limitations.
3Speed
If running data is transferred without compression, then the transfer speed may be faster, but the storage object experiences high read and write pressure
Solution Approach 1:
The system changes the parameter of data density by applying compression algorithms to memory pages before storage. Although compression requires processing time, it significantly reduces the volume of data that needs to be transferred and stored, thereby reducing storage object pressure. The compressed state is maintained during hibernation, and decompression is performed only when resuming, optimizing the balance between transfer speed and storage pressure.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
This application relates to a hibernation method and a related device thereof. The method includes: determining, when running of a to-be-hibernated virtual machine is stopped by a first virtual machine monitor of a first target server in which the to-be-hibernated virtual machine is located, a memory page used for storing running data of the to-be-hibernated virtual machine in memory of the to-be-hibernated virtual machine in the first target server; and transferring the running data to a predetermined storage object after the running data stored in the memory page is compressed by the first virtual machine monitor, so that when the to-be-hibernated virtual machine is woken up, the woken to-be-hibernated virtual machine inherits, by using the running data read from the storage object, a running state that is before the running of the to-be-hibernated virtual machine is stopped. According to this application, problems that hibernation of a virtual machine depends on a capability of an operating system of the virtual machine, resulting in hibernation failure and a long hibernation delay are resolved.