Live Migration Throttling for VM Downtime Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current live migration techniques for virtual machines often result in excessively long convergence periods or performance degradation, especially when migrating a virtual machine in use, due to inefficient data copying and throttling methods.
Innovation Solution
Implementing a live migration tool that sets adjustable throttling and downtime limits to manage the migration process, allowing for controlled data transfer and minimizing downtime and performance impact on the virtual machine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If incremental data copying is performed during live migration, then the virtual machine can remain in use during migration, but the convergence period becomes excessively long
Solution Approach 1:
The system performs preliminary actions by capturing memory access information and identifying dirty pages before the final migration cutoff. This advance preparation allows the migration to converge faster by having the dirty page set ready for transfer, reducing the convergence period while maintaining availability.
Solution Approach 2:
The migration process is segmented into distinct phases: initial memory copying, dirty page tracking during runtime, and final synchronized transfer. This segmentation allows the system to maintain the virtual machine during migration while systematically transferring data in manageable portions, reducing overall convergence time.
2Loss of time
If throttling is applied to reduce data transfer volume, then convergence time decreases, but virtual machine performance degrades
Solution Approach 1:
The system implements feedback by continuously monitoring memory access patterns and dirty page generation rates. Based on this feedback, the migration tool dynamically adjusts the migration timing and throttling level, applying increased throttling only when necessary to meet the downtime limit while minimizing impact on virtual machine performance.
Solution Approach 2:
The throttling level is made dynamic rather than static. The system adjusts throttling in real-time based on the actual rate of dirty page generation and the remaining migration time, applying stronger throttling only when needed to ensure convergence within the downtime limit, thereby minimizing performance degradation.
3Loss of time
If aggressive throttling is applied to meet downtime limits, then migration convergence improves, but the throttling amount exceeds acceptable limits
Solution Approach 1:
The migration tool continuously monitors the actual downtime and throttling levels, using this feedback to adjust the migration strategy. If aggressive throttling would exceed the throttling limit, the system modifies the approach by optimizing dirty page identification or adjusting the migration timing, thereby meeting downtime limits without exceeding acceptable throttling amounts.
Data Source
AI summary
A system and method include initiating a live migration of a virtual machine from a first host machine to a second host machine. The system and method include copying an initial copy of data stored in a memory of the first host machine to a memory of the second host machine, determining a set of modified data stored in the memory of the first host machine, determining that transferring the set of modified data stored in the memory of the first host machine exceeds a downtime limit, determining that an increased throttling of the virtual machine executing on the first host machine does not exceed a throttling limit, applying the increased throttling to the virtual machine executing on the first host machine, and copying the set of modified data while the increased throttling is applied to the virtual machine executing on the first host machine.