vGPU Scheduling Policy-Aware VM Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to consider vGPU scheduling policies during virtual machine migration, leading to sub-optimal placement and performance degradation in vGPU-enabled systems.
Innovation Solution
The virtual machine scheduler takes into account vGPU scheduling policies by maintaining tables and data structures that map GPU profiles to identifiers and scheduling policies, using vGPU scheduling migration rules to identify the best available policy for destination GPUs during migration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If virtual machines are migrated without considering vGPU scheduling policies, then migration speed is improved, but performance stability deteriorates
Solution Approach 1:
The system performs preliminary classification of destination hosts based on their vGPU scheduling policies before migration. By pre-identifying hosts with matching or compatible scheduling policies, the system ensures that performance stability is maintained without adding significant overhead to the migration process itself.
Solution Approach 2:
The system uses feedback from the scheduling policy classification to guide migration decisions. By continuously monitoring and classifying destination hosts according to their vGPU scheduling policies, the system can make informed migration decisions that maintain performance stability while still achieving efficient migration.
2Reliability
If vGPU scheduling policies are considered during migration, then performance stability is improved, but system complexity increases
Solution Approach 1:
The system changes the parameter used for host selection from generic resource availability to specific vGPU scheduling policy compatibility. By classifying hosts based on their scheduling policy parameters and matching them with source hosts, the system maintains performance stability while using a straightforward classification approach rather than complex optimization algorithms.
3Manufacturing precision
If comprehensive host information is collected for migration decisions, then placement quality is improved, but information processing overhead increases
Solution Approach 1:
The system extracts only the critical vGPU scheduling policy information from comprehensive host data, rather than processing all available host information. By focusing specifically on scheduling policy compatibility, the system achieves sufficient placement quality without the overhead of analyzing every host parameter.
Data Source
AI summary
Disclosed are aspects of virtual graphics processing unit (vGPU) scheduling-aware virtual machine migration. Graphics processing units (GPUs) that are compatible with a current virtual GPU (vGPU) profile for a virtual machine are identified. A scheduling policy matching order for a migration of the virtual machine is determined based on a current vGPU scheduling policy for the virtual machine. A destination GPU is selected based on a vGPU scheduling policy of the destination GPU being identified as a best available vGPU scheduling policy according to the scheduling policy matching order. The virtual machine is migrated to the destination GPU.