Virtual Machine Context Scheduling with Skew Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In virtual machine environments, strict coscheduling can lead to resource underutilization and performance reduction due to idle physical processors, as some virtual machines cannot run when others require simultaneous scheduling of multiple virtual processors.

Innovation Solution

The solution involves tracking and managing the skew between coscheduled contexts by recording transition times and implementing corrective actions such as costop, costart, and coswap policies to reduce skew, allowing for relaxed coscheduling and efficient resource allocation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If strict coscheduling is implemented to ensure simultaneous execution of multiple virtual processors, then synchronization and correctness are improved, but physical processor utilization deteriorates due to idle processors

Engineering Contradiction:

ImprovesynchronizationVSAvoidprocessor utilization

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system dynamically adjusts coscheduling constraints based on system state. When skew detection indicates sufficient synchronization, the system relaxes coscheduling requirements to allow individual virtual processors to run independently, improving processor utilization while maintaining necessary synchronization when required

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes the scheduling parameter from strict simultaneous execution to flexible execution with skew thresholds. By adjusting the skew parameter and transition criteria, the system optimizes the balance between synchronization requirements and processor utilization efficiency

Inventive Principle:

Principle #35Parameter changes

2Reliability

If strict coscheduling is enforced for SMP virtual machines, then inter-processor operation correctness is improved, but resource fragmentation increases causing idle physical processors

Engineering Contradiction:

ImprovecorrectnessVSAvoidresource fragmentation

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The system applies partial coscheduling by allowing some virtual processors to execute independently when skew thresholds are satisfied, rather than requiring all virtual processors to execute simultaneously. This partial action reduces resource fragmentation while maintaining correctness for critical inter-processor operations

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system implements feedback through skew detection mechanisms that monitor the state of coscheduled contexts. Based on this feedback, the scheduler dynamically adjusts scheduling decisions to prevent resource fragmentation while ensuring correctness when skew exceeds acceptable thresholds

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS8127301B1Scheduling selected contexts in response to detecting skew between coscheduled contexts

Publication Date: 2012.02.28 VMWARE INC

AI summary

Management of contexts that execute on a computer system is described. More specifically, context scheduling in a virtual machine environment is described. A set of coscheduled contexts is monitored. If a skew metric associated with any one of the coscheduled contexts fails to satisfy a condition, then all coscheduled contexts in the set of coscheduled contexts not already descheduled are descheduled. After the contexts are descheduled, a subset of the set of coscheduled contexts is scheduled before the remainder of the set of coscheduled contexts.