IOMMU Cache QoS Policy Remapping for VM Priority

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In virtualized datacenters, virtual machines sharing peripheral devices often experience undesirable latency and throughput due to shared resource access, particularly when one VM runs a critical workload.

Innovation Solution

A software-driven approach that uses an IOMMU to set a QoS policy by sending a priority descriptor, reserving resources for critical workloads, thereby prioritizing access to peripheral devices and reducing latency and increasing throughput.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple virtual machines share a peripheral device, then resource utilization is improved, but latency and throughput deteriorate

Engineering Contradiction:

Improveresource utilizationVSAvoidlatency

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent changes the QoS policy parameters of the IOMMU remapping hardware cache based on VM priority. When a VM is identified as having critical workload, the system modifies cache allocation parameters to favor that VM, ensuring lower latency while maintaining high resource utilization through dynamic parameter adjustment rather than static allocation

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system dynamically adjusts the IOMMU cache QoS policy based on real-time VM workload conditions. The virtualization software monitors VM performance and actively modifies hardware cache behavior, transitioning from static resource sharing to dynamic resource allocation that adapts to changing workload demands, thereby reducing latency for critical VMs while maintaining overall resource utilization

Inventive Principle:

Principle #15Dynamics

2Productivity

If multiple virtual machines share a peripheral device, then resource utilization is improved, but throughput deteriorates

Engineering Contradiction:

Improveresource utilizationVSAvoidthroughput

Core Design Contradiction:

ProductivityVSSpeed

Solution Approach 1:

The patent modifies the IOMMU remapping hardware cache QoS policy parameters based on VM priority levels. For VMs running critical workloads, the system adjusts cache hit ratio targets and allocation parameters to ensure higher throughput, while maintaining efficient resource utilization across all VMs through software-controlled parameter optimization

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system implements dynamic QoS policy adjustment in the IOMMU hardware cache based on real-time VM throughput requirements. The virtualization software monitors throughput performance and actively reconfigures hardware cache behavior, enabling the system to maintain high resource utilization while providing guaranteed throughput for critical VMs through continuous adaptation

Inventive Principle:

Principle #15Dynamics

3Loss of time

If the IOMMU remapping hardware cache QoS policy is dynamically adjusted based on VM priority, then latency and throughput are improved, but device complexity increases

Engineering Contradiction:

ImprovelatencyVSAvoidIOMMU configuration complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The patent introduces virtualization software as an intermediary layer between the VMs and the IOMMU hardware. This software intermediary manages the complexity of QoS policy configuration by automatically monitoring VM workloads and translating priority requirements into appropriate IOMMU cache parameters, thereby reducing latency without requiring direct complex configuration of the hardware itself

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements self-service QoS management where the virtualization software automatically monitors VM performance metrics, identifies critical workloads, and adjusts IOMMU cache policies without manual intervention. This automation reduces the operational complexity of managing dynamic QoS settings while achieving the desired latency improvements for priority VMs

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20220334991A1Software-driven remapping hardware cache quality-of-service policy based on virtual machine priority

Publication Date: 2022.10.20 INTEL CORP

AI summary

Systems, methods, and devices for software-driven resource reservation of an input/output memory management unit (IOMMU) are provided. A system may include a peripheral device and a processing device. The peripheral device may be accessible to a virtual machine running on the processing device via direct memory access (DMA) that is translated by an IOMMU). The processing device may run the virtual machine and a virtual machine manager. The processing device also includes the IOMMU, which is configurable to reserve a subset of resources of the IOMMU to the virtual machine based on a descriptor provided by the virtual machine manager.