Virtualization Core Reset via Segmented Processor Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multicore processor systems using virtualization software, a failure in the virtualization software can inadvertently reset all processor cores, including those running independent block server OS, leading to unnecessary resets and performance deterioration.
Innovation Solution
A method is implemented where processor cores are divided into groups with hardware-assisted virtualization enabled or disabled, allowing the first OS to independently manage and reset only the virtualization module's processor cores without affecting the first OS, by using interrupt handlers and LAPIC IDs to control the reboot process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtualization software is used to run multiple OSes on a multicore processor, then resource utilization and versatility are improved, but system reliability deteriorates because a failure in the virtualization software causes all processor cores to reset
Solution Approach 1:
The processor cores are divided into two distinct groups: a first processor core group where hardware-assisted virtualization is disabled and a second processor core group where hardware-assisted virtualization is enabled. This segmentation isolates the virtualization software to specific cores, allowing other cores to continue operating independently even when virtualization software fails, thus resolving the contradiction between versatility and reliability.
2Reliability
If all processor cores are reset when virtualization software fails, then system stability is restored, but productivity deteriorates due to unnecessary resets of independent OSes
Solution Approach 1:
By segmenting processor cores into those dedicated to virtualization and those dedicated to independent OS operation, the system can selectively reset only the affected virtualization cores while maintaining operational continuity of independent OSes, thus resolving the contradiction between reliability restoration and productivity maintenance.
Solution Approach 2:
The virtualization software and its associated processor cores are extracted as a separate, isolated component from the overall system. This extraction allows the independent OSes to be decoupled from the virtualization software's failure mode, enabling selective resetting of only the virtualization components while preserving the productivity of independent OS operations.
3Productivity
If hardware-assisted virtualization is enabled on all processor cores, then virtualization performance is improved, but device complexity increases and independent OS operation becomes difficult
Solution Approach 1:
Hardware-assisted virtualization is selectively enabled only on the second processor core group dedicated to virtualization software, while the first processor core group maintains hardware-assisted virtualization disabled for independent OS operation. This segmented approach optimizes virtualization performance where needed while simplifying the overall system configuration and avoiding the complexity of managing virtualization across all cores.
Data Source
AI summary
A computer, having: a first OS to which a first processor core group is allocated; and a virtualization module to which a second processor core group is allocated, wherein a virtualization module registers interrupt handler processing for resetting the second processor core group, wherein a first OS has: a monitoring module for monitoring the virtualization module; and an interrupt control module for obtaining identifiers of the second processor core group at a time of booting of the virtualization module, and, when the monitoring module determines to reboot the virtualization module, issuing resetting interrupt to the processor cores of the second processor core group that are associated with the kept identifiers, wherein the second processor core group receives the resetting interrupt and executes the interrupt handler processing to reset its own processor cores, wherein the interrupt control module issues startup interrupt to the second processor core group.