Virtual Machine Host Compute Layer Updates Without Guest OS Reboots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hypervisor-based virtualization technologies face challenges in updating the host compute layer (HCL) of guest virtual machines (VMs) due to disruptive reboot requirements, which are time-consuming and difficult to manage across multiple VM hosts, leading to inconsistencies and delays in bug fixes.
Innovation Solution
A method to update the HCL of guest VMs by pausing the guest OS, persisting its operating state, copying updated HCL code, and resuming the VM without shutting down the guest OS, allowing for transparent and non-disruptive updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the HCL of guest VMs is updated using traditional reboot methods, then the updated HCL code is deployed, but the guest OS must be shut down and the update process is time-consuming and disruptive
Solution Approach 1:
The patent applies preliminary action by persisting the guest OS operating state before stopping the virtual processor system, and by preparing the updated HCL code in advance. This allows the update to be applied without requiring a full reboot, as the system has already captured the necessary state information and prepared the updated code, significantly reducing update deployment time while maintaining consistency across multiple VM hosts.
2Reliability
If the virtual processor system is stopped to copy updated HCL code, then the HCL can be updated, but the guest OS execution is interrupted
Solution Approach 1:
The patent segments the update process into distinct phases: persisting the operating state, stopping the virtual processor system, copying the updated HCL code, and resuming execution. This segmentation allows the system to maintain guest OS availability by quickly completing each phase, particularly by using the persisted state to enable rapid resumption, thereby updating HCL code without significant interruption to guest OS productivity.
Solution Approach 2:
The patent uses copying by creating a persistent copy of the guest OS operating state before updating the HCL code. This copied state is stored and later used to restore the guest OS after the HCL update, allowing the virtual processor system to be stopped and updated without permanent loss of guest OS execution context, thus maintaining high availability during updates.
3Stability of the object's composition
If multiple VM hosts are updated simultaneously, then consistent HCL versions are achieved, but management complexity increases
Solution Approach 1:
The patent applies self-service by enabling each VM host to autonomously persist its guest OS state and update its HCL code using the same standardized process. Each host independently manages its own update cycle by persisting state, stopping the virtual processor, copying updated code, and resuming execution, which simplifies centralized management while ensuring consistent HCL versions across multiple hosts through uniform update procedures.
Data Source
AI summary
A method is disclosed for updating a host compute layer (HCL) in a virtual machine (VM) host computer system. The method involves determining the availability of an update for the HCL of a guest VM operating on the VM host system. A message is sent to the HCL to persist the HCL's operating state, including pausing the execution of the guest operating system (OS) and then persisting the operating state. Subsequently, the virtual processor (VP) system associated with the guest VM is stopped, a register is set to a power-on or reset value, and updated HCL code is copied into the memory space of the HCL. The VP system is then resumed, booting the updated HCL, which restores the operating state and resumes the execution of the guest OS.