SCP Subsystem Snapshot Engine for Live Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional live migration and high availability techniques for server devices with System Control Processor (SCP) subsystems are inadequate as they cannot access SCP subsystem state information, preventing complete replication of virtual machines during migration or backup operations.
Innovation Solution
A System Control Processor (SCP) subsystem with an SCP snapshot engine that retrieves and generates SCP component state information, allowing for the creation of SCP subsystem snapshots that can be transmitted to another SCP subsystem for seamless operation of a virtual machine on a different system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional live migration techniques are used, then virtual machine migration can be performed, but SCP subsystem state information cannot be accessed or retrieved
Solution Approach 1:
The patent segments the state information collection process into two distinct phases: a first phase that collects initial state information from various components, and a second phase that collects updated state information after changes occur. This segmentation allows the system to systematically retrieve SCP subsystem state information that was previously inaccessible, resolving the contradiction between information completeness and operational feasibility.
Solution Approach 2:
The patent implements preliminary actions by collecting state information from components before the actual migration operation. The system performs state information collection in advance, generating snapshots that can be applied during migration. This preliminary action ensures that all necessary SCP subsystem state information is captured before migration begins, eliminating the information loss problem.
2Loss of information
If multiple snapshot commands are issued, then complete state information can be retrieved, but the process becomes more complex
Solution Approach 1:
The patent employs feedback mechanisms where the system receives feedback about which state information has been collected and which updates are needed. The first snapshot command collects initial state information, and the second snapshot command collects updates based on feedback about changes since the first snapshot. This feedback-driven approach systematically retrieves complete state information while managing process complexity through structured iterations.
Solution Approach 2:
The patent uses periodic action by issuing multiple snapshot commands at different stages: a first snapshot command to collect initial state information, and subsequent second snapshot commands to collect updates after changes occur. This periodic retrieval of state information ensures completeness while organizing the complex process into manageable periodic cycles rather than a single complex operation.
3Productivity
If differential snapshots are used, then migration speed improves, but accuracy of state replication decreases
Solution Approach 1:
The patent merges the benefits of differential snapshots with complete state information retrieval by combining first snapshots (which capture complete state at a given time) with second snapshots (which capture differential updates). This merging approach allows the system to achieve both speed through differential updates and accuracy through comprehensive state capture, resolving the contradiction between migration speed and replication precision.
Solution Approach 2:
The patent performs preliminary complete state information collection through first snapshot commands before utilizing differential updates in subsequent second snapshot commands. This preliminary action ensures that the baseline state information is accurate and complete, while subsequent differential snapshots build upon this foundation to achieve both speed and accuracy in the overall migration process.
Data Source
AI summary
A live migration/high availability system includes a first computing system having a first SCP subsystem coupled to first computing system components and a first hypervisor subsystem that provides a first virtual machine. Each time the first SCP subsystem receives snapshot commands from the hypervisor subsystem, it retrieves respective SCP component state information that was not retrieved in response to a previous snapshot command from each first SCP component included in the first SCP subsystem, and uses the respective SCP component state information to generate a respective SCP subsystem snapshot based on that snapshot command. The first SCP subsystem then transmits the SCP subsystem snapshots to a second SCP subsystem in a second computing system, and the second SCP subsystem uses the SCP subsystem snapshots to allow a second hypervisor subsystem on the second computing system to provide a second virtual machine that operates the same as the first virtual machine.