Live Migration Downtime Control via Phase Interruption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computing environments that share resources struggle with accurately tracking and limiting downtime during the migration of computing processes between virtual machines, often resulting in excessive downtime due to delays in initializing execution on new virtual machines.
Innovation Solution
A method is introduced that determines a migration time limit for the stop and copy phase of the migration process by subtracting the predicted worst-case start time from the allowed downtime, allowing for the detection and interruption of the migration if it exceeds this limit, ensuring the computing process resumes on the original virtual machine to enforce a guaranteed maximum downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the migration process is allowed to proceed without time limits, then the computing process can be fully migrated to the new virtual machine, but the downtime exceeds the allowed period and reliability decreases
Solution Approach 1:
The system performs preliminary actions by calculating the migration time limit before the migration phase begins. The formula (allowed downtime - predicted worst case start time) = migration time limit is computed in advance, and this pre-determined limit is then enforced during the migration phase to guarantee maximum downtime requirements are met.
Solution Approach 2:
The system implements feedback by monitoring the actual migration phase duration and comparing it against the pre-calculated migration time limit. If the monitored duration approaches or exceeds the limit, the system intervenes to halt the migration, ensuring the allowed downtime is not violated. This closed-loop control provides feedback to maintain reliability constraints.
2Reliability
If the migration time limit is set too strictly, then downtime is controlled, but the migration may be interrupted even when completion is imminent, reducing productivity
Solution Approach 1:
The system changes the parameter of time limit calculation by using the formula (allowed downtime - predicted worst case start time) instead of simply using the full allowed downtime. This parameter adjustment accounts for the predicted initialization delay on the new virtual machine, effectively extending the usable migration window while still guaranteeing the maximum allowed downtime requirement.
3Reliability
If the system monitors and enforces migration time limits, then downtime is guaranteed, but the complexity of the migration management system increases
Solution Approach 1:
The migration process is segmented into distinct phases: a first phase where the computing process continues executing on the original virtual machine, and a second phase where execution is halted and the migration time limit is enforced. This segmentation allows the system to apply different control strategies to different phases, simplifying the overall management complexity while maintaining reliability guarantees.
Data Source
AI summary
Methods and systems for improved live migration of computing processes with guaranteed maximum downtime are provided. In a first embodiment, a method is provided that includes migrating a computing process between two virtual machines according to a first migration phase. The computing process may continue executing during the first migration phase. A second migration phase may begin, in which execution of the computing process may stop. It may be detected that a duration of the second migration phase exceeds a predetermined period of time, and the second migration phase may be halted such that the computing process continues executing on an original virtual machine. The predetermined period of time can be determined based on a downtime for resuming execution of the computing process and a predicted worst case start time for the computing process.