Cloud VM Live Migration Prioritization Under Bandwidth Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud service migration technologies are inadequate in managing the live migration of multiple virtual machines (VMs) during disruptions, as it can take a long time to migrate a large number of VMs, which may not be feasible during cloud service disruptions, necessitating a system to prioritize and optimize VM migration to maximize high-priority service availability.
Innovation Solution
An automatic live migration system that uses a framework with a VM Monitor, Anomaly Detector, Cloud Service Selector, Target Selector, and Migration Trigger to identify and prioritize VMs for migration, determine the best migration location, and manage the migration process under time, memory, and bandwidth constraints, employing machine learning algorithms for anomaly detection and resource optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all VMs are migrated during cloud disruption, then service availability is improved, but migration time becomes excessively long and may exceed the disruption window
Solution Approach 1:
The patent segments the set of VMs into priority groups (high priority, medium priority, low priority) based on service criticality. The migration system processes high-priority VMs first to ensure critical services are restored quickly, then progressively migrates lower-priority VMs as resources and time permit. This segmentation resolves the contradiction by ensuring service availability for critical functions without requiring migration of all VMs within a fixed time window.
Solution Approach 2:
The patent implements partial migration by selecting and migrating only the most critical VMs during acute disruption periods when time is constrained. The system dynamically adjusts the migration scope based on available time, network bandwidth, and resource availability, performing sufficient migration action to restore essential services without attempting to migrate the entire VM fleet, thus resolving the time availability contradiction.
2Ease of operation
If VM migration is performed manually, then migration control is improved, but operational complexity and response time worsen during disruptions
Solution Approach 1:
The patent implements an automated migration system that autonomously monitors cloud infrastructure health, detects disruptions, prioritizes VMs based on service criticality, selects appropriate destination hosts, and executes migration without human intervention. The system self-manages the entire migration workflow, resolving the contradiction by eliminating manual operational complexity while maintaining precise control through algorithmic decision-making and automated orchestration.
Solution Approach 2:
The patent incorporates continuous feedback loops where the system monitors VM performance metrics, service health status, and resource availability in real-time. Based on this feedback, the system dynamically adjusts migration decisions, prioritization criteria, and resource allocation. This feedback mechanism enables automated control while adapting to changing conditions, resolving the contradiction between automation and control.
3Use of energy by moving object
If non-live migration is used, then migration resource consumption is reduced, but service downtime increases during migration
Solution Approach 1:
The patent performs preliminary actions by pre-selecting destination hosts and pre-establishing migration pathways before actual VM migration begins. The system pre-allocates resources at destination hosts and prepares migration environments in advance, so that when disruption occurs and migration is triggered, the process can proceed efficiently with minimal downtime and optimized resource utilization during the actual migration execution.
Data Source
AI summary
A system and method maximize the availability of cloud services in the event of a disruption to one or more cloud servers by automatically triggering the live migration of selected cloud services and automatically performing the triggered migration of such services. The operational state of virtual machines operating on a cloud server and the services that are associated with the virtual machines is monitored. Different techniques are used for deciding when to migrate cloud services based on the automatic detection of anomalies, for deciding what cloud services to migrate so as to maximize the availability of high priority services during migration under time and network bandwidth constraints, and for deciding where to migrate the selected cloud services based on a determination of the best location for migration.