Dynamic Virtual Machine Consolidation via Distributed Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current virtual machine consolidation techniques in data centers are limited by centralized planning tools that do not consider dynamic computing resource requirements, leading to issues like over-subscription and under-subscription of resources, resulting in inefficient power usage and performance problems.
Innovation Solution
A distributed and dynamic mechanism for consolidating virtual machines based on real-time calculation of virtual and physical machine metrics, allowing for the migration of virtual machines to optimize resource utilization and reduce the number of active physical machines, thereby reducing power consumption and operational costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If centralized planning tools are used for virtual machine consolidation, then resource allocation can be planned in advance, but the system cannot adapt to dynamic computing resource requirements leading to over-subscription and under-subscription issues
Solution Approach 1:
The patent implements dynamic virtual machine consolidation by enabling physical machines to autonomously make migration decisions based on real-time workload metrics. This transforms the static, centralized planning approach into a dynamic, distributed system where consolidation behavior adapts continuously to changing computing resource requirements, resolving the contradiction between adaptability and system complexity
Solution Approach 2:
The patent applies self-service by empowering individual physical machines to autonomously evaluate their own workload metrics and make independent decisions about virtual machine migration. Each machine monitors its own resource utilization and autonomously determines when to migrate VMs to other machines, eliminating the need for complex centralized coordination while maintaining adaptive response to dynamic conditions
2Reliability
If more physical machines are kept active to handle peak loads, then service availability is improved, but power consumption and operational costs increase
Solution Approach 1:
The patent implements dynamic consolidation that adjusts the number of active physical machines based on real-time workload conditions. During low-utilization periods, VMs are consolidated onto fewer machines allowing others to be powered down, reducing power consumption. During peak loads, the system dynamically activates additional machines, maintaining service availability while optimizing energy usage through adaptive resource allocation
Solution Approach 2:
The patent changes the operational state parameter of physical machines dynamically by transitioning them between active and powered-down states based on workload metrics. This parameter change enables the system to reduce power consumption during low-demand periods while maintaining service availability during peak periods, directly addressing the contradiction between reliability and energy consumption
3Use of energy by stationary object
If virtual machines are consolidated to fewer physical machines, then power consumption is reduced, but resource over-subscription can occur leading to performance problems
Solution Approach 1:
The patent implements feedback mechanisms where physical machines continuously monitor workload metrics and use this information to make informed migration decisions. This feedback loop prevents over-subscription by detecting when consolidation would lead to performance degradation and adjusting consolidation levels accordingly, maintaining performance stability while achieving power consumption reduction through optimized VM placement
Solution Approach 2:
The patent applies partial action by implementing conservative consolidation strategies that migrate VMs only when workload metrics indicate sufficient capacity on target machines. This partial approach to consolidation prevents over-subscription and performance degradation while still achieving meaningful power consumption reductions, balancing the trade-off between energy efficiency and performance stability
Data Source
AI summary
At a first physical computing machine executing a plurality of virtual machines and connected to a network, one or more virtual machine metrics for each virtual machine are calculated. Each virtual machine metric represents a workload of a resource of the first physical computing machine due to the execution of a corresponding virtual machine. Additionally, one or more corresponding physical machine metrics that represent a total workload of the corresponding resource of the first physical computing machine due to the execution of the plurality of virtual machines are also calculated. Based on the one or more physical machine metrics, a determination is made that at least one of the plurality of virtual machines should be migrated to one of a plurality of other physical computing machines connected to the network. A first virtual machine is selected for migration to a selected second physical computing machine.