Virtual Machine Co-location for Power Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional load balancing techniques for virtual machines across physical hardware resources lead to inefficient power consumption due to underutilized resources, as they do not account for communication loads between virtual machines, resulting in high overhead power usage.
Innovation Solution
A system and method that identifies affinity relationships between virtual machines through communication monitoring and co-locates them on common processing resources, reducing power consumption by optimizing resource allocation and communication loads.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional load balancing techniques are used to spread processing and memory usage evenly across all available physical hardware resources, then hardware resource utilization is improved, but power consumption increases due to underutilized resources running at high overhead
Solution Approach 1:
The patent merges multiple virtual machines that have affinity relationships (frequent communication patterns) onto the same physical hardware resource. This consolidation reduces the total number of active physical resources needed, thereby lowering power consumption while maintaining or improving overall system productivity through optimized resource utilization.
Solution Approach 2:
The patent applies local quality by making load balancing decisions based on local communication patterns between virtual machines. By monitoring and identifying affinity relationships, the system creates localized optimizations where frequently communicating virtual machines are co-located, rather than applying uniform load balancing across all resources. This localized approach reduces unnecessary network traffic and improves energy efficiency.
2Reliability
If virtual machines are migrated to other physical processing resources when over-utilized, then load balancing is improved, but communication efficiency deteriorates due to separation of affinity-related virtual machines
Solution Approach 1:
The patent implements feedback by continuously monitoring communication patterns between virtual machines to identify affinity relationships. This feedback mechanism allows the system to make informed migration decisions, keeping frequently communicating virtual machines together on the same physical resource, thereby maintaining communication efficiency while still achieving load balancing when necessary.
Solution Approach 2:
The patent applies preliminary action by proactively identifying affinity relationships between virtual machines before migration decisions are made. By pre-establishing which virtual machines should be co-located based on their communication patterns, the system can make faster and more accurate placement decisions, reducing communication overhead and improving overall system efficiency.
Data Source
AI summary
Communications between virtual machines are monitored to identify virtual machines that have an affinity with each other, such as where the virtual machines have greater than a threshold of communication between each other. An affinity table tracks virtual machines having an affinity relationship and is referenced upon start-up or migration of a virtual machine so that a starting-up or migrating virtual machine will run on the same processing resource as virtual machines with which it has an affinity relationship.