Virtual Machine Placement Map Construction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inefficient placement algorithms for guest virtual machines in distributed computer systems lead to significant time consumption and system slowdowns, especially in large or complex environments, affecting system availability and performance.
Innovation Solution
A placer process constructs a placement map representing available capacity and latency between host machines, iteratively solving and legalizing it to determine optimal placements while satisfying constraints, using techniques like metaheuristic optimization and phantom capacity requests to manage resource allocation efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional placement algorithms are used to allocate guest virtual machines to host computer environments, then the system can handle basic placement requirements, but the placement process consumes significant time and becomes increasingly slow as the number of guest virtual machines and host systems grows
Solution Approach 1:
The patent pre-calculates and stores compatibility information between guest virtual machines and host computer environments in a data structure before placement is needed. This preliminary action allows the placement algorithm to quickly retrieve pre-computed compatibility data instead of performing complex calculations during the actual placement process, significantly reducing placement time while maintaining accuracy even as system size grows
Solution Approach 2:
The patent segments the placement problem by dividing it into distinct components: guest virtual machine requirements, host environment capabilities, and compatibility rules. By organizing these elements separately in a structured data format, the system can efficiently query and match components without processing the entire system state, improving scalability and placement speed
2Reliability
If traditional placement algorithms are used in large or crowded host system environments, then basic placement can be achieved, but system slowdowns and outages occur adversely affecting system availability and performance
Solution Approach 1:
The patent pre-identifies and stores compatibility constraints and requirements for each guest virtual machine and host environment combination. By having this compatibility information pre-computed and stored in an efficient data structure, the system can make rapid placement decisions that avoid incompatible configurations, preventing system slowdowns and outages while maintaining high availability and performance even in large or crowded environments
Solution Approach 2:
The patent implements a feedback mechanism where the placement algorithm continuously monitors system state and uses pre-stored compatibility data to make informed placement decisions. This feedback loop ensures that placements are made that maintain system stability and performance, avoiding configurations that would lead to slowdowns or outages
Data Source
AI summary
Techniques for placing virtual machines at hardware locations are described herein. A first assignment of virtual machines to hardware locations that fails to satisfy a set of placement criteria is detected, a second assignment of virtual machines to hardware locations that assigns a subset of the set of virtual machines to new hardware locations is generated, and it is determined if the second assignment of virtual machines to hardware locations satisfies a subset of the set of placement criteria.