Virtual Machine Scheduling via Matching Score Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual machine scheduling methods in cloud computing face challenges in achieving accurate and efficient resource allocation, leading to high scheduling costs and negative impacts on service experience due to poor accuracy in primary scheduling and lengthy secondary scheduling processes.
Innovation Solution
A virtual machine scheduling method that acquires optimal resource distribution information across physical machines in a cloud computing system, determines a matching score between physical machines and virtual machine instances, and allocates resources to the physical machine with the highest matching score for improved resource utilization and scheduling efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If primary scheduling is performed to allocate physical machines for virtual machine instances, then real-time performance is ensured, but scheduling accuracy deteriorates leading to poor resource distribution
Solution Approach 1:
The patent performs preliminary scheduling actions by pre-calculating optimal resource distribution and identifying candidate physical machines before actual virtual machine deployment. This preliminary action includes analyzing resource metrics, computing suitability scores, and preparing allocation plans in advance, which improves both real-time performance and scheduling accuracy by reducing the complexity of real-time decision-making.
2Manufacturing precision
If secondary scheduling is performed to rearrange deployed virtual machines, then resource distribution合理性 is improved, but time consumption increases due to lengthy migration processes
Solution Approach 1:
The patent performs preliminary optimization by pre-calculating optimal resource distribution plans and identifying migration targets before actual virtual machine migration. This includes computing suitability scores for candidate physical machines, determining optimal migration sequences, and preparing allocation plans in advance, which reduces the time consumption of secondary scheduling while maintaining resource distribution合理性.
Solution Approach 2:
The patent changes the parameter of resource allocation by introducing a suitability score metric that combines multiple resource metrics (CPU utilization, memory availability, storage capacity, network bandwidth) into a single composite parameter. This parameter transformation enables more efficient decision-making during secondary scheduling, improving resource distribution合理性 without proportionally increasing time consumption.
3Manufacturing precision
If more virtual machines are migrated during secondary scheduling, then overall resource distribution is improved, but scheduling cost increases
Solution Approach 1:
The patent transforms multiple resource metrics into a single suitability score parameter that quantifies the optimal placement of virtual machines. This parameter change enables precise identification of high-value migration opportunities, allowing the system to achieve improved overall resource distribution by migrating only the most beneficial virtual machines rather than performing exhaustive migrations, thus reducing scheduling cost.
Solution Approach 2:
The patent creates a virtual model or copy of the resource distribution state to perform simulations and evaluations before actual migration. By copying the current resource allocation state and testing different migration scenarios in this virtual model, the system can identify the most cost-effective migration plans that improve resource distribution with minimal actual migration operations.
Data Source
AI summary
A virtual machine scheduling method, an electronic device and a storage medium. The method includes: acquiring optimal resource distribution information of a plurality of physical machines in a cloud computing system, wherein the optimal resource distribution information includes specifications and numbers of specifications of virtual machine instances expected to be deployed on idle resources of the plurality of physical machines; for a first to-be-deployed virtual machine instance, determining a score of matching degree between each physical machine and the first to-be-deployed virtual machine instance based on the optimal resource distribution information; and determining the physical machine with the highest score of matching degree as a host machine of the first to-be-deployed virtual machine instance, for resource allocation.