Virtual Machine Migration in Cluster Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional virtualization technologies face challenges in optimizing resource utilization and administrative efficiency in cluster systems, as they often result in low server utilization and high administrative costs due to the need for maintaining multiple server platforms, and lack effective mechanisms for dynamically reallocating resources and migrating virtual machines based on performance data.
Innovation Solution
A cluster system architecture with Single System Image (SSI) functionality that allows for the execution of virtual machines across multiple nodes, utilizing workload management algorithms to analyze performance data against service level objectives, reallocating resources and migrating virtual machines to optimize resource utilization and achieve desired performance levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If virtual machines are executed on multiple discrete server platforms, then administrative flexibility and fault tolerance are improved, but resource utilization deteriorates and administrative costs increase
Solution Approach 1:
The patent merges multiple discrete server platforms into a unified cluster system with shared storage and coordinated management. Virtual machines can be migrated across cluster members, combining the resources of multiple servers while maintaining the administrative flexibility and fault tolerance of having multiple platforms. This resolves the contradiction by achieving both high reliability through cluster redundancy and high resource utilization through centralized management and load balancing.
Solution Approach 2:
The cluster system creates a universal platform where any server can host any virtual machine through standardized interfaces and shared storage access. This multi-functionality allows resources to be dynamically allocated across different servers based on demand, improving overall resource utilization while maintaining the ability to provide multiple independent server functions for fault tolerance and administrative flexibility.
2Productivity
If virtual machines are migrated between server platforms, then resource utilization and performance optimization are improved, but system complexity increases
Solution Approach 1:
The patent introduces a gateway server as an intermediary that manages virtual machine migration between cluster members. The gateway server handles the complexity of coordination, state management, and resource allocation, while presenting a simplified interface to administrators. This mediator approach enables resource utilization optimization through migration without requiring each server to implement complex migration logic, thus reducing overall system complexity.
Solution Approach 2:
The system implements performance monitoring and analysis that provides feedback to the migration management process. The gateway server continuously monitors virtual machine performance and resource utilization, using this feedback to make intelligent migration decisions that optimize resource allocation. This feedback mechanism automates the complexity of migration management, reducing the need for manual intervention while achieving high resource utilization.
3Ease of operation
If performance monitoring and analysis are implemented, then virtual machine migration decisions are improved, but processing overhead increases
Solution Approach 1:
The patent implements selective performance monitoring that focuses on key metrics relevant to migration decisions rather than comprehensive monitoring of all system parameters. The gateway server analyzes performance data from multiple virtual machines and makes migration decisions based on aggregated trends rather than continuous real-time analysis of every metric. This partial action approach provides sufficient information for accurate migration decisions while minimizing processing overhead and energy consumption.
Data Source
AI summary
In one embodiment, a method comprises executing a plurality of virtual machines on a plurality of nodes of a cluster computing system, wherein at least one application is executed within each of the plurality of virtual machines, generating data that is related to performance of applications in the virtual machines, analyzing, by a management process, the data in view of parameters that encode desired performance levels of applications, and migrating, by the management process, a virtual machine on a first node to a second node of the plurality of nodes in response to the analyzing.