Job Routing Controller for Physical and Virtual Nodes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Information processing systems face inefficiencies in job processing due to the difference in processing loads between arithmetic processing jobs and input-output device access jobs, with virtual machines having lower arithmetic processing performance than physical machines, leading to suboptimal execution efficiency.
Innovation Solution
A system that determines the type of a job based on processing load and allocates it to either a physical or virtual calculation node, with high-load jobs executed by physical nodes and low-load jobs by virtual nodes, allowing for efficient resource utilization and reduced wait times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtual machines are used to execute all jobs, then resource consolidation and flexibility are improved, but arithmetic processing performance and execution efficiency deteriorate
Solution Approach 1:
The system segments jobs into two categories: arithmetic processing jobs and I/O access jobs. This segmentation allows different execution strategies to be applied to each job type, with physical machines handling arithmetic jobs and virtual machines handling I/O jobs, thereby resolving the contradiction between resource consolidation and arithmetic processing performance
Solution Approach 2:
The system applies different quality characteristics to different execution environments: physical machines provide high arithmetic processing performance for specific job types, while virtual machines provide resource consolidation and flexibility for other job types. This local differentiation resolves the contradiction by optimizing each environment for its intended purpose
2Productivity
If physical machines are used to execute all jobs, then arithmetic processing performance is improved, but resource utilization efficiency and flexibility deteriorate
Solution Approach 1:
The system segments the workload between physical and virtual machines based on job type characteristics. I/O-intensive jobs are routed to virtual machines, improving resource utilization efficiency, while compute-intensive jobs are routed to physical machines, maintaining high arithmetic processing performance
Solution Approach 2:
The system creates a universal execution platform that can handle both arithmetic processing jobs and I/O access jobs through different execution paths. Physical machines and virtual machines serve multiple functions depending on the job type, optimizing overall resource utilization while maintaining performance for different workloads
3Device complexity
If jobs are executed without type-based allocation, then system simplicity is maintained, but job processing efficiency and wait times deteriorate
Solution Approach 1:
The system performs preliminary classification of jobs into arithmetic processing jobs and I/O access jobs before execution. This preliminary action enables optimal routing decisions to be made in advance, improving job processing efficiency without significantly increasing system complexity
Solution Approach 2:
The system uses feedback from job type analysis to dynamically route jobs to appropriate execution environments. The controller receives information about job characteristics and uses this feedback to make intelligent allocation decisions, improving processing efficiency while maintaining manageable system complexity
Data Source
AI summary
A system includes a plurality of physical calculation nodes and a controller. The controller is configured to execute a process including: determining a type of a target job based on a processing load for execution of the target job, first instructing a first physical calculation node among the plurality of physical calculation nodes to execute the target job when it is determined that the type of the target job is a first type, and second instructing a virtual machine running on a second physical calculation node among the plurality of physical calculation nodes to execute the target job when it is determined that the type of the target job is a second type, a job of the second type being required a lower processing load for the execution of the job than a job of the first type.