Container Cluster Resource Sizing for Fault-Tolerant CPU Allocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing management systems fail to assign an appropriate number of calculation resources corresponding to the appropriate number of CPUs to containers, leading to decreased operation rates of applications implemented by each container.

Innovation Solution

A calculation resource cluster management apparatus that includes a container configuration reception unit and a calculation resource calculation unit to determine an optimal combination of calculation resources and resource amounts, ensuring that the total resource amount exceeds the requested amount, the number of resources is sufficient, and each container can be assigned to at least three resources, with spares for fault-tolerance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the management system assigns calculation resources to containers without optimizing the number of CPUs, then the resource assignment process is simple, but the operation rate of the application decreases

Engineering Contradiction:
Improveoperation rateVSAvoidresource assignment complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The invention changes the parameters of resource assignment by introducing a calculation resource calculation unit that computes optimal combinations of number of calculation resources and resource amount per resource. The system adjusts parameters such as total requested resource amount, requested resource maximum amount, and constraint conditions to determine optimal resource allocation that maximizes operation rate while maintaining manageable complexity through automated calculation.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If the number of calculation resources is increased to ensure fault-tolerance, then the reliability improves, but the resource utilization efficiency decreases

Engineering Contradiction:
Improvefault-toleranceVSAvoidresource utilization efficiency
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The invention applies preliminary action by pre-calculating optimal resource combinations before actual container deployment. The calculation resource calculation unit performs advance computation to determine the exact number of calculation resources and resource amounts needed, incorporating fault-tolerance requirements from the outset. This preliminary calculation ensures that sufficient resources are allocated for reliability while avoiding excessive resource provision that would waste energy and reduce utilization efficiency.

Inventive Principle:
Principle #10Preliminary action

3Quantity of substance

If the resource amount per calculation resource is increased, then the total resource capacity increases, but the number of assignable containers decreases

Engineering Contradiction:
Improvetotal resource capacityVSAvoidnumber of assignable containers
Core Design Contradiction:
Quantity of substanceVSAdaptability or versatility

Solution Approach 1:

The invention implements dynamics by enabling flexible adjustment of resource allocation parameters. The calculation resource calculation unit dynamically determines optimal combinations of number of calculation resources and resource amount per resource based on real-time constraints including total requested resource amount, requested resource maximum amount, and specific constraint conditions. This dynamic adaptation allows the system to optimize resource capacity while maintaining high adaptability to assign the maximum number of containers efficiently.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS12530242B2Computational resource cluster management device, computational resource cluster management method, and computational resource cluster management program
Publication Date: 2026.01.20 NT T INC
  • US12530242B2 patent drawing
  • US12530242B2 patent drawing
  • US12530242B2 patent drawing

AI summary

A cluster management apparatus includes: a container configuration reception unit that receives a configuration definition of an application including a total requested resource amount for a calculation resource cluster and a requested resource maximum amount of a single container which operates on the calculation resource cluster; and a calculation resource calculation unit that calculates a combination of the number of calculation resources and a resource amount per one calculation resource, the combination satisfying constraint conditions that a resource amount of the calculation resource cluster is larger than the total requested resource amount, that the number of calculation resources of the calculation resource cluster is equal to or larger than three, and that the number of calculation resources to which a requested resource maximum amount of a single container included in the calculation resource cluster can be assigned is equal to or larger than three.