HPC Resource Allocation Decision Support for Fast Instance Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The increasing complexity of high-performance computing (HPC) infrastructures and the lack of IT expertise among users make it difficult to quickly identify and allocate the necessary computing resources, especially with the growing size and diversity of clusters, leading to manual and inefficient resource allocation processes.

Innovation Solution

A decision support method and device for allocating computing resources on HPC infrastructure, utilizing a communication module, data processing module, and reservation management module to identify and manage instances meeting specific resource requirements, including graphical representation and alternative solutions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If manual resource allocation is used, then flexibility and user-specific customization are maintained, but allocation efficiency and speed deteriorate as infrastructure size increases

Engineering Contradiction:

Improveallocation speedVSAvoidallocation process complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary device positioned between users and the HPC system that automates resource allocation. This intermediary translates user requirements into automated allocation decisions, resolving the contradiction by providing both speed (automated processing) and flexibility (user-specific criteria) without requiring manual intervention for each allocation request.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The allocation device enables self-service automation where the system automatically identifies and allocates resources based on predefined user requirements and availability criteria. This eliminates manual allocation processes while maintaining adaptability to different user needs, thereby improving productivity without proportionally increasing operational complexity.

Inventive Principle:

Principle #25Self-service

2Productivity

If automated resource allocation is implemented, then allocation speed and efficiency improve, but adaptability to diverse user needs and resource diversity deteriorate

Engineering Contradiction:

Improveallocation speedVSAvoiduser requirement adaptability

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic adaptability where the allocation device can adjust its behavior based on user-specific requirements, resource availability, and changing conditions. The system dynamically selects allocation strategies and criteria tailored to each user request, maintaining high adaptability while operating at automated speeds through programmable decision logic.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system changes operational parameters based on user requirements and resource states. By dynamically adjusting allocation parameters such as resource selection criteria, availability thresholds, and matching rules, the system maintains versatility across diverse user needs while operating through automated processes that ensure speed and efficiency.

Inventive Principle:

Principle #35Parameter changes

3Measurement precision

If comprehensive resource tracking is implemented, then allocation accuracy and user needs matching improve, but system complexity and data processing requirements worsen

Engineering Contradiction:

Improveresource availability accuracyVSAvoidtracking system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the tracking system into modular components that monitor specific resource attributes independently. By dividing the comprehensive tracking function into separate monitoring modules for different resource types and attributes, the system achieves high measurement precision while managing complexity through modular architecture and specialized data structures.

Inventive Principle:

Principle #1Segmentation

4Ease of operation

If manual resource allocation is used, then user expertise requirements are minimized, but allocation time and operational burden increase

Engineering Contradiction:

Improveuser operation simplicityVSAvoidallocation time

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The allocation device enables users to specify their requirements in simple terms without needing HPC expertise, while the system automatically performs the complex allocation process. This self-service approach maintains ease of operation by allowing straightforward user input while eliminating time losses through automated processing of the allocation logic and resource matching.

Inventive Principle:

Principle #25Self-service

Data Source

PatentEP3674897B1Decision support method and device for allocating computation means on a high performance computing infrastructure

Publication Date: 2026.05.20 BULL SA

Figure 1~2
Figure 3A~3C
Figure 4

AI summary

The invention relates to a decision support method for resource allocation on an HPC-type infrastructure, enabling the identification of a set of instances meeting a resource requirement. The invention further relates to a computer device (1) comprising a storage module (11) configured to store a resource class repository, a reservation repository, and a keyword repository, a data processing module (13), a reservation management module (14), and a communication module (12), said storage module (11) containing instructions for a program whose execution by said data processing module (13) triggers the implementation of the method (100).