Distributed Display Sessions for Scalable HPC GUI Access

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing remote desktop tools for HPC systems are centralized and lack efficient methods for distributed graphical user interface (GUI) management across multiple compute nodes, limiting user interaction and resource utilization.

Innovation Solution

A distributed display system utilizing a dispatching system grid, dedicated virtual display managers and proxies, and a distributed display manager to assign compute nodes, initiate display managers, and generate visual representations for user interaction, enabling seamless GUI-based application execution and management across a distributed computer architecture.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If a centralized Display Manager is used for remote GUI access, then user interaction capability is provided, but system scalability and resource utilization are limited

Engineering Contradiction:

Improveuser interaction capabilityVSAvoidsystem scalability

Core Design Contradiction:

Ease of operationVSAdaptability or versatility

Solution Approach 1:

The centralized Display Manager is segmented into multiple distributed Display Managers, each running on separate compute nodes. This allows the system to scale by adding more compute nodes with Display Managers without requiring a single centralized management system, thereby improving scalability while maintaining user interaction capability.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

A display proxy server is introduced as an intermediary component between the user's local machine and the distributed Display Managers. This proxy manages the connection pooling and routing, enabling multiple users to access multiple compute nodes efficiently, thus improving both scalability and ease of operation.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of operation

If a centralized Display Manager is used, then GUI access is enabled, but resource utilization across multiple compute nodes is inefficient

Engineering Contradiction:

ImproveGUI accessVSAvoidresource utilization

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The system segments the display management function across multiple compute nodes, allowing each node to independently manage its own Display Manager. This enables efficient resource utilization as compute nodes can be dynamically assigned to users based on availability and requirements, rather than relying on a single centralized manager.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system implements dynamic resource allocation where compute nodes and Display Managers can be dynamically assigned and reassigned based on user requests and node availability. This dynamic behavior optimizes resource utilization while maintaining convenient GUI access for users.

Inventive Principle:

Principle #15Dynamics

3Adaptability or versatility

If dedicated Display Managers are initiated on each compute node, then distributed GUI management is enabled, but system complexity increases

Engineering Contradiction:

Improvedistributed GUI managementVSAvoidsystem complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The display proxy server acts as an intermediary that simplifies the complexity of managing multiple distributed Display Managers. It handles connection pooling, routing, and communication protocols, thereby enabling distributed GUI management while reducing the operational complexity for users and administrators.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The display proxy server performs multiple functions including connection management, authentication, routing, and protocol translation. This multi-functionality consolidates complex management tasks into a single universal component, reducing overall system complexity while enabling distributed GUI management.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Ease of operation

If remote viewer tools are used for HPC systems, then user interaction is improved, but the tools lack integrated display management capabilities

Engineering Contradiction:

Improveuser interactionVSAvoiddisplay management integration

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The invention merges the remote viewer functionality with integrated display management capabilities into a unified system. The display proxy server combines connection management, authentication, and display routing functions, providing both improved user interaction and integrated management in a single solution rather than separate tools.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20250362940A1Method and system for providing distributed display system for HPC and grid compute environments

Publication Date: 2025.11.27 SNIR SHAMIR SHMUEL

AI summary

The present invention relates generally to a system designed to provide application graphical display while using High-Performance Computing systems which are based on GRID systems, in which applications can be executed on any one out of multiple compute nodes while the user interactive session (also known as an X session) is viewed from a control session. In specific, the invention relates generally to a system which creates a separate selfsustained X Display session for each launched application, while a list of available sessions is provided to the main user graphical session allowing interacting with the remote application without directly affecting its integrity in case of disconnection.