Distributed Computing Task Scheduling via Portal Server
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed High Performance Computing (HPC) systems are difficult to use, limiting access to users who need to process large data sets in fields like genomics and nuclear physics, as they require significant computational resources that users may not possess.
Innovation Solution
A distributed computing system that allows clients to schedule and execute applications using cloud-based resources, where clients provide parameters for operations, and a server module selects appropriate computing resources and generates workflows to execute applications on nodes, facilitating access to HPC services without the need for expensive infrastructure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If users directly access distributed HPC systems, then computational power and processing capability are improved, but system complexity and difficulty of operation increase
Solution Approach 1:
The patent introduces a portal server as an intermediary between users and the distributed HPC system. The portal server provides a simplified interface that accepts user requests and automatically translates them into appropriate computing tasks distributed across the HPC network, shielding users from system complexity while maintaining access to high computational power
Solution Approach 2:
The system implements automated resource allocation and task scheduling where the HPC system itself manages the complexity of distributing and managing computational tasks across multiple nodes without requiring user intervention, allowing users to simply submit requests while the system handles the complex coordination automatically
2Productivity
If distributed HPC systems are implemented, then processing capability for large data sets is improved, but infrastructure cost and resource requirements increase
Solution Approach 1:
The patent creates a universal HPC platform that can handle multiple types of computational tasks across different scientific domains (genomics, nuclear physics, analytics) through a single distributed infrastructure, maximizing the utilization of available resources and reducing the need for specialized hardware for each application type
Solution Approach 2:
The system dynamically allocates computational resources based on real-time demand and task requirements, allowing the same infrastructure to adapt to different workloads and user needs, thereby optimizing resource utilization and reducing the total amount of hardware needed compared to static allocations
Data Source
AI summary
A method, system, and computer-readable storage medium for a reconfigurable computing system are disclosed. One method involves configuring one or more computing resources (selected according to a workflow that specifies an application to be executed) of a computing node and executing, using the one or more computing resources, at least a portion of an application at the computing node. At least one of the one or more computing resources is a reconfigurable logic device, and the configuring, at least in part, configures the reconfigurable logic device according to a configuration script of the workflow. The executing comprises performing one or more operations. The one or more operations are performed by the reconfigurable logic device. The reconfigurable logic device is configured to perform the one or more operations by virtue of having been configured according to the configuration script.