Cross-data center computing power resource scheduling method and device based on container

By adopting a container-based cross-data center computing resource scheduling method, combined with KubeSphere and Volcano Scheduler, and employing improved scheduling algorithms and resource mounting technology, the problems of low resource utilization and insufficient security in existing technologies are solved, achieving efficient and flexible computing resource management.

CN118916164APending Publication Date: 2024-11-08CHANGSHA RUISUAN TECHNOLOGY CO LTD +3
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202411003177.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Existing computing task scheduling methods cannot meet the needs of different types of computing tasks, have low resource utilization, lack rapid deployment, portability and expansion capabilities, and have insufficient security and isolation.

Method used

A container-based cross-datacenter computing resource scheduling method is adopted. By combining KubeSphere with computing scheduling, the Volcano Scheduler component is used to parse scheduling requests. Combined with LeastRequestedPriority and an improved Binpack scheduling algorithm, container instances are created and resources such as GPUs are mounted, providing SSH login access.

Benefits of technology

It achieves lightweight, flexible, and secure computing power scheduling, improves resource utilization and ease of use, supports multiple storage methods, and is easy to port and expand.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916164A_ABST
    Figure CN118916164A_ABST
Patent Text Reader

Abstract

The embodiment of the invention discloses a container-based cross-data center computing power resource scheduling method and device, and relates to the technical field of computing power scheduling, and the method comprises the steps: receiving a computing power scheduling request sent by a user based on a computing power scheduling task; analyzing the computing power scheduling request, and obtaining a target physical resource node corresponding to the computing power scheduling request according to an analysis result; creating a container instance at the target physical resource node, and operating a scheduling algorithm to assign the container instance to a target worker node in the target physical resource node; and mounting a target computing power resource on the container instance according to the computing power resource information, and providing an entrance for external login access on the container instance. According to the embodiment of the invention, the container arrangement service platform KubeSphere is combined with the computing power scheduling, so that the lightweight, flexible access and release of the computing power scheduling can be realized, the computing power scheduling is easy to transplant and expand, and the isolation and security are good.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • Large model layered caching loading method oriented to Plevatory capacity intelligent computing center

    CN120216463A

  • Large model hierarchical cache loading method for inclusive computing power intelligent computing center

    CN120216463B

  • Intelligent computing resource sharing scheduling method and device based on non-invasive technology

    CN121807573A

  • A method and device for sharing and scheduling intelligent computing resources based on non-invasive technology

    CN121807573B