Method for managing local high-performance storage based on logical volumes in a cluster

By employing logical volume management technology and container interface service plugins in the cluster to dynamically create and bind logical volumes, the problems of low storage resource utilization and insufficient security isolation in cloud-native storage solutions are solved, achieving efficient storage resource management and billing, and improving the return on infrastructure investment.

CN122308730APending Publication Date: 2026-06-30BEIJING INBO DIGITAL TECH CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING INBO DIGITAL TECH CO LTD
Filing Date
2026-03-18
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing cloud-native storage solutions suffer from problems such as low storage resource utilization, inaccurate allocation, insufficient security isolation, difficulty in billing and accounting, and risk of resource overuse when using local high-performance solid-state drives on GPU servers.

Method used

Logical Volume Management (LVM) technology is used to aggregate disk space in the node cluster, and a container interface service plugin is deployed. Logical volumes are dynamically created and bound through the cluster scheduler to achieve dynamic allocation and management of storage resources.

Benefits of technology

It improves storage hardware utilization, enables collaborative scheduling of storage and GPU computing power, simplifies storage lifecycle management, enhances operational flexibility, and supports secure isolation and accurate billing in multi-user environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122308730A_ABST
    Figure CN122308730A_ABST
Patent Text Reader

Abstract

This disclosure presents a method for local high-performance storage processing based on logical volume management in a cluster. One specific implementation of this method includes: aggregating the disk space corresponding to each specified cluster node in the node cluster to obtain a resource storage pool; deploying a corresponding container interface service plugin for each cluster node in the node cluster; performing resource validity verification on the resource object to be constructed; in response to the successful resource validity verification of the resource object creation task, controlling the cluster scheduler to select a cluster node adapted to the resource object to be constructed as the target cluster node; controlling the container interface service plugin in the target cluster node to dynamically create the resource object and its corresponding logical volume in the resource storage pool of the target cluster node according to the resource object creation task, and binding the logical volume to the resource object. This implementation ensures efficient recycling of storage space.
Need to check novelty before this filing date? Find Prior Art