JBOD Resource Scheduling for Dual-Head Node Failover

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing data storage systems using Just a Bunch of Disks (JBOD) suffer from low scheduling efficiency and poor reliability due to misalignment of JBODs mounted on storage service nodes, especially in dual-head systems, leading to inconsistent mapping information and inefficiencies during node failures.

Innovation Solution

A control and scheduling service node acquires attribution data of JBODs to manage mapping relationships and detects abnormal storage service nodes, scheduling JBODs from abnormal to normal nodes within dual-head groups, ensuring quick and reliable resource allocation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If JBODs are mounted on storage service nodes in a dual-head system, then data storage availability is improved, but scheduling efficiency and reliability deteriorate due to misalignment of JBOD mapping information

Engineering Contradiction:
Improvedata storage availabilityVSAvoidscheduling efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent introduces a dedicated control and scheduling service node as an intermediary between storage service nodes and JBODs. This mediator manages all JBOD mounting, dismounting, and migration operations, centralizing control logic and eliminating the need for storage service nodes to independently manage JBOD mapping. The control node maintains accurate attribution data and coordinates all scheduling decisions, thereby improving scheduling efficiency while preserving the high availability benefits of dual-head systems.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The control and scheduling service node performs multiple functions: it manages JBOD mounting/dismounting operations, maintains attribution data mapping, detects abnormal storage service nodes, and executes JBOD migration. By consolidating these diverse scheduling tasks into a single multi-functional node, the system achieves coordinated control over JBOD resources across multiple storage service nodes, resolving the misalignment issues that plague distributed scheduling approaches.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Stability of the object's composition

If JBODs are mounted on storage service nodes for high availability, then system stability is improved, but mapping information consistency deteriorates leading to poor reliability

Engineering Contradiction:
Improvesystem stabilityVSAvoidmapping information consistency
Core Design Contradiction:
Stability of the object's compositionVSReliability

Solution Approach 1:

The control and scheduling service node continuously monitors the mounting status of JBODs on storage service nodes and maintains an updated attribution data mapping. When a storage service node is added, removed, or experiences failures, the control node receives feedback about these changes and automatically updates the JBOD mapping information. This feedback mechanism ensures that the centralized attribution data remains consistent with the actual system state, preventing mapping inconsistencies even as the system configuration changes.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

By positioning the control and scheduling service node as an intermediary between storage service nodes and JBODs, the patent creates a single source of truth for JBOD attribution data. This mediator validates and coordinates all mounting and dismounting operations, ensuring that mapping information remains consistent across the system. The intermediary role prevents divergent mapping states that would otherwise occur in distributed scheduling scenarios.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Quantity of substance

If multiple JBODs are managed by multiple storage service nodes, then storage capacity is improved, but scheduling complexity increases leading to low efficiency

Engineering Contradiction:
Improvestorage capacityVSAvoidscheduling complexity
Core Design Contradiction:
Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent extracts the complex scheduling logic from individual storage service nodes and consolidates it into a dedicated control and scheduling service node. Storage service nodes are relieved of JBOD management responsibilities and focus solely on data storage operations. The control node assumes full responsibility for JBOD mounting, dismounting, migration, and attribution data management. This extraction of scheduling functions reduces the operational complexity at each storage node while maintaining the ability to manage multiple JBODs across the system.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The system architecture is segmented into distinct functional components: storage service nodes handle data operations, while the control and scheduling service node handles all JBOD resource management. This segmentation separates concerns and allows each component to specialize in its designated function. The control node manages the complexity of coordinating multiple JBODs across multiple storage nodes through centralized attribution data and coordinated scheduling decisions, while storage nodes maintain simpler, focused operations.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4671951A1Method, apparatus, system, device and storage medium for scheduling resource for just a bunch of disks
Publication Date: 2025.12.31 BEIJING VOLCANO ENGINE TECH CO LTD
  • EP4671951A1 patent drawingFigure 1~2
  • EP4671951A1 patent drawingFigure 3~4
  • EP4671951A1 patent drawingFigure 5~6

AI summary

The disclosed embodiments provide a method, an apparatus, a system, a device and a storage medium for scheduling a resource for Just a Bunch of Disks (JBOD). The method comprises: acquiring (S101), by a control and scheduling service node, attribution data of the JBOD, wherein the attribution data of the JBOD represents a mapping relationship between at least one JBOD and a storage service node to which the at least one JBOD belongs; and detecting (S102), by the control and scheduling service node, an abnormal storage service node in the dual-head storage service node group, and scheduling, by the control and scheduling service node, a JBOD mounted on the abnormal storage service node to a normal storage service node in the dual-head storage service node group based on the attribution data of the JBOD.