Cluster File System Message Transmission via Storage Channel

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Cluster file systems face service interruptions and data loss due to unreliable message networks, leading to frequent host restarts and high management costs, as existing solutions either restart hosts or temporarily make them unavailable, causing service disruptions and data unavailability.

Innovation Solution

Implementing a dual-channel message transmission system where hosts switch from a message network channel to a storage channel when the message network fails, allowing continuous communication through the storage channel until the message network is restored, ensuring uninterrupted service and data access.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If hosts use message network for communication in cluster file system, then message transmission is enabled, but network reliability is poor and service interruptions occur

Engineering Contradiction:
Improvemessage network reliabilityVSAvoidservice continuity
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent introduces a storage channel as an intermediary communication path between hosts. When the message network fails, hosts automatically switch to using the storage channel for message transmission, ensuring continuous communication without service interruption. This mediator approach allows the system to maintain functionality through an alternative pathway when the primary communication channel becomes unavailable.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the communication parameter from message network to storage channel based on network status. The system dynamically adjusts the communication pathway by detecting message network failures and switching to the storage channel, thereby adapting to changing conditions and maintaining service continuity despite network reliability issues.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If hosts are restarted when message network fault is detected, then communication reliability is improved, but service interruption and data loss occur

Engineering Contradiction:
Improvecommunication reliabilityVSAvoidservice interruption time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent prepares an alternative storage channel in advance for message transmission. Instead of restarting hosts when network failure occurs, the system has pre-configured the storage channel as a backup communication path, allowing immediate switching without service interruption or data loss when the message network fails.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If hosts exit cluster when message network faults, then fault isolation is achieved, but virtual machine availability decreases and management complexity increases

Engineering Contradiction:
Improvefault isolationVSAvoidcluster management complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The storage channel serves as an intermediary that enables continued cluster operation during network failures. Instead of isolating hosts by exiting the cluster, the storage channel maintains communication pathways, allowing hosts to remain in the cluster and continue operations without increasing management complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3579501B1Message transmission method and device in cluster file system
Publication Date: 2024.05.22 HUAWEI TECH CO LTD
  • EP3579501B1 patent drawingFigure 1~2
  • EP3579501B1 patent drawingFigure 3~4
  • EP3579501B1 patent drawingFigure 5

AI summary

This application relates to the field of communications technologies, and discloses a message transmission method and apparatus in a cluster file system, so as to ensure normal communication between hosts when a message network is faulty. The method includes: determining, by a first host in the cluster file system, that the message network link is faulty; switching from the message network link for message transmission to a storage channel; generating a first message; and recording the first message in a first location area of a data storage medium, where the first location area is used by a second host to access and read the first message.