Cloud Storage Client Caching for Master Node Failure

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Data cloud storage systems are highly sensitive to the failure of the master node or its standby, leading to system downtime even if the master node has not fully recovered, as existing hot standby methods do not account for simultaneous failures.

Innovation Solution

A method where client terminals query and cache address information of storage servers based on scheduling management service center availability, allowing data storage, query, and download operations to continue even if the service center fails by redirecting requests to available storage servers.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a hot standby method with one master node and one standby master node is used, then the system can continue operation when the master node fails, but the system remains highly sensitive to failures and cannot handle simultaneous failures of both master and standby nodes

Engineering Contradiction:
Improvesystem continuityVSAvoidnode management complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the centralized master node into multiple master nodes (first master node and second master node), distributing the scheduling management functions across multiple independent units. This segmentation eliminates the single point of failure inherent in the traditional hot standby architecture, as each master node can independently handle scheduling tasks when others are unavailable.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent merges the functions of multiple master nodes into a unified scheduling management system where all master nodes share the same scheduling management service center database. This allows seamless failover and load balancing while maintaining system-wide coordination, combining the advantages of both distributed resilience and centralized management.

Inventive Principle:
Principle #5Merging (Combining)

2Device complexity

If a centralized scheduling management service center is used to manage storage servers, then the system structure is simple and easy to manage, but the entire system fails when the scheduling management service center fails

Engineering Contradiction:
Improvesystem structure simplicityVSAvoidsystem availability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent enables each master node to independently access and manage the scheduling management service center database, giving each node local autonomy in handling scheduling tasks. This local quality approach allows the system to maintain simplified centralized database management while achieving distributed fault tolerance, as each master node can independently continue operations even when others fail.

Inventive Principle:
Principle #3Local quality

3Ease of operation

If the system uses a master node architecture for centralized management, then ease of operation is improved, but the system becomes sensitive to master node running state

Engineering Contradiction:
Improvecentralized management easeVSAvoidsensitivity to master node failure
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent implements a dynamic master node architecture where the system can automatically adapt to changes in master node availability. Multiple master nodes can dynamically take over scheduling management tasks based on the running state of other nodes, providing both ease of operation through automated failover and reliability by eliminating single points of failure.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3386168B1Data cloud storage methods, client terminal and data storage server
Publication Date: 2021.03.03 HANGZHOU HIKVISION DIGITAL TECHNOLOGY CO LTD
  • EP3386168B1 patent drawingFigure 1
  • EP3386168B1 patent drawingFigure 2
  • EP3386168B1 patent drawingFigure 3

AI summary

A data cloud storage system, client terminal, storage server and application method, wherein the data storage method includes: sending a storage server query request to a scheduling management service center; receiving and caching first address information when the scheduling management service center is under normal operation, and selecting first address information from the cached first address information; sending a resource request to a storage server corresponding to the first address information; such that the storage server allocates a storage unit for the target data and generates storage resource information; sending a data storage request to the storage server, such that the storage server generates data index information based on attribute information of the target data, and storing the target data in the storage unit allocated to the target data. The embodiments of the present application can guarantee data storage, query and downloading in the event of a failure of the scheduling management service center. Thus, the sensitivity of the data cloud storage system to the failure of the scheduling management service center is reduced.