Cloud Platform Storage Backend Segmentation for Disaster Tolerance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing cloud computing systems have low disaster tolerance and stability due to reliance on a single storage backend, which can lead to system collapse and poor scalability when additional storage nodes are added.

Innovation Solution

Configuring at least two storage backends in the cloud platform, where compute nodes are associated with different storage backends, and image files are stored differently based on their type, ensuring that base images are stored across all backends and exported images are stored with their respective compute nodes, allowing for enhanced disaster tolerance and scalable storage capacity without affecting online virtual machines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a single storage backend is used to store virtual machine image files, then the system structure is simple, but the disaster tolerance is low and system stability deteriorates when the storage backend fails

Engineering Contradiction:
Improvedisaster toleranceVSAvoidstorage backend configuration
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The storage backend is segmented into multiple independent storage backends, each capable of storing virtual machine image files. Compute nodes are associated with specific storage backends, and when one storage backend fails, other storage backends can continue to operate independently, preventing system-wide collapse and improving disaster tolerance.

Inventive Principle:
Principle #1Segmentation

2Quantity of substance

If storage nodes are added to scale out the storage backend, then the storage capacity increases, but the system stability deteriorates and virtual machines may collapse

Engineering Contradiction:
Improvestorage capacityVSAvoidsystem stability
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The storage system is divided into multiple independent storage backends that can be scaled out without affecting each other. When new storage backends are added, they operate independently from existing ones, allowing storage capacity to increase while maintaining system stability and avoiding virtual machine collapse.

Inventive Principle:
Principle #1Segmentation

3Reliability

If all image files are stored in every storage backend, then the disaster tolerance improves, but the storage space is wasted

Engineering Contradiction:
Improvedisaster toleranceVSAvoidstorage space waste
Core Design Contradiction:
ReliabilityVSLoss of substance

Solution Approach 1:

Different image files are stored in different storage backends based on their type and the compute node's association. Base image files are stored in all storage backends for high availability, while exported image files are stored only in the associated storage backend, optimizing both disaster tolerance and storage space utilization.

Inventive Principle:
Principle #3Local quality

4Reliability

If base image files are stored in every storage backend, then the disaster tolerance improves, but the storage space consumption increases

Engineering Contradiction:
Improvedisaster toleranceVSAvoidstorage space consumption
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

Base image files, which are essential for system operation and need to be accessible from any storage backend, are stored in every storage backend. This ensures high disaster tolerance for critical operations while accepting the necessary storage space consumption for these fundamental files.

Inventive Principle:
Principle #3Local quality

5Ease of operation

If exported image files are stored in every storage backend, then the accessibility improves, but the storage efficiency deteriorates

Engineering Contradiction:
Improveimage file accessibilityVSAvoidstorage efficiency
Core Design Contradiction:
Ease of operationVSLoss of substance

Solution Approach 1:

Exported image files, which are specific to certain compute nodes and their associated storage backends, are stored only in the associated storage backend rather than being replicated across all backends. This maintains storage efficiency while providing adequate accessibility for operations within the associated compute node's availability zone.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3442201B1Cloud platform construction method and cloud platform
Publication Date: 2019.11.06 CHINANETCENT TECH
  • EP3442201B1 patent drawingFigure 1~2
  • EP3442201B1 patent drawingFigure 3~4

AI summary

The present disclosure provides a cloud platform and a construction method thereof. The cloud platform includes a plurality of compute nodes and at least two storage backends. The construction method includes: establishing an association relationship between the compute nodes and the at least two storage backends, such that each compute node is associated with one and only one storage backend. Different storage backends are associated with different compute nodes. When the compute node generates an image file, a type of the image file is determined. The type of the image file includes a base image file or an exported image file. The image file is stored according to the type of the image file and the storage backend associated with the compute node that generates the image file. The technical solutions of the present disclosure enhance the overall disaster tolerance and stability of the cloud platform.