Cloud Service Migration via Asynchronous Disk Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud service migration methods result in prolonged service interruptions and complex operations due to the lack of comprehensive solutions for synchronizing large data sets across geographic regions, leading to inefficient cloud service distribution and user experience enhancement.
Innovation Solution
A method for cloud service migration that involves obtaining a migration request, scheduling the migration, instantiating a new cloud server instance in the target cluster, configuring resources, and asynchronously copying disk data using a migration priority order, allowing for live migration and minimizing service interruption time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If disk data is mirrored using traditional copying methods, then data synchronization is achieved, but service interruption time is prolonged due to the large size of disk data (hundreds of gigabytes or terabytes)
Solution Approach 1:
The patent segments the large disk data into multiple data blocks, each with a unique identifier. This segmentation allows parallel processing and asynchronous copying of different blocks, significantly reducing the total migration time compared to traditional sequential copying methods while ensuring complete data synchronization.
Solution Approach 2:
The patent performs preliminary actions by creating mirrored images of data blocks in advance and storing them in a migration pool before actual migration. This preliminary preparation enables the migration process to start with ready-to-copy blocks, reducing the critical path time and allowing service to be restored faster once migration begins.
2Adaptability or versatility
If comprehensive cloud service migration is implemented, then service distribution and user experience are improved, but operational complexity increases due to multiple configuration tasks
Solution Approach 1:
The patent provides a universal migration system that handles multiple cloud service components (disk data, cloud service instances, configuration information) through a single integrated process. The system can migrate different types of data and services using the same underlying mechanism, reducing operational complexity while maintaining comprehensive migration capabilities.
Solution Approach 2:
The migration system automatically performs configuration tasks such as creating cloud service instances, allocating resources, and setting up data operations without requiring manual intervention for each step. The system self-manages the complex coordination between source and target clusters, reducing the operational burden on users.
3Reliability
If disk data is copied synchronously, then data consistency is ensured, but migration speed is reduced due to the need to wait for complete data transfer
Solution Approach 1:
The patent implements dynamic migration by allowing the migration process to proceed asynchronously with different data blocks being copied at different rates. The system dynamically adjusts the migration pace based on available resources and service requirements, using priority levels to ensure critical data is migrated first while maintaining overall data consistency through metadata tracking.
Solution Approach 2:
The patent introduces a migration pool as an intermediary storage area that holds mirrored data blocks during the migration process. This intermediary allows the source system to continue operating while data is being prepared and transferred to the target, enabling asynchronous migration without compromising data consistency through the use of block identifiers and metadata.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The disclosure provides a method and for cloud service migration. The method comprises: obtaining a migration request related to a cloud service hosted in a source cluster, the migration request comprising a scheduled migration time to migrate the cloud service from the source cluster to a target cluster; migrating, based on the scheduled migration time, disk data associated with an original instance of the cloud service to a disk for servicing a new instance of the cloud service instantiated in the target cluster, the migration of the disk data being performed based on a migration priority order of the disk data; and configuring a data operation of the cloud service for a disk for servicing the original instance as a data operation for the disk for servicing the new instance.