Server-Side Cache Push for Consistent Cloud Data Sync
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud databases face performance degradation when pushing data changes to multiple clients due to sequential processing of operations, which affects data consistency.
Innovation Solution
Store operation types and data identifiers in a cache space, allowing the server to determine data changes and push relevant information to clients based on current database states, reducing reliance on direct database queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the cloud database sequentially pushes data changes to each client based on request execution sequence, then data consistency is ensured, but database performance is greatly affected
Solution Approach 1:
The patent introduces a message queue as an intermediary component between the cloud database and clients. The message queue receives data change notifications from the database and manages client subscriptions, acting as a buffer that decouples the database from direct client communication. This allows the database to write changes asynchronously without being blocked by sequential push operations to multiple clients, thus maintaining data consistency while improving database performance.
Solution Approach 2:
The patent segments the data push function by separating it into two independent parts: (1) the cloud database that focuses on data processing and change detection, and (2) the message queue that handles the distribution of changes to clients. This segmentation allows each component to operate independently at its own pace, with the database not being burdened by the overhead of managing multiple client connections and push sequences, thereby resolving the contradiction between consistency and performance.
2Reliability
If the cloud database pushes changes to all clients after each operation, then data consistency is maintained, but system overhead increases
Solution Approach 1:
The patent implements a subscription-based mechanism where clients actively register their interests in specific data types or records with the message queue. Instead of the system pushing to all clients unconditionally, only clients who have subscribed to relevant data changes receive notifications. This self-service approach allows clients to control their own information flow, reducing unnecessary system overhead while ensuring that all interested parties receive consistent data updates.
Solution Approach 2:
The patent applies partial action by implementing selective data pushing based on client subscriptions. Rather than pushing all data changes to all clients (excessive action), the system pushes only the specific changes that subscribed clients are interested in (partial action). This optimization maintains data consistency for subscribed clients while significantly reducing system overhead by avoiding unnecessary push operations to clients whose data has not changed or who are not interested in the changes.
Data Source
AI summary
An information pushing method includes a server side that stores, in a target cache space after a cloud database successfully performs an operation, a type of the operation and an identifier of data associated with the operation. Then, when pushing to a client, the server side pushes related information of the data to the client based on the identifier of the data, the type of the operation, and whether the data exists in the cloud database currently, where the related information of the data is used by the client to determine a change between the data in the cloud database currently and that before the operation is performed such that an impact on performance of the cloud database can be reduced, and the client can obtain data information consistent with that in the cloud database.