Data Center Read Request Forwarding for Disaster Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a multi-data center disaster recovery system, data synchronization is often asynchronous, leading to situations where data written in one data center cannot be read from another, reducing reading and writing reliability due to potential unsynchronized data across centers.
Innovation Solution
A method where a first data center forwards data reading requests to a second data center if the user identification is found in a writing user identification set, ensuring that data written is processed and synchronized in the second data center, thereby allowing for accurate and reliable data reading across centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is written in one data center with asynchronous synchronization to another data center, then data writing speed is improved, but data reading reliability deteriorates because written data may not be accessible in other data centers
Solution Approach 1:
The system performs preliminary actions by maintaining a writing user identification set that tracks users who have written data, and proactively determines whether reading requests should be forwarded to other data centers before the actual read operation occurs. This ensures data availability without waiting for asynchronous synchronization to complete.
Solution Approach 2:
The patent introduces an intermediary mechanism (the writing user identification set and forwarding logic) between the data writing operation and data reading operation. This intermediary tracks writing operations and mediates reading requests by determining whether to forward them to other data centers, thus ensuring data consistency without requiring direct synchronization between all data centers.
2Reliability
If all data reading requests are forwarded to the data center where data is written, then data reading reliability is improved, but network bandwidth pressure increases
Solution Approach 1:
The system applies local quality by making reading requests local (processing in the current data center) when possible, and only forwarding requests to other data centers when necessary (when the user identification is in the writing user identification set). This selective forwarding approach minimizes network bandwidth consumption while maintaining data reading reliability.
Solution Approach 2:
Instead of forwarding all reading requests or none, the system performs partial action by forwarding only those reading requests where the user identification exists in the writing user identification set. This partial forwarding approach optimizes the balance between reliability and network bandwidth usage.
3Stability of the object's composition
If data synchronization is performed between multiple data centers, then data consistency is improved, but system complexity increases
Solution Approach 1:
The patent extracts the essential information needed for data consistency (writing user identification) and separates it from the actual data synchronization process. By tracking only user identifications rather than synchronizing complete data sets, the system maintains data consistency while significantly reducing system complexity and synchronization overhead.
Solution Approach 2:
The system changes the parameter being synchronized from complete data sets to simplified user identification markers. This parameter change allows the system to maintain data consistency through a much simpler mechanism, reducing system complexity while achieving the same consistency goal.
Data Source
AI summary
The application provides a method for data reading and writing processing, data center, disaster recovery system and storage medium. The method includes receiving, by a first data center, a data reading request including a first user identification; determining, in response to the data reading request, whether a writing user identification set stored in the first data center includes the first user identification, the writing user identification set including user identifications corresponding to data writing requests; and under a condition that the writing user identification set includes the first user identification, forwarding the data reading request to a second data center, to read, in the second data center, data to be read as indicated by the data reading request, the second data center being configured to receive and process a data writing request sent by a user terminal or forwarded by the first data center.