GBase 8s Enables Same-City Active-Active Disaster Recovery

Published on 2020-05-12


 

Customer Requirements

As business continues to evolve across industries, customers' existing systems can no longer meet demand, requiring the construction of a second data center in the same city. The new secondary data center must combine performance gains and enhanced security—it must handle part of the workload to offload the primary data center, and it must enable mutual data backup between the two city sites to achieve same-city active-active disaster recovery.

 

Solution

The database is the foundation of the entire same-city active-active architecture. Database deployment focuses on efficiency, integrity, and stability to provide complete and accurate data sources for the upper-layer applications. A cluster must be configured in each data center to eliminate single points of failure. The databases in the two centers run independently, each serving its own database workload. Bidirectional data synchronization between the two centers enables same-city active-active disaster recovery.
 

The HDR+ER cluster deployment solution of the GBase 8s transactional database easily enables same-city active-active disaster recovery, providing robust support for applications. This solution deploys GBase 8s on four servers: the HDR high-availability cluster provides mutual failover between two nodes, while ER enterprise replication shares specific data across different services.

 

 


 

HDR and ER are both built-in components of GBase 8s, offering simple configuration and proven reliability.
 

HDR (High-Availability Data Replication) is a built-in disaster recovery technology within the GBase 8s engine that requires no separate purchase. It continuously synchronizes the primary and standby servers by replicating logical logs. When the primary server fails, the standby server can take over application connections, ensuring business continuity and delivering highly reliable database services.
 

ER (Enterprise Replication) is a built-in, asynchronous, log-based mechanism that distributes changes to specific tables and rows at the transaction level across any number of participating nodes. ER can operate across different platforms, and even different minor versions of GBase 8s, offering more flexibility for new database builds without requiring upgrades to the original system. Replication begins after the user transaction is committed, using an asynchronous approach that tolerates network latency. ER supports multiple data models, such as update-anywhere, consolidation, dissemination, and workload partitioning. These models can be freely mixed within a single replication system to implement one-to-one, one-to-many, many-to-one, or many-to-many unidirectional or bidirectional replication.

 

Key Benefits

Low cost: The database system in the existing data center requires no changes. Once the new data center database is built, simply configuring data replication between the two centers completes the same-city active-active setup. This protects existing hardware and software investments while offering excellent scalability. Both HDR and ER components are included and require no separate purchase. HDR does not require shared storage, saving hardware resources.
 

High Availability: The HDR cluster high-availability solution automatically fails over to another node when one node encounters an issue—no manual intervention required—ensuring continuous and secure business operations.
 

Easy Development: The ER enterprise replication feature synchronizes shared data between different applications in real time, reducing application development effort and system complexity.

 

Use Case

This solution has been applied in real-world projects, including a city branch of China Mobile and a provincial branch of China Telecom, effectively addressing customers’ needs for workload distribution and same-city active-active disaster recovery during the construction of a second data center.