How Multi‑Active Database Architecture Is Redefining Bank Disaster Recovery

We all know that the core requirements for bank operations are data security and system stability; today, automation and intelligent operations are also essential. This article interviews Chen Chunsheng, a database expert at Huaxia Bank, who has twelve years of experience managing both traditional relational and open‑source databases.

Initially, the bank used storage‑level replication to build a two‑site, three‑center disaster‑recovery architecture, with a production center for read/write services and a disaster‑recovery center for cold backup. This approach left the backup resources idle and made it difficult to verify real‑time data availability.

To address these issues, the bank introduced logical and physical replication, creating a primary‑center read/write model and a backup‑center read‑only model, improving resource utilization and failover speed.

However, most multi‑center setups still assign unequal roles to primary and backup centers, causing brief service interruptions during failover and limiting consistency to eventual rather than real‑time. Huaxia Bank is now experimenting with a truly equal, multi‑active architecture where all data centers provide read/write services simultaneously, achieving real‑time data consistency across centers and eliminating the need for failover actions.

Chen explains that this equal, multi‑active model does not increase operational complexity; if one node fails, other nodes continue serving requests, reducing response time pressure on operators and better utilizing backup resources.

He also highlights the evolution of operations from manual scripts and real‑time monitoring to automated, intelligent workflows, including remote mobile app maintenance. The mobile app (DB Magic) enables engineers to diagnose, locate, and resolve database incidents from anywhere, crucial during emergencies like the COVID‑19 pandemic.

Security for mobile operations is emphasized: strict access controls, multi‑factor authentication, and encryption of critical data ensure that remote access does not compromise information security.

Huaxia Bank is building a data‑centered operations platform that aggregates hardware and software metrics, performs correlation analysis, and provides real‑time alerts and automated root‑cause analysis, further enhancing reliability.

Chen concludes that responsible, secure, and well‑documented operations are vital, and continuous learning and innovation remain essential as technology evolves.