Database Disaster Recovery Technologies for the Financial Sector

Databases are core to information systems, and financial institutions rely heavily on disaster‑recovery (DR) and backup solutions to ensure business continuity.

With the growing diversity of database workloads and the shift to cloud environments, DR strategies in finance have evolved. This article focuses on the technical aspects of database DR in the financial sector, covering definitions, classifications, current practices, market conditions, and mainstream DR architectures.

1. DR and Backup Overview Disaster recovery (DR) refers to systematic, pre‑planned measures that keep an information system operational during a disaster, while backup is the process of copying data to prevent loss. DR ensures continuity; backup provides a restore point.

2. History and Definitions DR originated in the 1970s in the United States; the first DR center was built by SunGard in 1979. Over time, DR evolved from simple data and system backups to comprehensive disaster‑recovery planning (DRP) that includes business‑level recovery objectives such as RTO and RPO.

3. DR Classification DR can be classified by distance (local vs. remote) and by protection level (data‑level, application‑level, business‑level).

4. Backup Classification Backups are categorized by data volume (full, incremental, differential), frequency (scheduled vs. real‑time), and granularity (byte‑level, block‑level, file‑level).

5. Difference Between DR and Backup Backup is the foundation of DR; DR replicates production changes to a standby site, which may be synchronous or asynchronous, but cannot replace a proper backup strategy.

6. DR Metrics Reliability of a DR solution is measured by Recovery Time Objective (RTO) and Recovery Point Objective (RPO).

Financial Institution DR Status and Requirements Banks, securities firms, and other financial entities demand zero data loss and high availability, employing architectures such as two‑site three‑center, active‑passive, or active‑active setups, with technologies like dense‑wavelength division multiplexing (DWDM) for real‑time data sync.

Key requirements include massive data backup, real‑time replication, cross‑platform migration, high‑availability of primary‑secondary applications, intelligent operation and maintenance, and automated large‑scale DR validation.

Market and Trend Overview The Chinese DR market grew from ¥4.98 billion in 2010 to nearly ¥18 billion in 2018, projected to exceed ¥30 billion by 2022. Gartner predicts that by 2021, 50 % of enterprises will manage long‑term data with backup rather than archiving.

DR technologies are expanding from local storage replication to multi‑site, cloud‑based solutions, covering host, OS, database, file, and network replication, as well as data‑masking, big‑data management, and read/write separation.

Distributed Database DR Architectures

(1) Single‑Center DR – Deploys multiple instances across availability zones within a single data center for high availability, but cannot survive a site‑wide disaster.

(2) Intra‑City Active‑Passive (Mutual Backup) – Mirrors the production environment in a standby city‑level site; the standby remains idle until a disaster triggers a switchover.

(3) Intra‑City Active‑Active (Dual‑Active) – Both production and standby sites serve live traffic simultaneously, providing load balancing and automatic failover with strong data consistency.

(4) Two‑Site Three‑Center – Adds a remote third center to the active‑active or active‑passive setup, offering protection against regional catastrophes; typically uses asynchronous replication.

These architectures are essential for maintaining continuous financial services.

Source: China Academy of Information and Communications Technology & Baidu Netcom (2021).