The "three data centres" architecture for backup and disaster recovery.
That is, three data centres are deployed in two geographical locations, usually including: Intra-city active-active centre and remote disaster recovery centre, which is the key infrastructure to ensure enterprise business continuity and data security. Here's how necessary it is from multiple dimensions:
---
1. Respond to different levels of disaster risk
- Local-level disasters (e.g., data centre fires, power failures, network outages):
The same-city double-living center (usually 30-100 kilometers away) can realize real-time data synchronization and rapid service switching, ensuring that business is not interrupted in the event of local failures.
- Regional disasters (such as earthquakes, floods, wars):
The off-site disaster recovery center (more than hundreds of kilometers away) stores key data by asynchronously copying and in the event that extreme disasters cause the same city center to be paralyzed, core business can still be restored through the off-site center.
Case: The 2011 Japan earthquake caused a large-scale downtime in Tokyo and surrounding data centers, and off-site disaster recovery centers became the best choice for enterprises to resume operations.
---
2. Achieve business continuity (RTO/RPO optimization)
- RTO (recovery time target):
The same-city dual-activity activity can realize second-level switching (such as financial trading systems), and the disaster recovery in other places is usually restored at the hourly level to meet the needs of different business priorities.
- RPO (data recovery point target):
The same-city center achieves RPO≈0 (zero data loss) through synchronous replication, and the asynchronous replication of the off-site center tolerate minute-level data differences, balancing costs and risks.
Technical support: Graded protection is achieved based on technologies such as storage-level replication (such as synchronous SAN), database log synchronization (such as Oracle Data Guard).
---
3. Avoid single point of failure and systemic risks
- Hardware/software failure:
The multi-center architecture eliminates the risk of single point of failure in a single data center in power supply, network, storage and other links.
- Human operation risk:
Logical errors (such as accidentally deleting data) can be quickly rolled back through the off-site backup version to avoid damage to the entire disk data.
-Supply Chain Risk:
Dispersed geographical locations to avoid complete paralysis due to supplier service interruptions (such as regional failures of cloud service providers).
Case: Amazon Cloud Services (AWS) failed in the eastern U.S. region in 2021, and many global enterprises that rely on a single region were out of service for hours.
---
4. Meet compliance and industry regulatory requirements
- Financial Industry:
The People's Bank of China's "Specifications on the Power System of Computer Rooms of Financial Information Systems" clearly requires "double active life in the same city + disaster recovery in other places".
- Medical Industry:
The HIPAA Act requires that patient data must be recoverable offsite.
- Multinational corporations:
Data sovereignty regulations such as GDPR may force data to be retained in local areas, and a multi-center architecture can meet compliance in multiple places.
---
5. Support business expansion and technological innovation**
- Flow peak response:
The Double Live Center can achieve load balancing and alleviate the pressure of sudden access to e-commerce promotions and government affairs systems.
- Hybrid cloud architecture integration:
Off-site centers can be deployed as private clouds or public clouds to form hybrid cloud disaster recovery (such as Alibaba Cloud hybrid cloud disaster recovery solution).
-New technology test site:
The disaster recovery center can undertake tasks such as grayscale release and disaster recovery drills, without affecting the stability of the production environment.
---
6. Economic Benefit Analysis
- Direct cost:
The initial construction cost is high, but the investment in traditional self-built data centers can be reduced through cloud disaster recovery (such as DRaaS).
- Risk cost:
According to Gartner statistics, the average loss of a company's key system downtime by one hour is US$300,000, far exceeding the investment in disaster recovery.
- Insurance discount:
A complete disaster recovery system can reduce corporate commercial insurance premiums (such as property insurance and network security insurance).
---
Typical architecture comparison
plan | RTO | RPO | cost | Applicable scenarios |
Local backup | Hours | Hours | Low | Non-critical business |
Two lives in the same city | Seconds | 0 | Medium-high | Financial core system |
Two places and three centers | Grade recovery | Grade protection | high | Government, energy and other key areas |
---
Summarize
The "two places and three centers" have become the core strategy for enterprises to resist systemic risks in the digital era through the triple guarantee of spatial dispersion, data redundancy and business activity. Especially under the trend of intelligence driven by cloud computing and AI, its value has shifted from a "cost center" to a "core component of business empowerment", which is an indispensable cornerstone for enterprise digital transformation.