Improved high availability. If one region goes down, traffic can be routed to other
Disaster recovery and business continuity. If one region goes down, traffic can be routed to other in fast time. Companies has disaster recovery plan in that case.
Reduced latency.
Enhanced data redundancy and security. Improve mitigating of risk of data loss, sensitive data can be stored in specific region to comply region regulations.
Compliance and risk management. Hosting data in multiple regions helps to reduce risks associated with region specific legal and compliance issues.
Regulatory compliance and the need for data homing and encryption¶
User's data is located in home country.
Proof of Concept
❌ We can't replicate user data in another country. It means when country data center is failed, user can't login
❌ If user move to another country, for example from UK to Australia. His latency will increase because traffic must go form Australia to UK
❌ Encryption
This solution is raw yet
Challenges of mutlti region architecture relates with:¶
Handling different type of data
Metadata - non-personal data session data, tokens, data which is not contain identifiable information
Transient data - data which can contain personal information but is short-lived
Persistent data - data which can contain personal information and is long-lived
Choosing the right distributed database
AWS Aurora
Google Cloud Spanner
CockroachDB,
YugaByte
other
Sharding and replicating a database across regions