For a multi-region application requiring shared state with low latency, which pattern is suitable?

For a multi-region application requiring shared state with low latency, which pattern is suitable?

• A. DynamoDB Global Tables for cross-region data replication and low-latency reads; or use S3 cross-region replication.
• B. RDS Cross-Region Read Replicas.
• C. Aurora Global Database only.
• D. CloudFront.

Answer: (A)

When you need a globally distributed app to share state with low latency, you want a data store that is replicated across regions and allows reads and writes in any region. DynamoDB Global Tables provides this exactly: a multi-region, fully replicated table model that lets your application read and write in the region closest to the user, while updates are replicated to other regions. This gives a single logical dataset with local access latency and automatic synchronization, which is ideal for keeping state consistent across geographies. Using S3 cross-region replication is useful for keeping copies of static assets, but S3 is object storage and not a fast, writable database for shared mutable state, so it isn’t ideal for maintaining low-latency global writes. RDS cross-region read replicas offer additional regions with read-only copies; they don’t support writes in those regions, so they cannot provide a truly shared, writable state globally. Aurora Global Database improves cross-region read performance but maintains a primary writer in one region, with replication to others, which isn’t as true a multi-master solution as Global Tables for global write access.

When you need a globally distributed app to share state with low latency, you want a data store that is replicated across regions and allows reads and writes in any region. DynamoDB Global Tables provides this exactly: a multi-region, fully replicated table model that lets your application read and write in the region closest to the user, while updates are replicated to other regions. This gives a single logical dataset with local access latency and automatic synchronization, which is ideal for keeping state consistent across geographies.

Using S3 cross-region replication is useful for keeping copies of static assets, but S3 is object storage and not a fast, writable database for shared mutable state, so it isn’t ideal for maintaining low-latency global writes. RDS cross-region read replicas offer additional regions with read-only copies; they don’t support writes in those regions, so they cannot provide a truly shared, writable state globally. Aurora Global Database improves cross-region read performance but maintains a primary writer in one region, with replication to others, which isn’t as true a multi-master solution as Global Tables for global write access.