Course contentsShow
← Cloud and Infrastructure
Lesson 2533 of 2,98954. Serverless ArchitecturesFree lesson

DynamoDB Global Tables and Streams

Implement multi-region replication with global tables and process change events using DynamoDB Streams.

DynamoDB Global Tables and Streams

You'll learn how to replicate DynamoDB data across multiple AWS regions automatically and process change events in real-time using streams.

What Are Global Tables and Streams?

Global Tables enable multi-region, fully replicated DynamoDB tables. Think of them as synchronized copies of the same table living in different AWS regions—write to any region, and changes propagate automatically to all others. This builds on the cross-region replication and active-active patterns you've already learned.

DynamoDB Streams capture a time-ordered sequence of item-level changes (inserts, updates, deletes) in your tables. They're like a changelog that other services can read and react to—perfect for event-driven architectures and keeping global tables in sync.

How They Work Together

When you enable a global table, DynamoDB Streams work behind the scenes to replicate changes. Each regional table publishes its changes to a stream, which propagates updates to other regions. You can also consume these streams yourself with Lambda functions to trigger workflows, update caches, or send notifications.

Region A (us-east-1)     Region B (eu-west-1)
     Table                    Table
       |                        ↑
       | write                  | replicate
       ↓                        |
    Stream  -----------------> Stream
       |                        |
       ↓                        ↓
  Lambda (optional)       Lambda (optional)

Key Characteristics

  • Last-writer-wins conflict resolution for concurrent updates
  • Sub-second latency for cross-region replication
  • Stream retention: 24 hours of change records
  • Use streams for custom processing beyond replication (audit logs, analytics, downstream updates)

Key Takeaway: DynamoDB Global Tables provide active-active multi-region replication through streams, enabling low-latency global access while letting you react to every data change with event-driven serverless patterns.