Cross-region replication (Global Tables) in DynamoDB - Step-by-Step Execution

Concept Flow - Cross-region replication (Global Tables)

Create Global Table

↓

Write Data in Region A

↓

Replication Service Detects Change

↓

Replicate Data to Region B

↓

Data Available in Region B

↓

Read/Write in Any Region

↓

Conflict Resolution if Needed

↓

Consistent Data Across Regions

Data written in one region is automatically copied to other regions, allowing reads and writes anywhere with conflict resolution to keep data consistent.

Execution Sample

DynamoDB

1. Create Global Table with Region A and B
2. Write item {id:1, name:'Alice'} in Region A
3. Replication service copies item to Region B
4. Read item from Region B returns {id:1, name:'Alice'}

Shows how writing data in one region replicates it to another region automatically.

Execution Table

Step	Action	Region A Data	Region B Data	Replication Status	Notes
1	Create Global Table	{}	{}	Not started	Table created in both regions, empty data
2	Write item {id:1, name:'Alice'} in Region A	{id:1, name:'Alice'}	{}	Pending	Data written only in Region A
3	Replication service detects change	{id:1, name:'Alice'}	{}	In progress	Change detected, replication started
4	Replicate data to Region B	{id:1, name:'Alice'}	{id:1, name:'Alice'}	Completed	Data replicated to Region B
5	Read item from Region B	{id:1, name:'Alice'}	{id:1, name:'Alice'}	Completed	Data available and consistent in Region B
6	Write item {id:1, name:'Alice Smith'} in Region B	{id:1, name:'Alice'}	{id:1, name:'Alice Smith'}	Pending	Update made in Region B
7	Replication service detects change	{id:1, name:'Alice'}	{id:1, name:'Alice Smith'}	In progress	Change detected, replicating back to Region A
8	Conflict resolution if needed	{id:1, name:'Alice Smith'}	{id:1, name:'Alice Smith'}	Completed	Conflict resolved, data consistent
9	Final data in both regions	{id:1, name:'Alice Smith'}	{id:1, name:'Alice Smith'}	Completed	Data synchronized across regions

💡 Replication completes and data is consistent across all regions

Variable Tracker

Variable	Start	After Step 2	After Step 4	After Step 6	After Step 8	Final
Region A Data	{}	{id:1, name:'Alice'}	{id:1, name:'Alice'}	{id:1, name:'Alice'}	{id:1, name:'Alice Smith'}	{id:1, name:'Alice Smith'}
Region B Data	{}	{}	{id:1, name:'Alice'}	{id:1, name:'Alice Smith'}	{id:1, name:'Alice Smith'}	{id:1, name:'Alice Smith'}
Replication Status	Not started	Pending	Completed	Pending	Completed	Completed

Key Moments - 3 Insights

Why does Region B data remain empty immediately after writing in Region A (Step 2)?

Because replication happens asynchronously, the data is first written only in Region A. The replication service detects the change next (Step 3) before copying it to Region B (Step 4).

How is data consistency maintained when updates happen in both regions?

The replication service detects changes in each region and uses conflict resolution rules to synchronize data, as shown in Steps 6 to 8 where an update in Region B is replicated back to Region A and conflicts are resolved.

What happens if replication is delayed or interrupted?

Data in regions may temporarily differ, but the replication service retries until data is consistent, ensuring eventual consistency as seen by the 'Replication Status' moving from 'Pending' to 'Completed'.

Visual Quiz - 3 Questions

Look at the execution_table at Step 4. What is the replication status?

A Pending

B Completed

C Not started

D In progress

💡 Hint

Check the 'Replication Status' column at Step 4 in the execution_table.

At which step does Region B first have the item {id:1, name:'Alice'}?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Look at the 'Region B Data' column in the execution_table to see when data appears.

If the replication service failed after Step 3, what would be the Region B data at Step 5?

A {} (empty)

B {id:1, name:'Alice'}

C {id:1, name:'Alice Smith'}

D Replication status would be 'Completed'

💡 Hint

Refer to the replication status and Region B data columns in the execution_table for Step 5.

Concept Snapshot

Cross-region replication (Global Tables) in DynamoDB:
- Create a global table spanning multiple regions.
- Writes in any region replicate asynchronously to others.
- Reads can be served locally from any region.
- Conflict resolution ensures data consistency.
- Enables low-latency global access and disaster recovery.

Full Transcript

Cross-region replication with DynamoDB Global Tables means when you write data in one region, it automatically copies to other regions. This lets you read and write data anywhere with low delay. The replication service detects changes and copies them asynchronously. If updates happen in multiple regions, conflict resolution keeps data consistent. The execution table shows steps from creating the table, writing data in Region A, replicating to Region B, and updating data in Region B with replication back to Region A. Variables track data states in each region and replication status. Key moments clarify why data appears later in Region B and how conflicts are resolved. The quiz tests understanding of replication timing and data consistency. This helps build globally available, fault-tolerant applications.