DMS is a fully managed service that enables migration and continuous data replication from on-premises, cloud, or hybrid databases to AWS with minimal downtime. It supports heterogeneous and homogeneous migrations.
Setting Up a Replication Instance
First, we must set up a replication instance. One advantage of AWS DMS is that it does not require a license—we only pay for the virtual machine (VM) used.
For production purposes, we can choose a multi-AZ (Availability Zone) deployment to enhance availability and fault tolerance.
Configuring the Source Endpoint
Next, we must create a source endpoint for AWS DMS to connect to. In my case, the source is located in an AWS VPC, but we can also use an on-premises server or another cloud provider.
Once AWS DMS successfully connects, the source endpoint status will change to Active. We must also create a target endpoint and ensure that both the source and target endpoints have an Active status.
To verify connectivity, we can use the Test Endpoint feature in AWS DMS.
Creating a Migration Task
After configuring the endpoints, we need to set up an AWS DMS migration task by selecting the source and target databases.
For the migration type, we can choose from:
- Full Load – Migrates all data in one go.
- Full Load and CDC (Change Data Capture) – Performs a full migration and continues replicating changes in real-time.
- CDC Only – Captures and replicates only the incremental changes. I used Full Load and CDC for real-time replication of delta changes. It is recommended to enable CloudWatch monitoring to track the performance of the replication task.
Filtering and Migrating Data
We can filter and specify which schemas or tables to migrate to AWS and which ones to exclude.
Once everything is set up, the migration task will start, and the data will begin replicating to the target database.
Top comments (0)