AWS Database Migration Service

AWS Database Migration Service is a managed service for migrating and replicating data between homogeneous and heterogeneous database engines and data stores. It operates using customer-provisioned replication instances and migration tasks to perform initial full load and optionally ongoing change data capture where supported. It typically sits between source databases (on-premises or cloud) and target data stores in AWS as a migration and replication layer.

🗂 Resource Category

Migration and Transfer • Database

🧠 Exam Memory Hook

Think: "Source database + replication tasks + VPC replication instance = AWS Database Migration Service"

📖 Ownership

Classification: AWS Managed Service

AWS responsibilities: AWS operates and maintains the DMS control plane, the underlying physical infrastructure, the hypervisor layer, and the managed replication instance platform and service software; AWS also provides the DMS APIs, console, and automated maintenance for the managed service platform. AWS secures the service control plane and the infrastructure that hosts the managed replication instances, and provides integration points for logging and metrics. AWS does not manage the customer's source or target databases, their data, or database schemas.

Customer responsibilities: The customer configures replication instances, source and target endpoints, migration tasks, table mappings, and network access; the customer is responsible for data, database schema, access control for source and target databases, test validation, and any application changes needed after migration. The customer secures network connectivity (VPC, security groups, VPN/Direct Connect), configures IAM permissions for DMS actions, monitors migration tasks and CloudWatch metrics, and maintains any migration tooling they run outside DMS (for example on EC2 or on-premises).

Patching responsibilities: AWS patches the physical infrastructure, the hypervisor, and the managed DMS service platform including the managed replication instance runtime and DMS service software. The customer patches guest operating systems and runtimes only when they run migration or connectivity components on their own hosts (for example Amazon EC2 or on-premises servers). The customer is responsible for patching application libraries, drivers, dependencies, and any migration scripts or tools they deploy themselves.

🏗 Typical Architecture

💡 Top 5 Features

Performs full-load plus optional ongoing change data capture (CDC) replication for supported sources and targets.
Runs managed replication instances in the customer VPC to process migration tasks and apply changes.
Supports heterogeneous and homogeneous migrations with configurable table mappings and limited transformations within tasks.
Integrates with Amazon CloudWatch for logging and metrics and AWS Identity and Access Management for access control.
Allows migration to object storage targets such as Amazon S3 and to many database engines as supported endpoints.

✅ Top 5 Use Cases

One-time migration of an on-premises relational database to Amazon RDS using a managed replication instance to perform the copy.
Continuous replication from a production database to a read-only target for reporting or analytics with CDC enabled.
Consolidating multiple legacy databases into a single cloud-hosted database by migrating selected schemas and tables.
Migrating between different database engines where schema conversion is performed separately and DMS performs data movement.
Bootstrapping a data lake by migrating transactional data into Amazon S3 for downstream analytics.

🏗 Architecture Placement

DMS sits in the migration/replication layer inside a customer VPC (when using managed replication instances) and connects to source databases over network links such as VPN or Direct Connect and to AWS targets such as Amazon RDS or Amazon S3. Clients and on-premises databases are the typical sources that send data to DMS endpoints, while DMS sends data to configured target endpoints. DMS is a regional service and replication instances are deployed into the customer's VPC subnets when launched. Configuration, monitoring, and task management occur at the account and regional level.

🎯 Commonly Used With

- Amazon RDS
- Amazon S3
- Amazon EC2
- Amazon VPC
- AWS Identity and Access Management

🌍 5 Real-World Examples

A bank migrating customer account data from an on-premises Oracle database to Amazon RDS for PostgreSQL to standardize operations and reduce on-premises footprint.
A healthcare provider continuously replicating transactional updates to Amazon S3 for analytics pipelines without impacting the primary OLTP system.
A retail company consolidating several regional MySQL databases into a single Amazon RDS instance to simplify reporting and operations.
A telecommunications operator using DMS to replicate billing records into Amazon RDS while performing a phased cutover to the cloud.
A government education department bootstrapping a central analytics store by migrating historical student records from legacy databases into S3 for analysis.

🎓 AWS Exam Clues

Use DMS when you need managed data migration or change-data-capture replication between supported source and target endpoints.
Consider network access and reachability because DMS replication instances must access source and target databases over the network.
DMS handles data movement but not complex schema conversion; use the AWS Schema Conversion Tool for heterogeneous schema tasks.
Choose DMS when you require a managed replication instance in your VPC rather than building custom replication on EC2.
Monitor DMS tasks and replication instance resource metrics in CloudWatch to identify performance bottlenecks and task failures.

📝 Quick Revision

AWS Database Migration Service moves data between supported source and target stores using managed replication instances and tasks for full load and optional CDC. Use it when you need managed migration or continuous replication, and plan for schema conversion, network connectivity, endpoint permissions, monitoring, and task tuning.

🏷 Keywords

change data capture • replication instance • replication task • source endpoint • target endpoint • table mapping • full load • ongoing replication • schema conversion • AWS SCT • CloudWatch metrics • VPC placement