Skip to main content
In this lesson/article we cover AWS Database Migration Service (DMS): a managed service designed to migrate and replicate databases and analytics workloads to AWS quickly, securely, and with minimal downtime or data loss. Think of DMS as the database-focused counterpart to AWS Application Migration Service: DMS specializes in migrating schemas, data, and ongoing changes between databases. DMS supports migrations across 20+ database and analytics engines, including MySQL, PostgreSQL, Oracle, SQL Server, MongoDB, Amazon Aurora, Amazon Redshift, and Amazon DocumentDB.

​
What is Database Migration Service?

AWS DMS helps you move data from a source database to a target database while minimizing downtime. It handles full data loads, ongoing replication via Change Data Capture (CDC), and can work alongside the AWS Schema Conversion Tool (SCT) for schema and code conversion during heterogeneous migrations (for example, Oracle → PostgreSQL). Key benefits:
  • Minimal downtime migrations using CDC
  • Support for homogeneous and heterogeneous migrations
  • Managed replication compute and monitoring
  • Integration with SCT for schema and SQL/code conversion

​
DMS Components

DMS has several components that together enable discovery, conversion, migration, and continuous replication.

​
DMS Fleet Advisor

DMS Fleet Advisor helps discover and analyze database inventory across your environment. It collects configuration and inventory information from database servers (for example, an on-premises MySQL server) using a data collector you install on the source systems. The collected metadata helps plan migrations and can be exported to CSV for further analysis. This discovery step is analogous to installing an agent to profile servers and applications before a migration.

​
DMS Schema Conversion Tool (SCT)

The AWS Schema Conversion Tool (SCT) assesses and converts database schemas and application SQL code when migrating between different database engines. SCT evaluates migration complexity, converts schemas, procedures, functions, and other database objects where possible, and lets you create custom type mappings. SCT is essential for heterogeneous migrations (for example, Oracle → PostgreSQL) where data types and SQL dialects differ.

​
Replication Instance and Endpoints

During migration DMS uses a replication instance to run replication tasks that perform extract, (optional) transform, and load operations. Endpoints anchor the source and target databases; each endpoint stores connection and authentication details the replication instance uses to access the database.
The image is a slide titled "DMS Components – Replication Instance" showing a DMS Endpoint icon and a row of endpoint info icons labeled Source or Target, DB Type, Server Name and port, Encryption, and Credentials. It visually outlines the key connection details needed for a database migration service endpoint.
Typical endpoint configuration items: Make sure the replication instance has network connectivity (VPC, security groups, routing) and the correct credentials to access each endpoint.
Always secure endpoint credentials and verify encryption settings. Incorrect network or permission configurations are the most common causes of migration failures.

​
Replication Tasks

A replication task ties together:
  • the replication instance that performs the work,
  • the source endpoint, and
  • the target endpoint.
When you create a replication task you select the replication instance, choose the source and target endpoints, and configure task settings such as table mappings, transformations, and logging. The task type determines whether DMS does an initial bulk load, continuous replication, or both.
A technical diagram titled "DMS Components – Replication Task" showing Source Endpoint, DMS Endpoint, a Task connecting to a Target Endpoint, and a Replication Instance. The replication instance lists three task types: Full Load, Full load + CDC, and CDC Only.

​
Migration Types

DMS supports three primary migration task types:
  • Full load
    Perform a one-time bulk copy of the dataset. DMS migrates table data and can create tables in the target if needed.
  • Full load + CDC (Change Data Capture)
    Perform an initial full load and capture ongoing changes simultaneously. After the full load completes, the captured changes are applied to the target so it remains synchronized with the source.
  • CDC only
    Use when bulk data has already been transferred by another method (for example, native export/import tools) and you only need ongoing change replication until cutover.
CDC stands for Change Data Capture. CDC captures and replicates data changes (inserts, updates, deletes) from the source so the target stays in sync with minimal downtime.

​
Migration Patterns

DMS supports:
  • Homogeneous migrations — same engine to same engine (for example, Oracle → Oracle).
  • Heterogeneous migrations — different engines (for example, Oracle → PostgreSQL or MongoDB → Amazon DocumentDB). Heterogeneous migrations typically involve SCT for schema and code conversion.
Using CDC with DMS enables near-continuous replication so you can test, validate, and cut over with minimal interruption.

​
Schema Conversion Example

SCT can automatically convert schemas and let you define custom mappings for types and database objects. For example, SCT can map Oracle VARCHAR2 to PostgreSQL VARCHAR, or rework data warehouse schemas for Amazon Redshift targets. Use SCT to reduce manual schema and code changes in heterogeneous migrations.
A diagram showing schema conversion from a corporate on‑premises Oracle instance through a central Schema Conversion Tool (SCT) to cloud targets. The AWS Cloud side lists an Oracle instance and Amazon RedShift as destinations.

​
Terminology recap

​
Use Cases

DMS is commonly used for:
  1. Migrating data — Move on‑premises or legacy databases to AWS-managed databases, with tools to convert schema and migrate data.
  2. CDC/replication — Continuously replicate changes between databases for minimal-downtime migrations, reporting replicas, or cross-region replication.
  3. Data lake ingestion — Stream or periodically load database changes into data lakes for analytics and real-time processing.
A slide titled "Use Cases" showing three cards labeled 01 Migrate Data, 02 CDC, and 03 Data Lake. Each card has a purple circular icon: cloud/server migration, a gear with sync arrows, and stacked servers over water.

​
Summary

AWS Database Migration Service provides a managed framework to discover, convert, migrate, and continuously replicate databases to AWS with minimal downtime. Core components include:
  • DMS Fleet Advisor (discovery and inventory),
  • Schema Conversion Tool (SCT) for heterogeneous migrations,
  • Replication instances that run tasks,
  • Endpoints that define database connections, and
  • Replication tasks for full load and CDC-based migrations.

Watch Video