AWS DMS

Database migration — migrate databases to AWS with minimal downtime using continuous data replication and optional schema conversion.

Overview

AWS Database Migration Service (DMS) migrates databases to AWS quickly and securely — the source database remains operational during migration, minimising downtime for applications. DMS supports homogeneous (e.g., Oracle to Oracle) and heterogeneous (e.g., Oracle to Aurora PostgreSQL) migrations.

Core Concepts

Concept	Description
Replication Instance	An EC2 instance that runs the migration tasks; sized based on workload
Source Endpoint	Connection to the source database (on-premises, EC2, RDS, or other cloud)
Target Endpoint	Connection to the target database (RDS, Aurora, DynamoDB, S3, Redshift, etc.)
Migration Task	Defines what data to migrate, the migration type, and table mappings
Full Load	One-time bulk copy of all existing data from source to target
Change Data Capture (CDC)	Continuous replication of ongoing changes after the full load
Table Mapping	Rules that select, filter, rename, or transform tables and columns during migration
Schema Conversion Tool (SCT)	Standalone tool that converts source database schema and code to target-compatible format
Premigration Assessment	Automated checks that identify potential issues before starting migration

How DMS Works

Source Database                Replication Instance              Target Database
(on-prem / EC2 / RDS)         (runs in a VPC)                   (RDS / Aurora / S3 / etc.)
       │                              │                                 │
       │── Full Load ────────────────▶│────────────────────────────────▶│
       │                              │                                 │
       │── CDC (ongoing changes) ────▶│────────────────────────────────▶│
       │                              │                                 │

Full Load — Bulk-copies all existing rows from source tables to the target
CDC — Reads the source database transaction log and continuously applies changes
Full Load + CDC — Most common; combines both for a complete migration with minimal downtime

Migration Types

Type	Description	Downtime
Full Load only	One-time copy of all data; no ongoing replication	Higher (cutover gap)
CDC only	Replicate only ongoing changes (assumes target already loaded)	Minimal
Full Load + CDC	Bulk copy then continuous replication until cutover	Minimal (recommended)

Homogeneous vs Heterogeneous Migration

Aspect	Homogeneous	Heterogeneous
Example	MySQL → Aurora MySQL	Oracle → Aurora PostgreSQL
Schema conversion	Not needed (same engine)	Required — use AWS SCT
DMS role	Data replication only	Data replication (after SCT schema convert)
Complexity	Lower	Higher — stored procedures, triggers, types

Supported Sources and Targets

Direction	Supported Engines
Sources	Oracle, SQL Server, MySQL, MariaDB, PostgreSQL, MongoDB, SAP ASE, IBM Db2, Azure SQL, S3
Targets	RDS (all engines), Aurora, Redshift, DynamoDB, S3, OpenSearch, Kinesis Data Streams, Kafka, DocumentDB, Neptune

AWS Schema Conversion Tool (SCT)

Converts source schema objects (tables, views, stored procedures, functions) to target-compatible format
Provides an assessment report showing conversion complexity and manual intervention items
Handles data type mapping, code translation, and object dependency resolution
Required for heterogeneous migrations; not needed for homogeneous

Common Use Cases

Cloud migration — Migrate on-premises databases to RDS or Aurora with minimal downtime using Full Load + CDC.
Database engine change — Move from commercial (Oracle, SQL Server) to open-source (PostgreSQL, MySQL) to reduce licensing costs.
Continuous replication — Keep an on-premises database and AWS target in sync during a migration window using CDC.
Data lake ingestion — Replicate database changes to S3 or Kinesis for analytics and data lake architectures.
Database consolidation — Merge multiple source databases into a single Aurora target.

SAA/SAP Exam Tips

SAA Tip: "Migrate database to AWS with minimal downtime" → DMS with Full Load + CDC. This is the standard answer for database migration questions.

Exam Trap: DMS handles data migration. For schema conversion in heterogeneous migrations (e.g., Oracle → PostgreSQL), the AWS Schema Conversion Tool (SCT) is required in addition to DMS.

SAP Tip: DMS can replicate to S3, Kinesis, or Kafka as targets — enabling real-time data lake ingestion from relational databases without building custom CDC pipelines.

Cross-Cloud Equivalents

Provider	Service / Solution	Notes
AWS	AWS DMS	Baseline
Azure	Azure Database Migration Service	Similar concept; supports offline + online modes
GCP	Database Migration Service	Focused on MySQL and PostgreSQL to Cloud SQL
On-Premises	Oracle GoldenGate, Attunity Replicate, Striim	Enterprise CDC and replication tools

Pricing Model

Dimension	Unit	Notes
Replication instance	Per hour	Based on instance class (e.g., dms.r5.large)
Data transfer	Per GB out	Standard AWS data transfer rates apply
Log storage	Per GB-month	Task logs stored on the replication instance

Amazon RDS — common migration target for relational databases
AWS DataSync and Snow Family — for file/storage migrations (not database-specific)
Amazon Kinesis and Flink — DMS can stream CDC to Kinesis for real-time analytics
Amazon S3 — DMS can replicate to S3 for data lake architectures

AWS DMS

On this page