Application & Database Migration Services
Migrating to AWS involves moving digital assets—data, workloads, and IT resources—from on-premises or other clouds into the AWS ecosystem. For the SAA-C03 exam, the focus is on selecting the right service to minimize downtime and ensure data integrity during the transition.
The Analogy: Moving to a New Smart Home
Imagine you are moving from an old, cluttered house to a modern smart home. AWS Application Migration Service (MGN) is like a professional moving crew that packs your furniture exactly as it is and places it in the new house (Lift & Shift). AWS Database Migration Service (DMS) is like a specialist who moves your high-tech home theater system; they don’t just move the boxes, they ensure the wiring is compatible with the new house’s electrical system, and if the plugs don’t match, they provide an adapter (Schema Conversion).
Core Concepts & The Well-Architected Framework
- Operational Excellence: Use AWS MGN to automate the conversion of physical, virtual, or cloud-based servers to run natively on AWS, reducing manual errors.
- Performance Efficiency: Use AWS DMS for minimal-downtime migrations. It allows the source database to remain operational during the transfer.
- Reliability: Implement Multi-AZ for DMS replication instances to ensure the migration process itself is highly available.
Service Comparison Table
| Service | Best For… | Key Feature | Downtime |
|---|---|---|---|
| AWS MGN | Server Migration (Rehost) | Block-level replication | Minimal (Cutover only) |
| AWS DMS | Database Migration | Supports CDC (Change Data Capture) | Near-Zero |
| AWS SCT | Heterogeneous DBs | Schema translation (e.g., Oracle to Aurora) | N/A (Tool) |
| AWS DataSync | Large scale file transfer | Automates data movement to S3/EFS/FSx | N/A (Background) |
Scenario-Based Decision Matrix
If/Then Strategy
- IF you need to migrate a SQL Server to Amazon RDS with zero downtime THEN use AWS DMS with Change Data Capture (CDC).
- IF you are migrating an Oracle DB to Amazon Aurora THEN use AWS SCT first to convert the schema, then AWS DMS to move the data.
- IF you have 100TB of data and limited internet bandwidth THEN use AWS Snowball Edge for the initial bulk transfer.
- IF you want to move 500 physical servers to EC2 with minimal reconfiguration THEN use AWS Application Migration Service (MGN).
Exam Tips: Golden Nuggets
- DMS Replication Instance: Remember that DMS requires a replication instance. You must size this instance based on the volume of data and the complexity of transformations.
- SCT vs. DMS: If the source and target database engines are DIFFERENT (e.g., SQL Server to MySQL), you MUST use the Schema Conversion Tool (SCT) before using DMS.
- MGN vs. SMS: AWS Server Migration Service (SMS) is deprecated in many contexts; Application Migration Service (MGN) is now the primary recommendation for rehosting.
- DMS is not just for AWS: You can use DMS to migrate data out of AWS or between two on-premises databases (though the replication instance must run on AWS).
Migration Architecture Visualized
From Source to Target: The Migration Pipeline
Key Services
- MGN: Rehost servers at scale.
- DMS: Migrate DBs with CDC.
- SCT: Convert DB schemas.
- DataSync: NFS/SMB/S3 sync.
Common Pitfalls
- Ignoring Network Latency.
- Skipping SCT for heterogeneous DBs.
- Not testing cutover windows.
- Under-provisioning DMS instances.
Quick Patterns
- Rehost: Use MGN (Fastest).
- Replatform: Use DMS to RDS.
- Relocate: VMware Cloud on AWS.
- Refactor: Rewrite for Lambda.