Migration — Concept (DMS, AMS, MGN, DataSync, Storage Gateway, Snow, Application Discovery)
Why this group
SAA-C03 tests the standard migration toolbox. Memorize which tool moves what.
AWS Database Migration Service (DMS)
- Migrates databases with minimal downtime.
- Source and target can be different engines (e.g. Oracle → Aurora Postgres) using AWS Schema Conversion Tool (SCT) for schema.
- Supports one-time + ongoing CDC (Change Data Capture) replication.
- Sources: relational DBs, MongoDB, Kafka, S3.
- Targets: RDS / Aurora / Redshift / DynamoDB / S3 / Kinesis / OpenSearch / DocumentDB / Kafka.
- Multi-AZ replication instance possible.
AWS Application Migration Service (MGN, formerly CloudEndure)
- Lift-and-shift of on-prem / VM-based servers into AWS as EC2.
- Continuous block-level replication; cutover at chosen time.
- Best for migrating large fleets of physical / virtual servers.
AWS Server Migration Service (SMS)
- Older service; functionality merged into MGN. Don't pick SMS for new designs unless the exam item specifies it.
AWS DataSync
- Move files / objects between on-prem (NFS / SMB / HDFS / Object) and AWS (S3 / EFS / FSx).
- Or between AWS services (S3 ↔ EFS, cross-region S3).
- Scheduled or one-time; encrypted, includes data validation.
- Up to 10× faster than open-source tools.
AWS Storage Gateway (recap)
- Permanent on-prem service connecting NFS/SMB/iSCSI to S3 / EBS / FSx with local cache.
- Use for ongoing hybrid storage, not pure migration.
AWS Snow Family (recap)
- Physical devices for petabyte-scale offline data transfer + edge compute.
AWS Application Discovery Service
- Inventory & dependency of on-prem servers/applications.
- Agentless discovery (VMware vCenter) or Agent-based (Windows/Linux).
- Feeds Migration Hub for planning.
AWS Migration Hub
- Central dashboard for migration progress (DMS, MGN, Discovery, Refactor Spaces).
AWS Mainframe Modernization
- Modernize mainframe apps (COBOL, JCL) to AWS — niche, but exists.
Pick the right tool
| Source / scenario | Tool |
|---|---|
| Relational / NoSQL DB → AWS DB (different engine) | DMS (+ SCT) |
| Lift & shift on-prem VMs | MGN |
| Ongoing file replication (NFS/SMB) | DataSync |
| One-time petabyte transfer with poor network | Snowball / Snow Family |
| Continuous hybrid storage (apps still on-prem) | Storage Gateway |
| Inventory on-prem servers + dependencies | Application Discovery Service |
| Track progress across many tools | Migration Hub |
The 7 Rs of migration (commonly referenced)
- Retire — turn off
- Retain — keep on-prem
- Rehost ("lift & shift") — MGN
- Relocate — VMware Cloud on AWS
- Repurchase — switch to SaaS
- Replatform — small tweaks to use AWS managed services (e.g. lift & shift to RDS)
- Refactor — rebuild as microservices / serverless
Common exam scenarios
- "Migrate Oracle DB to Aurora Postgres with minimal downtime" → SCT for schema + DMS ongoing replication, then cutover.
- "Move 500 TB of files to S3, limited bandwidth" → Snowball Edge (or multiple).
- "Continuously sync on-prem NFS share to EFS" → DataSync.
- "Lift entire VM fleet from VMware to AWS" → MGN.
- "Inventory and analyze on-prem servers before migration" → Application Discovery Service.
- "Centralized dashboard for migration progress" → Migration Hub.
- "Ongoing NFS share to apps on-prem + cloud backup" → Storage Gateway (File).
Exam tip
- Database migrating → DMS (+ SCT if engine changes).
- VM lift & shift → MGN.
- File sync → DataSync.
- Bulk offline transfer → Snow Family.
- Permanent hybrid storage → Storage Gateway.