Resilience, Migration, and Business Continuity · Topic 1 of 6
Migration Assessment, Rehosting, and Data Migration
Plan AWS migrations at scale: discovery, dependency mapping, migration waves, rehosting with AWS Transform MGN, and data moves with AWS DMS and DataSync.
- 5 min
- Discovery and dependency mapping come before wave planning; waves keep chatty components together.
- AWS Transform MGN is the rehost answer: block-level replication, test launch, short cutover.
- Rehosting is a fast first step; you can replatform or refactor after the move.
What it is
A large migration starts with knowing what you have. Discovery builds an inventory of servers, applications and the network traffic between them, and portfolio assessment decides what to do with each application.
Rehosting (lift and shift) then moves servers to Amazon EC2 with as little change as possible. Databases and file data usually move on their own track, with services built for each kind of data.
How it works
Discovery and portfolio assessment
AWS Transform uses AI agents to help with the discovery, planning and execution phases of a migration. For server migrations it collects and analyzes on-premises inventory from multiple sources, groups applications, identifies dependencies and builds prioritized migration waves.
Dependencies decide the waves. Servers that talk to each other constantly, such as an application tier and its database, should move in the same wave, or the traffic between them will cross the hybrid link and add latency.
Each application gets one of the "7 Rs" migration strategies.
| Strategy | What happens | Typical trigger |
|---|---|---|
| Retire | Switch it off | Unused or duplicated application |
| Retain | Leave it on premises for now | Recent upgrade, hard dependency, compliance |
| Rehost | Move servers as they are to EC2 | Data center exit deadline |
| Relocate | Move a hypervisor environment without changing it | Large VMware estate, fast exit |
| Repurchase | Move to a SaaS product | Commodity application such as CRM |
| Replatform | Small changes to use managed services | Self-managed database to Amazon RDS |
| Refactor | Re-architect for cloud native design | Scaling or agility limits |
Wave planning
A good plan starts with a small pilot wave of low-risk applications to prove the tooling, network and runbooks. Later waves grow in size and criticality.
AWS Transform can also design a multi-account landing zone and translate the source network configuration into an Amazon VPC architecture. That matters because the target accounts and networks must exist before the first wave cuts over.
Rehosting with AWS Transform MGN
AWS Transform MGN, previously known as AWS Application Migration Service, migrates physical, virtual and cloud servers to AWS. An agent on each source server performs continuous block-level replication into AWS, and MGN converts the servers to boot natively on EC2.
Three templates (replication, launch and post-launch) control how servers replicate, launch and are configured after launch. Servers can be grouped into applications and applications into waves, so launch, cutover and archive actions run in bulk.
Because replication keeps running until cutover, the cutover window is typically minutes. You launch test instances first, validate them, then do a final cutover when the business agrees.
AWS Transform can orchestrate the rehost step itself, with guided replication, testing and cutover workflows, as part of a larger migration job.
Database and data migration
| Need | Service | Key point |
|---|---|---|
| Move a database with minimal downtime | AWS DMS | Full load, then change data capture (CDC) until cutover |
| Change database engine | AWS DMS with DMS Schema Conversion or AWS SCT | Convert schema and code first, then move data |
| Move files or objects online | AWS DataSync | Agent based, scheduled, verifies transferred data |
| Keep file access on premises with cloud storage | AWS Storage Gateway | Hybrid access, not a one-time move |
| Bring partner file transfers to AWS | AWS Transfer Family | Managed SFTP, FTPS and FTP into S3 or EFS |
| Upload large data sets from your own devices | AWS Data Transfer Terminal | Physical location with a high speed connection to AWS |
AWS DMS keeps the source database online while it copies existing data and then replicates ongoing changes. Cutover becomes a short switch of the connection string after replication lag reaches near zero.
A heterogeneous migration, such as Oracle to Amazon Aurora PostgreSQL, needs schema and stored code converted first. DMS Schema Conversion or AWS SCT handles that, and DMS moves the data.
AWS DataSync copies data from NFS, SMB and other storage systems into Amazon S3, Amazon EFS or Amazon FSx. It can run on a schedule, which suits an initial bulk copy followed by incremental syncs before cutover.
How the exam asks about this
- "Hundreds of servers must move before a data center lease ends, with minimal change and cutover windows of minutes." -> AWS Transform MGN with continuous replication, test launches, then cutover by wave.
- "An Oracle database must move to Aurora PostgreSQL with minimal downtime." -> Convert the schema with DMS Schema Conversion or AWS SCT, then AWS DMS full load plus CDC.
- "Applications must be grouped so tightly coupled servers move together." -> Discovery with dependency mapping in AWS Transform, then wave planning based on those groups.
What to remember for the exam
- Discovery and dependency mapping come before wave planning; waves keep chatty components together.
- AWS Transform MGN is the rehost answer: block-level replication, test launch, short cutover.
- Rehosting is a fast first step; you can replatform or refactor after the move.
- AWS DMS full load plus CDC gives minimal downtime database moves.
- Engine changes need schema conversion (DMS Schema Conversion or AWS SCT) before the data moves.
- DataSync moves file data online; Storage Gateway is for ongoing hybrid access.
Do not choose a one-time backup and restore when the question says "minimal downtime" for a database. Only continuous replication (DMS CDC, or MGN for whole servers) keeps the cutover window short.
Common mistakes
Migrating by server instead of by application. Moving half of an application leaves the other half calling across the hybrid link, which adds latency and breaks during the cutover.
Using MGN to change the database engine. MGN copies whole servers block by block, so the database engine stays the same. An engine change is a DMS plus schema conversion job.
Starting with the most critical application. The first wave should prove the process on low-risk workloads, so problems are found before they hit revenue systems.
Check yourself
Why does a rehost with AWS Transform MGN have a short cutover window?
Replication is continuous and block level, so at cutover only the final changes need to sync before the EC2 instance launches.
What does an engine change from Oracle to Aurora PostgreSQL need besides AWS DMS?
Schema and code conversion with DMS Schema Conversion or AWS SCT before the data is moved.
How should dependency data shape migration waves?
Servers that communicate heavily belong in the same wave so their traffic does not cross the hybrid link after cutover.
Topic 1 of 6 in Resilience, Migration, and Business Continuity
Test yourself on Resilience, Migration, and Business Continuity
The full AWS SAP-C03 question bank has 850 questions, with timed practice exams and an explanation for every answer.
This study guide is independent and is not affiliated with, endorsed by, or sponsored by AWS. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.