This chapter covers AWS Migration Services, a suite of tools and services designed to help organizations move their existing workloads—servers, databases, applications, and data—to the AWS Cloud. For the CLF-C02 exam, this objective falls under Domain 3: Cloud Technology Services, and while it's not the heaviest weighted domain, migration questions appear regularly, often as scenario-based multiple choice. Understanding the key services, their use cases, and the order of operations in a typical migration is critical to scoring well. By the end of this chapter, you'll be able to identify the right migration tool for a given scenario and explain the high-level migration process.
Jump to a section
Imagine your company is moving to a new office building. You have filing cabinets full of documents, servers humming in the basement, and employees who need to keep working during the move. You can't just shut down for a week. AWS Migration Services are like a professional moving crew that handles everything: they pack your files (data), label boxes (metadata), transport them in a secure truck (encrypted transfer), and unpack them exactly where they belong in the new office (target environment). Some items are small and can be moved over the internet (AWS DataSync), while large, heavy items like the mainframe need a physical truck (AWS Snowball). The moving crew also makes sure your phone system and internet work at the new location before you arrive (AWS Migration Hub tracks progress). And if you want to keep your old office running as a backup during the move, they can synchronize changes continuously (AWS DMS for database replication). The key mechanism: the movers don't just copy files; they validate each box, update your inventory (catalog), and provide a real-time dashboard so you know exactly what has been moved and what's still in transit. This mirrors how AWS Migration Services provide discovery, assessment, replication, and cutover planning—all from a single pane of glass.
What Are AWS Migration Services and What Problem Do They Solve?
AWS Migration Services are a collection of tools that help you move your on-premises workloads to AWS with minimal downtime, reduced risk, and automated processes. Before cloud migration, organizations had to manually reinstall operating systems, copy data over slow networks, and reconfigure applications—a process that could take months and often resulted in data loss or configuration errors. AWS provides a structured approach: assess, mobilize, migrate, and modernize.
The problem these services solve is threefold: (1) Discovery – you need to know exactly what you have (servers, databases, dependencies). (2) Replication – you need to copy data and applications without disrupting production. (3) Cutover – you need to switch traffic to the new environment cleanly.
How AWS Migration Services Work – The Mechanism
AWS Migration Services operate on a replication-based model. For server migrations, AWS Application Migration Service (MGN) installs an agent on the source server. This agent continuously replicates the server's disks to a staging area in AWS (typically an Amazon EBS volume in a temporary EC2 instance). When you're ready to cut over, the service takes a final sync, launches the target EC2 instance from the replicated volumes, and you update DNS or routing to point to the new server.
For databases, AWS Database Migration Service (DMS) uses a change data capture (CDC) mechanism. DMS reads the source database's transaction logs and applies changes to the target database in near real-time. This allows you to keep the source database running during migration and switch over with minimal downtime.
For large-scale data transfers (terabytes to petabytes), AWS Snow Family provides physical devices. You request a Snowball Edge device from AWS, load your data onto it, and ship it back. AWS then imports the data into Amazon S3.
Key Services, Tiers, and Configurations
AWS Migration Hub: A central tracking dashboard. It doesn't perform migrations itself but provides visibility into the progress of migrations using other tools like MGN or DMS. It also integrates with AWS Application Discovery Service to collect data about on-premises servers.
AWS Application Migration Service (MGN): The recommended service for lift-and-shift migrations of physical or virtual servers. It supports Windows and Linux, and can migrate from any source (VMware, Hyper-V, physical). Pricing is based on the number of source servers and the duration of replication.
AWS Database Migration Service (DMS): Supports homogeneous migrations (e.g., Oracle to Oracle) and heterogeneous migrations (e.g., Oracle to Amazon Aurora). It can also be used for continuous data replication for disaster recovery. DMS offers a free tier for the first 750 hours of instance usage per month (t2.micro).
AWS DataSync: An online data transfer service that simplifies moving large amounts of data between on-premises storage and AWS. It uses a software agent installed on your on-premises server and can schedule transfers, validate checksums, and encrypt data in transit.
AWS Snow Family: Snowcone (8 TB), Snowball Edge (80 TB), and Snowmobile (100 PB). Used when network transfer is impractical due to bandwidth or time constraints.
AWS Transfer Family: Managed file transfer protocols (SFTP, FTPS, FTP) for moving files directly into and out of Amazon S3 or Amazon EFS.
Comparison to On-Premises or Competing Approaches
On-premises migration typically involves manual scripting, rsync for data, and manual configuration of new servers. This is error-prone and slow. Competing cloud providers offer similar services (e.g., Azure Migrate, Google Cloud Migration). AWS's advantage is tight integration with its ecosystem—for example, DMS can replicate directly to Amazon Redshift, and MGN can automate post-migration optimization like right-sizing instances.
When to Use Each Service vs Alternatives
Use MGN when migrating hundreds of servers with minimal changes (lift-and-shift).
Use DMS when migrating databases, especially if you need minimal downtime or want to change database engines.
Use DataSync for one-time or recurring file transfers to S3 or EFS over the network.
Use Snow Family when transferring more than 10 TB over a slow internet connection, or when security policies forbid online transfer.
Use Transfer Family if you have legacy applications that only support FTP/SFTP and need to integrate with S3.
Exam tip: CLF-C02 often tests on the difference between online (DataSync, DMS, MGN) and offline (Snow Family) transfer methods. Also, remember that Migration Hub is a tracking tool, not a migration engine.
Discover and Assess
Use AWS Application Discovery Service to inventory your on-premises servers, including their specifications, dependencies, and performance metrics. You can deploy AWS Agentless Discovery Connector (for VMware) or install the AWS Application Discovery Agent on each server. The data is sent to AWS Migration Hub, where you can view a consolidated inventory. This step is crucial for planning: you'll know which servers are candidates for migration, their CPU/memory utilization, and inter-server dependencies. Without discovery, you risk missing a critical server or migrating a server that is underutilized.
Select Migration Strategy
Based on the assessment, choose a migration strategy: rehost (lift-and-shift) using MGN, replatform (lift and optimize) using MGN with instance right-sizing, or refactor (re-architect) using native AWS services. For databases, decide between homogeneous (same engine) or heterogeneous (different engine) migration using DMS. For data files, decide between online (DataSync) or offline (Snow Family). This step is often tested on the exam as scenario-based questions where you must recommend the appropriate service.
Set Up Replication
For server migration using MGN, install the MGN agent on each source server. The agent replicates the server's disks to a staging area in AWS (temporary EC2 instances and EBS volumes). For database migration using DMS, create a replication instance (an EC2 instance running DMS software), then define source and target endpoints. DMS then loads a full copy of the database and begins CDC. For data transfer using DataSync, deploy the DataSync agent (virtual appliance) on-premises, configure a task, and start the transfer. During replication, AWS charges for the staging resources (e.g., EBS storage for replicated disks).
Test and Validate
Before cutover, test the migrated environment. For MGN, you can launch a test instance from the replicated disks without affecting the source. For DMS, you can run validation tasks to compare source and target data. For DataSync, you can verify checksums. This step ensures that the migrated workload functions correctly and that there are no data inconsistencies. AWS recommends testing multiple times, especially for critical applications. The exam may ask about the purpose of test instances in MGN.
Cut Over and Go Live
When ready, perform the final sync: for MGN, stop replication, take a final snapshot, and launch the target EC2 instance. For DMS, stop CDC and switch applications to the target database. For DataSync, the final transfer is already complete. Then update DNS records or load balancer targets to point to the new AWS resources. Monitor the application to ensure it works. After a stable period, decommission the source servers. AWS Migration Hub provides a dashboard to track cutover status across all migrations.
Scenario 1: Enterprise Data Center Migration A financial services company is migrating 500 virtual machines from a VMware cluster to AWS. They use AWS Application Discovery Service to inventory servers and map dependencies. They then use AWS Application Migration Service (MGN) to replicate servers continuously over two weeks. During cutover weekend, they perform a final sync, launch EC2 instances (right-sized using MGN's recommendations), and update DNS. The migration completes with 15 minutes of downtime per server. Cost considerations: They pay for MGN per server per month during replication, plus the staging EBS volumes. They save on data center power and cooling costs post-migration.
Scenario 2: Database Migration with Minimal Downtime An e-commerce company is migrating a 2 TB Oracle database to Amazon Aurora MySQL. They use AWS DMS with CDC. First, DMS performs a full load (takes 8 hours). Then it continuously replicates changes from Oracle's redo logs. After a week of replication, they schedule a maintenance window: stop the application, apply final CDC changes (5 minutes), and switch the connection string to Aurora. The migration is successful with less than 10 minutes of downtime. What goes wrong if not properly configured: If the source database's retention period for redo logs is too short, DMS might lose changes and require a full reload.
Scenario 3: Large Data Transfer to S3 A media company has 50 TB of video archives on a NAS that needs to be moved to Amazon S3. Their internet connection is 100 Mbps, making online transfer take over 50 days. They use AWS Snowball Edge: request a device, copy data over NFS, ship it to AWS, and AWS imports it into S3. Total time: 2 weeks (including shipping). Cost: Snowball Edge usage fee plus shipping. Misconfiguration: If data is not encrypted at rest on the Snowball, it might be a security risk; AWS encrypts devices automatically, but the customer must manage the encryption keys.
What CLF-C02 Tests on This Objective Domain 3 (Cloud Technology Services), Objective 3.6: 'Identify AWS services for migration.' The exam tests your ability to match migration scenarios to the correct AWS service. You will NOT be asked to configure these services in depth. Instead, expect multiple-choice questions that present a business requirement (e.g., 'A company wants to move 10 TB of data over a slow internet connection') and ask which service to use.
Common Wrong Answers and Why 1. Choosing AWS DataSync for a 50 TB transfer over a 50 Mbps link – Candidates see 'large data' and think DataSync, but DataSync is online and would take weeks. The correct answer is Snowball (offline). 2. Selecting AWS Migration Hub as a migration tool – Candidates confuse it with MGN. Migration Hub is a tracking dashboard, not a replication tool. 3. Using AWS DMS for server migration – DMS is for databases only. Server migration requires MGN. 4. Confusing AWS Application Discovery Service with MGN – Discovery only collects data; MGN performs the actual migration.
Specific Terms That Appear on the Exam - AWS Application Migration Service (MGN) – the recommended lift-and-shift tool. - AWS Database Migration Service (DMS) – for database migrations, supports homogeneous and heterogeneous. - AWS DataSync – online data transfer to/from S3 and EFS. - AWS Snowball Edge – offline data transfer device. - AWS Migration Hub – central tracking dashboard.
Tricky Distinctions - Online vs Offline: If the scenario mentions 'slow network' or 'large dataset' (>10 TB), lean toward Snow Family. If 'fast network' or 'real-time replication', choose DataSync or DMS. - Server vs Database: If the question says 'migrate a SQL Server database', answer DMS. If it says 'migrate a Windows server', answer MGN. - One-time vs Continuous: DataSync can be one-time or scheduled; DMS can be one-time or continuous (CDC).
Decision Rule for Multiple Choice Read the scenario and identify the workload type (server, database, file). Then check the network bandwidth and data size. If it's a server, pick MGN. If it's a database, pick DMS. If it's files and network is fast, pick DataSync; if network is slow or data is huge, pick Snow Family. Eliminate any option that says 'Migration Hub' if the question asks for a migration tool.
AWS Migration Hub is a central dashboard for tracking migrations; it does not perform migrations.
AWS Application Migration Service (MGN) is the primary service for lift-and-shift server migrations.
AWS Database Migration Service (DMS) supports both homogeneous and heterogeneous database migrations with minimal downtime via CDC.
AWS DataSync is for online file transfers to/from S3 and EFS; it can be scheduled and validates checksums.
AWS Snow Family (Snowcone, Snowball Edge, Snowmobile) is for offline data transfer when network transfer is impractical.
The typical migration process is: discover (Application Discovery Service), track (Migration Hub), replicate (MGN/DMS/DataSync), cut over.
For CLF-C02, focus on matching the scenario to the correct service, not on configuration details.
These come up on the exam all the time. Here's how to tell them apart.
AWS DataSync
Online transfer over network
Best for datasets < 10 TB or fast network
Can schedule recurring transfers
Requires DataSync agent on-premises
Lower cost for small data, no shipping
AWS Snowball Edge
Offline transfer via physical device
Best for datasets > 10 TB or slow/unreliable network
One-time transfer (or multiple shipments)
No agent needed, just copy to device
Higher cost due to device usage and shipping
Mistake
AWS Migration Hub actually performs the migration of servers and databases.
Correct
Migration Hub is a tracking and monitoring tool only. It does not migrate anything. It integrates with other services like MGN and DMS to show progress.
Mistake
AWS DataSync can only transfer data from on-premises to AWS, not the other way around.
Correct
DataSync supports bidirectional transfers. You can move data from AWS to on-premises as well.
Mistake
AWS DMS requires the source database to be shut down during migration.
Correct
DMS supports live migrations with minimal downtime using change data capture (CDC). The source can remain operational during the full load and CDC phases.
Mistake
AWS Snowball is used for small data transfers (under 1 TB).
Correct
Snowball is designed for large data transfers (terabytes to petabytes). For small transfers, online methods like DataSync or AWS CLI are more cost-effective.
Mistake
AWS Application Migration Service can only migrate VMware VMs.
Correct
MGN supports migration from any source, including physical servers, Hyper-V, and other hypervisors, as long as the source OS is supported.
MGN is for migrating entire servers (OS and applications) to AWS EC2 instances. DMS is specifically for migrating databases (e.g., Oracle, SQL Server, MySQL) to AWS database services like Amazon RDS or Aurora. MGN uses block-level replication of server disks; DMS uses logical replication of database transactions. On the exam, if the question mentions 'server' or 'VM', choose MGN; if it mentions 'database', choose DMS.
Use DataSync when you have a fast and reliable network connection and the data size is manageable (typically under 10 TB). DataSync can also be used for ongoing, scheduled transfers. Use Snowball when the network is slow, unreliable, or expensive, or when the dataset is very large (tens of TB or more). Snowball is also useful if you need to transfer data from a location without internet access. Exam tip: look for keywords like 'slow network', 'limited bandwidth', or 'large dataset' to choose Snowball.
No. AWS Migration Hub is a tracking and monitoring service. It provides a single place to discover existing servers, plan migrations, and track the progress of migrations performed by other AWS services like MGN, DMS, and DataSync. It does not replicate or transfer any data itself. Many exam candidates mistakenly think Migration Hub is a migration tool—it is not.
Yes, DMS supports both homogeneous migrations (same engine, e.g., Oracle to Oracle) and heterogeneous migrations (different engine, e.g., Oracle to Amazon Aurora). For heterogeneous migrations, DMS uses built-in schema conversion from AWS Schema Conversion Tool (SCT) to convert the source schema to the target format. This is a key exam point: DMS handles the data movement, while SCT handles schema conversion.
The AWS Snow Family includes Snowcone (8 TB), Snowball Edge (80 TB), and Snowmobile (100 PB). Snowcone is for small edge computing and data transfer. Snowball Edge is the most common for large data migration; it also offers compute capabilities. Snowmobile is a shipping container pulled by a truck for exabyte-scale transfers. On the exam, know that Snowball Edge is the typical choice for multi-terabyte migrations.
MGN installs an agent on the source server. The agent continuously replicates disk changes to a staging area in AWS (temporary EC2 instances with EBS volumes). When you're ready to cut over, MGN stops replication, takes a final snapshot, and launches a fully provisioned EC2 instance from the replicated disks. You then update DNS or routing to point to the new instance. The source server remains online until cutover, minimizing downtime.
AWS Application Discovery Service helps you inventory your on-premises servers, including their configurations, performance metrics, and dependencies. It can be agentless (for VMware) or agent-based. The data is used to plan migrations, right-size instances, and identify dependencies. It integrates with Migration Hub. On the exam, know that Discovery is for assessment, not migration.
You've just covered AWS Migration Services — now see how well it sticks with free CLF-C02 practice questions. Full explanations included, no account needed.
Done with this chapter?