Study guide
Technical reference and lesson notes
Purpose of This Lesson
AWS provides several migration and transfer services for moving workloads, databases, and data between environments. Although these services can support migrations between AWS services, the most common exam scenario involves an organization moving existing on-premises infrastructure into AWS.
This lesson introduces the major service categories and the decision points used to select an appropriate migration tool. The concepts are primarily theoretical because practical implementation usually requires access to both an AWS environment and an on-premises or hybrid network.
Key Concepts
Migration is broader than moving into AWS
Migration services can be used for several patterns:
- Moving workloads from on-premises data centers to AWS
- Moving databases between different database platforms or environments
- Moving virtual machines into Amazon EC2
- Transferring large quantities of files or objects into or out of AWS
- Moving data between AWS services as part of modernization or restructuring
For the SAA-C03 exam, identify the type of workload being moved before selecting a service.
Database migration
AWS Database Migration Service (AWS DMS) is designed to migrate databases with minimal downtime. It supports migrations involving databases running on-premises, in Amazon EC2, or in managed AWS database services.
A common migration pattern is:
- Configure a source endpoint for the existing database.
- Configure a target endpoint for the destination database.
- Use a replication instance to perform the migration.
- Perform a full load of existing data.
- Continue replicating changes using change data capture (CDC).
- Cut over applications to the target database when synchronization is complete.
DMS is especially relevant when the source and target engines differ, such as migrating from a commercial database to an AWS-managed database engine. Schema and code conversion may require AWS Schema Conversion Tool (AWS SCT) before or alongside the data migration.
Server and virtual machine migration
The course terminology refers to AWS Server Migration Service (AWS SMS) for migrating virtual machines to Amazon EC2. AWS SMS was designed to replicate on-premises virtual machines as Amazon Machine Images (AMIs), which could then be launched as EC2 instances.
For current AWS implementations, AWS Application Migration Service (AWS MGN) is the primary lift-and-shift service for many server migration scenarios. It continuously replicates source servers and helps launch them in AWS with minimal changes.
On an exam question, focus on the migration requirement:
- If the requirement is to migrate a server or virtual machine with minimal modification, choose a server migration or application migration service.
- If the requirement is to redesign or refactor the application, a lift-and-shift service alone is not sufficient.
Online data transfer
AWS DataSync is used to move file-based data between on-premises storage and AWS storage services, or between supported AWS storage locations. It is appropriate when an organization has network connectivity and wants managed, accelerated, and automated data transfer.
Typical destinations include services such as:
- Amazon S3
- Amazon EFS
- Amazon FSx
DataSync is a strong choice when the scenario emphasizes recurring transfers, synchronization, scheduling, verification, or online movement over a network connection.
Offline bulk data transfer
AWS Snowball devices provide physical data transfer when moving data over the network would take too long, cost too much, or be impractical. Data is copied to the device, the device is shipped to AWS, and AWS imports the data into the selected destination.
Snowball is most appropriate when:
- The dataset is very large.
- Network bandwidth is limited.
- The migration has a fixed deadline.
- A one-time bulk transfer is preferable to a long-running online copy.
A useful exam comparison is the time required to transfer the data over the available network versus the time required to order, load, ship, and ingest a Snowball device.
Architecture Decision Guide
| Migration requirement | Likely service or approach | Main reason |
|---|---|---|
| Migrate a database while keeping downtime low | AWS DMS | Performs full load and ongoing change replication |
| Convert schemas or database code between engines | AWS SCT, often used with AWS DMS | Helps convert incompatible schemas and database objects |
| Move on-premises virtual machines to EC2 with minimal changes | AWS Application Migration Service; older course material may reference AWS SMS | Supports lift-and-shift server migration |
| Transfer file data repeatedly over an existing network connection | AWS DataSync | Managed, scheduled, and accelerated online transfer |
| Move a very large dataset when network transfer is too slow | AWS Snowball | Physical device-based data transfer |
| Move data between AWS storage services | DataSync or a service-specific migration method | Selection depends on storage type and transfer pattern |
Exam-Relevant Takeaways
- Start with the data or workload type: database, virtual machine, file system, or bulk dataset.
- AWS DMS is the database migration service and supports ongoing replication for reduced downtime.
- AWS SCT addresses schema and code conversion; DMS moves and replicates the data.
- AWS DataSync is an online transfer service for file-based data and recurring synchronization.
- AWS Snowball is designed for large offline data transfers when network migration is impractical.
- A server migration service is generally associated with lift-and-shift, not application modernization.
- Hybrid connectivity, such as a VPN or Direct Connect connection, may be required for online migration from an on-premises environment.
- The best answer often depends on minimizing downtime, minimizing network usage, or minimizing application changes.
Common Exam Traps
- Confusing DMS with DataSync: DMS migrates databases and understands database replication. DataSync transfers files and objects; it is not a database replication engine.
- Using Snowball for a small, recurring transfer: Snowball is optimized for large physical transfers, not routine low-volume synchronization.
- Assuming DMS converts all database code: DMS moves data and replicates changes. Use AWS SCT when schema or code conversion is required.
- Treating server migration as modernization: Replicating a virtual machine into EC2 does not automatically redesign the application for cloud-native operation.
- Ignoring downtime requirements: A one-time export/import may be insufficient when the business requires continuous operation and minimal downtime.
- Ignoring connectivity: Online migration services require suitable network access between the source environment and AWS.
- Selecting a service only because it is an AWS migration service: The service must match the workload and migration method described in the question.
Real-World Engineer Notes
- Create a migration plan that separates discovery, replication, testing, cutover, and rollback.
- Estimate transfer time using the dataset size and effective throughput, not the theoretical link speed.
- For databases, test application compatibility, latency, indexes, stored procedures, and character-set behavior before cutover.
- For server migrations, validate boot behavior, drivers, networking, licensing, security groups, IAM roles, and application dependencies.
- Use encryption and least-privilege access for migration endpoints and temporary storage.
- Treat migration as a hybrid architecture problem: routing, DNS, firewall rules, identity, monitoring, and operational ownership all matter.
- A successful initial copy is not the same as a successful production migration. Rehearse cutover and rollback procedures.
Quick Reference Summary
- AWS DMS: Database migration and ongoing change replication.
- AWS SCT: Schema and database-code conversion, commonly paired with DMS.
- AWS SMS: Older course terminology for virtual machine migration; AWS Application Migration Service is the current service to evaluate for many server lift-and-shift scenarios.
- AWS DataSync: Online, managed transfer and synchronization for file-based data.
- AWS Snowball: Offline physical transfer for large datasets when network transfer is impractical.
- Core decision: Match the service to the workload, transfer method, downtime target, and required degree of application change.
Flashcards
- Q: What is AWS DMS primarily used for?
A: Migrating and replicating databases between supported source and target environments.
- Q: What feature allows AWS DMS to reduce application downtime during migration?
A: Change data capture, which continuously replicates changes after the initial full load.
- Q: What is AWS SCT used for?
A: Converting database schemas and related code when moving between database engines.
- Q: What type of data is AWS DataSync designed to transfer?
A: File-based data between supported on-premises and AWS storage locations, or between AWS storage services.
- Q: When is Snowball preferable to an online transfer?
A: When the dataset is very large or network bandwidth makes online transfer too slow, expensive, or unreliable.
- Q: What migration strategy is typically associated with server migration services?
A: Lift-and-shift, where servers are moved with minimal application modification.
- Q: Why might an online migration require a VPN or Direct Connect connection?
A: The source environment must be able to communicate securely with AWS resources and migration endpoints.
- Q: What is the first decision when selecting a migration service?
A: Identify whether the workload is a database, virtual machine, file dataset, or large bulk dataset.
- Q: Does AWS DMS automatically modernize an application?
A: No. It migrates database data and supports replication; application redesign is a separate effort.
Practice Questions
Question 1
A company operates a business-critical database in its data center. It wants to migrate to an AWS database with only a short outage during final cutover. The source database must remain available while the initial data load is performed. Which solution best meets the requirement?
A. Copy database files to Amazon S3 using AWS DataSync and restore them later
B. Use AWS DMS with a full load followed by change data capture
C. Use AWS Snowball to transfer the database backup
D. Create an AMI of the database server and launch it in EC2
Correct answer: B
Explanation: AWS DMS can perform an initial full load and continue replicating changes with CDC, allowing the source database to remain operational until cutover. DataSync and Snowball are data-transfer tools, while an AMI migration does not provide database-aware replication.
Question 2
An organization must move 500 TB of file data from an on-premises file server to Amazon S3. Its WAN connection is heavily utilized, and transferring the data online would take several months. Which approach is most appropriate for the initial transfer?
A. AWS Snowball
B. AWS DMS
C. AWS SCT
D. Amazon EC2 instance store
Correct answer: A
Explanation: Snowball is designed for large offline transfers when network bandwidth makes online migration impractical. DMS and SCT are database-focused, and instance store is not a migration service or durable transfer destination.
Question 3
A company wants to move files from an on-premises storage system to Amazon EFS every night and verify that changes are synchronized. The company has reliable network connectivity to AWS. Which service should it use?
A. AWS DataSync
B. AWS Snowball
C. AWS DMS
D. AWS Application Migration Service
Correct answer: A
Explanation: DataSync supports managed online transfers, scheduling, synchronization, and verification for supported file-storage workflows. Snowball is more suitable for large offline transfers, while DMS and Application Migration Service target different workload types.
Question 4
A company wants to move several on-premises virtual machines to Amazon EC2 without redesigning the applications. Which migration approach best matches this requirement?
A. Use a database schema conversion tool
B. Use a server or application migration service for lift-and-shift replication
C. Use AWS DMS to replicate the virtual machine disks
D. Use DataSync to synchronize the operating system files to S3
Correct answer: B
Explanation: Server migration services are designed to replicate workloads into EC2 with minimal changes. This is a lift-and-shift strategy. DMS is for databases, SCT is for database conversion, and DataSync is not a complete virtual machine migration solution.