The 7 Rs of AWS Migration: Strategies, Services, and Exam Decisions – SAP-C02 Study Guide

Purpose of This Lesson

The 7 Rs provide a framework for choosing how each workload should move to AWS. The strategies range from eliminating a workload entirely to redesigning it as a cloud-native application.

The correct choice depends on business value, migration timeline, application complexity, modernization goals, operational risk, and available engineering capacity. On AWS exams, the key is to match the migration requirement to the strategy that provides the required outcome with the least unnecessary effort.

Key Concepts

The 7 Rs at a Glance

Strategy	Meaning	Typical AWS outcome	Relative effort
Retire	Decommission an application that is no longer needed	Shut down the workload and remove associated resources	Lowest
Retain	Keep the workload in its current environment for now	Continue operating on premises and reassess later	Very low
Relocate	Move a workload without changing its architecture or application	Move an entire environment or platform with minimal modification	Low
Rehost	Move the operating system and application to a new host	Migrate a server to Amazon EC2	Low to moderate
Repurchase	Replace the existing application with a different product or SaaS service	Replace a self-managed CRM with a SaaS CRM	Moderate to high
Replatform	Make limited changes to use managed AWS services	Move a database to Amazon RDS or an application to Elastic Beanstalk	Moderate
Refactor	Redesign the application to use cloud-native architecture	Use serverless, containers, managed databases, queues, and event-driven services	Highest

Some AWS material uses re-architect instead of refactor. In this context, both describe substantial redesign rather than a simple server move.

Retire

Retire workloads that provide little or no business value. Examples include obsolete applications, duplicate systems, unused development environments, and applications whose functionality has been replaced.

Retirement reduces migration scope, operating cost, licensing exposure, and security risk. It should be an intentional decision supported by dependency analysis and business-owner approval, not simply an assumption that an inactive-looking server is unused.

Retain

Retain means leaving a workload where it is temporarily or permanently. Common reasons include:

• The application is not a current migration priority.
• Existing hardware or software licenses still have useful life.
• The workload has unresolved dependencies.
• The application is stable and does not justify migration effort.
• Regulatory, latency, or business requirements favor keeping it on premises.

Retained workloads should be recorded in the migration portfolio with an owner and a future review date. Retain is often a deliberate deferral, not a failure to create a migration plan.

Relocate

Relocate moves a workload with little or no architectural change. The destination may be another data center, a hosted environment, or a platform designed to move an existing environment as a unit.

The main distinction from rehost is the migration unit:

• Relocate: move an existing environment or platform with minimal change.
• Rehost: move the operating system and application to a different host, such as EC2.

Networking, DNS, identity, storage connections, and monitoring may still require configuration changes even when the application itself is unchanged.

Rehost: Lift and Shift

Rehosting moves the existing operating system and application to new infrastructure. AWS commonly describes this as lift and shift. A typical target is an Amazon EC2 instance.

Rehosting is useful when the organization needs to:

• Exit a data center quickly.
• Minimize application changes.
• Reduce migration risk.
• Move workloads before a later modernization phase.
• Preserve behavior while gaining AWS infrastructure flexibility.

Rehosting is not always literally zero change. Drivers, agents, boot configuration, networking, monitoring, licensing, and integrations may need adjustment for the target environment.

#### AWS Application Migration Service

AWS Application Migration Service (AWS MGN) is the primary AWS service for lift-and-shift migrations to EC2. It continuously replicates source servers into AWS, allowing the migration team to perform testing before a final cutover.

A typical workflow is:

1. Install the replication agent on supported source servers, or use the appropriate VMware integration method.
2. Continuously replicate disks and changes to AWS.
3. Launch test instances in a staging environment.
4. Validate the application, connectivity, security groups, and operations.
5. Perform a final synchronization.
6. Launch the cutover instances and redirect users or traffic.

This approach minimizes downtime because the source environment can remain active during most of the replication and testing period.

Other migration utilities can support specific scenarios, including VM Import/Export. AWS Server Migration Service appears in older training material; AWS MGN is the current service to associate with most AWS server lift-and-shift questions.

Repurchase

Repurchase replaces an existing application with another product, often a SaaS service. Examples include replacing a self-managed customer relationship management system with a SaaS CRM or moving collaboration data to a managed cloud product.

Repurchase can reduce infrastructure operations and maintenance, but data migration, identity integration, customization, process changes, and vendor lock-in must be evaluated. A move between substantially different products may require significant transformation and integration work even though the strategy is labeled repurchase.

Replatform

Replatforming makes targeted changes to obtain cloud benefits without redesigning the entire application. The application’s business behavior remains largely the same, but selected components move to managed services.

Examples include:

• Moving a MySQL database to Amazon RDS for MySQL or Amazon Aurora.
• Moving an application to AWS Elastic Beanstalk.
• Replacing self-managed storage or middleware with managed AWS equivalents.
• Updating deployment and configuration processes to fit the target platform.

Replatforming typically requires more work than rehosting because schemas, drivers, connection strings, deployment processes, and operational procedures may change.

#### AWS Database Migration Services

AWS Database Migration Service (AWS DMS) supports migrations and ongoing replication between many database engines. It can help reduce downtime by keeping the target synchronized while the source remains operational.

When the source and target database engines differ, AWS Schema Conversion Tool (AWS SCT) can analyze schemas and assist with converting database objects and, where supported, application SQL code. DMS moves data; SCT helps address structural and code compatibility. Neither tool eliminates the need for testing and manual remediation.

Refactor or Re-architect

Refactoring redesigns an application to use cloud-native capabilities. Potential target components include:

• AWS Lambda or containers instead of long-running servers.
• Amazon Aurora, Amazon DynamoDB, or other managed databases instead of self-managed database servers.
• Amazon S3 or Amazon EFS instead of local file storage, where the access pattern is appropriate.
• Amazon SQS, Amazon SNS, and Amazon EventBridge for decoupling and event-driven workflows.
• AWS Step Functions for workflow orchestration.

This strategy can improve scalability, resilience, elasticity, deployment velocity, and operational efficiency. However, the primary cost is engineering effort: application decomposition, data-model changes, API redesign, testing, observability, security redesign, and operational retraining.

Refactoring is appropriate when long-term business value justifies substantial development work. It is usually a poor choice when the immediate objective is a rapid data-center exit with minimal application risk.

Exam-Relevant Takeaways

• Retire means the workload is no longer required; do not migrate it.
• Retain means defer migration or keep the workload where it is for a justified reason.
• Relocate moves an environment or platform with minimal changes.
• Rehost is AWS’s standard lift-and-shift strategy, commonly targeting EC2.
• Repurchase replaces an application, often with SaaS.
• Replatform introduces limited changes to use managed services such as Amazon RDS, Aurora, or Elastic Beanstalk.
• Refactor/re-architect involves significant redesign, commonly using serverless, containers, managed databases, and event-driven services.
• For server migrations to EC2 with minimal application changes, consider AWS MGN.
• For heterogeneous database migrations, consider AWS DMS plus AWS SCT.
• Continuous replication and a final cutover can reduce downtime compared with a one-time offline migration.
• The most modern architecture is not automatically the correct answer. Select the strategy that satisfies the stated business and technical constraints.

Architecture Decision Guide

Requirement or constraint	Best-fit strategy	Reasoning
Application is obsolete or duplicated	Retire	Avoid spending migration effort on a workload with no continuing value
Migration is not currently justified	Retain	Defer while documenting ownership, dependencies, and a review date
Entire environment must move with minimal modification	Relocate	Preserve the existing platform or environment structure
Fast data-center exit with minimal application change	Rehost	Move servers to EC2, commonly using AWS MGN
Organization wants to stop managing a packaged application	Repurchase	Adopt a replacement product or SaaS service
Application can remain mostly intact but database or platform operations should be managed	Replatform	Move selected components to services such as RDS, Aurora, or Elastic Beanstalk
Long-term scalability and agility justify major development	Refactor	Redesign around cloud-native and event-driven services

Common Exam Traps

• Confusing relocate and rehost: Relocate generally moves a platform or environment as a unit; rehost moves the workload to a new host, often EC2.
• Treating replatform as a complete rewrite: Replatforming makes focused changes. A major decomposition into serverless services is refactoring.
• Assuming lift and shift means zero changes: Source-server agents, drivers, networking, monitoring, licensing, and integrations may require updates.
• Choosing refactoring for a rapid migration: Refactoring offers long-term benefits but usually has the greatest development risk and timeline.
• Using DMS as a schema-conversion tool: DMS primarily migrates and replicates data. SCT helps convert schemas and identify incompatibilities.
• Assuming SaaS replacement is effortless: Repurchase can involve data transformation, identity integration, customization loss, and process changes.
• Ignoring retirement: Eliminating unnecessary workloads is often the lowest-cost and lowest-risk migration action.
• Migrating retained workloads by default: A workload with valid business, regulatory, licensing, or dependency reasons to remain on premises may be correctly marked retain.

Real-World Engineer Notes

• Assess application dependencies before selecting a strategy. A server that appears independent may rely on databases, file shares, DNS, identity systems, batch jobs, or hard-coded IP addresses.
• Treat rehost as a possible first phase. After stabilizing the application on AWS, a later program can replatform or refactor it based on measured value.
• Test the full operational model, not only application functionality. Validate backups, patching, monitoring, alerting, incident response, access control, and recovery procedures.
• For MGN migrations, plan source-agent deployment, replication bandwidth, staging and production instance sizing, cutover ownership, rollback, and post-cutover validation.
• For DMS migrations, validate data types, indexes, stored procedures, triggers, sequences, character sets, replication behavior, and application connection changes.
• Cloud-native refactoring should be driven by measurable outcomes such as throughput, deployment frequency, recovery objectives, or reduced operational effort—not by technology preference alone.
• A migration portfolio should record the selected R, rationale, dependencies, target date, owner, estimated effort, and any later modernization opportunity.

Quick Reference Summary

• Retire: Remove it.
• Retain: Leave it for now.
• Relocate: Move the environment with minimal change.
• Rehost: Lift and shift the server or application to a new host, usually EC2.
• Repurchase: Replace the application, often with SaaS.
• Replatform: Make limited changes to use managed AWS services.
• Refactor: Redesign for cloud-native architecture.
• AWS MGN: Primary AWS service for server lift-and-shift migrations to EC2.
• AWS DMS: Database migration and ongoing replication.
• AWS SCT: Schema and code assessment/conversion assistance for heterogeneous database migrations.

Flashcards

1. Q: What does the retire strategy mean?

A: Decommissioning a workload that no longer provides sufficient business value.

1. Q: Why might an organization retain a workload?

A: It may have unresolved dependencies, remaining hardware value, low migration priority, regulatory constraints, or insufficient migration benefit.

1. Q: What is the difference between relocate and rehost?

A: Relocate moves an environment or platform with minimal change, while rehost moves the operating system and application to a new host such as EC2.

1. Q: Which AWS service is commonly used for server lift and shift to EC2?

A: AWS Application Migration Service, also known as AWS MGN.

1. Q: What is repurchase?

A: Replacing an existing application with a different product or SaaS service.

1. Q: What is the goal of replatforming?

A: To make limited changes that allow an application to use managed cloud services without a complete redesign.

1. Q: Which AWS services are commonly associated with database replatforming?

A: Amazon RDS and Amazon Aurora.

1. Q: What is the role of AWS DMS?

A: It migrates and can continuously replicate data between supported database systems.

1. Q: What is the role of AWS SCT?

A: It assesses and helps convert database schemas and, where supported, application SQL between database engines.

1. Q: Which migration strategy usually requires the most development effort?

A: Refactor or re-architect.

1. Q: Name two patterns commonly used during refactoring.

A: Serverless functions, containers, managed databases, event-driven messaging, and workflow orchestration are examples.

1. Q: Why can rehosting still require application changes?

A: Drivers, agents, operating-system configuration, networking, monitoring, licensing, and integrations may differ in the target environment.

Practice Questions

Question 1

A company must leave its data center within six months. It has hundreds of Windows and Linux virtual machines, and the business wants minimal application changes and low downtime. The team wants to replicate servers continuously, test them in AWS, and perform a final cutover. Which option best meets the requirement?

A. Refactor the applications into Lambda functions
B. Use AWS MGN to rehost the servers on EC2
C. Use AWS SCT to convert the applications to containers
D. Repurchase all applications as SaaS products

Correct answer: B

Explanation: The requirement describes a lift-and-shift migration with continuous replication and a controlled cutover. AWS Application Migration Service is designed to rehost servers on EC2. Refactoring and repurchasing would introduce unnecessary scope and risk.

Question 2

An application uses a self-managed MySQL database. The company wants to preserve the application architecture but eliminate database-server patching and routine administration. Which migration strategy is most appropriate?

A. Retain
B. Repurchase
C. Replatform the database to Amazon RDS or Aurora
D. Refactor the entire application into an event-driven architecture

Correct answer: C

Explanation: Moving the database to a managed service while keeping the application largely intact is replatforming. A complete application redesign is not required by the stated objective.

Question 3

A company is moving from one database engine to another. It needs to convert incompatible schemas and continuously replicate changes until the final cutover. Which combination should it evaluate?

A. AWS MGN and VM Import/Export
B. AWS DMS and AWS SCT
C. Amazon S3 and Amazon CloudFront
D. AWS Elastic Beanstalk and AWS CodeDeploy only

Correct answer: B

Explanation: AWS SCT helps assess and convert schemas between database engines, while AWS DMS migrates data and supports ongoing replication for reduced downtime.

Question 4

During application portfolio analysis, a team finds a legacy reporting server that has had no users for two years. The business owner confirms that its reports have been replaced by a newer system. Which 7 R should be selected?

A. Retain
B. Rehost
C. Replatform
D. Retire

Correct answer: D

Explanation: The workload has been replaced and no longer provides business value. Retiring it avoids migration cost and reduces the attack surface.

Question 5

A company wants to improve elasticity and deployment speed. It plans to decompose a monolithic application into containerized services, move data to managed databases, and introduce queues and event-driven processing. Which strategy does this represent?

A. Relocate
B. Rehost
C. Replatform
D. Refactor or re-architect

Correct answer: D

Explanation: Decomposing the application and adopting cloud-native, event-driven components is a substantial redesign. That is refactoring or re-architecting, not a limited platform change.