Accelerate Workload Migration and Modernization
Migration questions test the 7 Rs (retire, retain, rehost, relocate, replatform, repurchase, refactor) and the tools that support each phase. Know AWS Application Migration Service for lift-and-shift server replication, AWS Database Migration Service with Schema Conversion Tool for heterogeneous database moves, and AWS DataSync or Snow Family for bulk data transfer. Understand the Migration Hub for tracking and the Migration Evaluator for business cases.
Modernization patterns
Modernization means decomposing monoliths into services, adopting managed runtimes, and automating operations. Know when to replatform (move to managed services with minimal code change) versus refactor (redesign for cloud-native). Understand strangler-fig migration, API facades, and event-driven integration. For cutover, plan for dual-running, validation, and rollback.
Traps to watch
- Using DMS for application code migration; it moves data, not logic.
- Assuming Schema Conversion Tool converts all procedural code automatically; some requires manual rewrite.
- Forgetting DataSync is for online transfer while Snowball suits limited-bandwidth bulk moves.
- Skipping a pilot and cutting over everything at once.
Worked example
A company moves a monolithic order system to AWS. First rehost the servers with Application Migration Service to establish a baseline, then migrate the Oracle database to Aurora PostgreSQL using DMS and Schema Conversion Tool, testing converted stored procedures manually. Next, apply the strangler-fig pattern: route new order APIs through an API Gateway facade while the monolith handles legacy paths, gradually shifting traffic. Use Migration Hub to track progress and validate each cutover with parallel runs. Rule of thumb: migrate in waves, pilot first, and modernize incrementally rather than rewriting everything at once.