In one sentence
Recognize compute, storage, database, networking, analytics, and AI services by use case rather than by name alone.
Key points
- S3 is the object store and data lake; Redshift is the warehouse; Athena queries S3 with SQL.
- Glue Data Catalog stores metadata; Glue ETL transforms; Kinesis Streams ingests; Firehose delivers.
- Use AWS AI Services for pretrained capabilities, SageMaker for custom models, and Bedrock for foundation-model APIs.
- RDS and Aurora handle relational transactions; DynamoDB handles large low-latency key-value workloads; DAX or ElastiCache caches.
Exam focus
- Polly speaks, Transcribe writes, Translate changes language, Comprehend understands text, Textract extracts documents, and Rekognition sees images.
- Multi-AZ is for availability, Read Replica for read scaling; Query is usually more efficient than Scan.
Common pitfalls
- Do not choose by product name alone; confirm data type, latency, control, operations, and cost constraints.
Remember
Start with the use case, then the service; separate responsibilities before eliminating similar options.
New services and migration map
| Need | Concept or service |
|---|---|
| Migration business case | Migration Evaluator |
| Server and dependency discovery | Application Discovery Service |
| Block-level server replication | AWS Transform MGN |
| Database full load and CDC | AWS DMS; add SCT for heterogeneous schema conversion |
| High-speed file migration | AWS DataSync |
| SFTP / FTPS / FTP / AS2 | AWS Transfer Family |
| GraphQL and real-time data | AWS AppSync |
| Web and mobile build and hosting | AWS Amplify |
| Full desktop / app streaming / isolated browsing | WorkSpaces / WorkSpaces Applications / Secure Browser |
| IoT certificates, MQTT, and rule routing | AWS IoT Core |
Remember Assess → Mobilize → Migrate & Modernize, then select the application strategy from the 7Rs.