Choose compute and data services | AWS Cloud Practitioner: foundational decisions

Understand the concept

EC2 provides virtual machines and guest-system control. Lambda executes functions in a managed event-driven model with limits and characteristics requiring assessment. Containers package applications but still need execution, observability, and security. Serverless does not mean servers are absent; it changes who manages them.

Apply and decide

RDS supports managed relational engines; DynamoDB offers a NoSQL database with a different access and design model. Choose based on data and operation patterns. SQS helps decouple processing through queues; SNS distributes notifications. Managed services reduce some tasks but do not choose your data model, functional consistency, or idempotency.

Workplace application

Compare continuous processing requiring operating-system control with a short event-triggered task. Do not choose solely by service name: identify needed control, limits, integration, and responsibilities. For an existing database, confirm engine and access model before assuming a NoSQL service is a direct substitute. When separating producer and consumer with a queue, document who handles failures and repeated work.

IN PRACTICE

A file-arrival event can trigger a short function. A legacy application needing specific system configuration may require another option.

Common pitfalls

Choosing a service without checking data model, limits, and responsibilities.

Related topics: Objects, blocks, files, and networking · Costs, observability, and auditing

Take this idea with you

Choose services by needs and limits, not just the “managed” label.

Create account

Reference: CLF-C02 Domain 3: Cloud Technology and Services · CLF-C02