6 AWS Compute Architecture and Practice Free Online FlashCards

What factors should guide an AWS compute choice?

Compare execution pattern, runtime and resource needs, scaling behavior, operational control, and team expertise.

When is Amazon EC2 a good fit?

EC2 fits applications needing operating-system access, specialized instances, persistent processes, or compatibility with existing software. You manage instances, patching, capacity, and much of scaling.

What limits and state-design considerations apply to AWS Lambda?

Lambda suits short, event-driven tasks. A standard invocation can run for at most 15 minutes, and durable state should live outside the function.

How do ECS and EKS differ?

ECS provides AWS-native container orchestration; EKS provides Kubernetes on AWS for workloads or teams that need Kubernetes compatibility and its ecosystem.

Which AWS services orchestrate containers, and which provide their compute capacity?

ECS and EKS orchestrate containers; Fargate and EC2 provide the compute capacity on which containers run.

What does AWS Fargate manage for container workloads?

Fargate runs containers without requiring you to manage EC2 hosts. You choose task or pod CPU and memory; Fargate works with ECS and EKS.

When might containers run on ECS or EKS using EC2 capacity?

Use ECS or EKS on EC2 when containers need host-level control, specialized instance capacity, or a capacity model suited to sustained utilization; you operate and scale the EC2 capacity.

Does packaging a Lambda function as a container make it a continuous container service?

A container image can package a Lambda function, but this does not make Lambda a general-purpose, continuously running container service.

What is a suitable starting design for irregular, quick image resizing after an S3 upload?

Use an S3 object-created event to trigger Lambda, resize the image, and write the output to S3. Keep durable state in storage.

How can a long report-generation job avoid keeping the client's request open?

Have the API accept the request and enqueue a job, then let a worker process it independently. Store job status and results outside the worker.

What architecture suits a continuously running legacy app requiring a specific OS and host-level access?

Run it on EC2. For higher availability, use an Auto Scaling group across multiple Availability Zones and a load balancer that checks target health.

How should durable state be handled when making compute replaceable?

Store durable data and job state in managed storage or databases rather than relying on a particular instance, task, or function environment.