AWS Cloud Practitioner: Infrastructure and Compute

24 questions with the correct answer and an explanation for each. Topic: Technology: Infrastructure and Compute. Free, no account needed.

These questions are written from AWS’s published exam guide and service documentation. They are not real exam questions — nobody who has sat the exam may reproduce those, and any site offering them is putting your certification at risk.

Take this test interactively

1. An engineer wants to browse an AWS account, see what resources currently exist, and make a single one-off change without writing any code. Which method fits best?

  1. AWS Management Console
  2. AWS CLI
  3. AWS SDKs
  4. Infrastructure as code

Answer: A — AWS Management Console

The console is a browser interface suited to exploring an account and doing one-off tasks.

2. A systems administrator wants to run the same set of AWS operations repeatedly as part of a script executed from a terminal. Which method fits best?

  1. AWS Management Console
  2. AWS CLI
  3. AWS SDKs
  4. Infrastructure as code

Answer: B — AWS CLI

The CLI is commands run at a terminal, suited to scripting and repeatable operations.

3. A development team is writing a Python application that needs to create and manage AWS resources directly from within its own code, rather than by shelling out to external commands. Which method fits best?

  1. AWS Management Console
  2. AWS CLI
  3. AWS SDKs
  4. Infrastructure as code

Answer: C — AWS SDKs

SDKs are libraries for languages such as Python used to call AWS from within your own application.

4. A company must build the same infrastructure identically across three separate environments and wants to eliminate configuration drift between them. Which approach fits best?

  1. AWS Management Console
  2. AWS CLI
  3. AWS SDKs
  4. AWS CloudFormation

Answer: D — AWS CloudFormation

Infrastructure as code, delivered by CloudFormation, describes infrastructure in a file so it is built identically and consistently, eliminating drift.

5. A company needs certain workloads to run on AWS infrastructure that is physically installed inside its own data centre, connected to resources it also runs in the cloud. Which AWS offering covers this, and which deployment model does it belong to?

  1. AWS Outposts, under the hybrid model
  2. AWS Outposts, under the cloud model
  3. AWS CloudFormation, under the hybrid model
  4. AWS CloudFormation, under the on-premises model

Answer: A — AWS Outposts, under the hybrid model

Outposts is named inside the hybrid bullet: hybrid connects cloud resources to infrastructure you still run yourself, and Outposts places AWS infrastructure in your own data centre.

6. A company runs part of its infrastructure in the cloud and keeps the rest running in facilities it operates itself, with the two connected together. What is this deployment model called?

  1. Cloud
  2. Hybrid
  3. On-premises
  4. Multi-cloud

Answer: B — Hybrid

Hybrid connects cloud resources to infrastructure the company still runs itself.

7. A company must keep all customer data physically within its home country because of a local law. Which factor should drive its choice of AWS Region?

  1. Latency
  2. Data sovereignty
  3. Disaster recovery
  4. Price

Answer: B — Data sovereignty

A legal requirement that data remain within a country is data sovereignty, one of the explicit reasons for choosing a Region.

8. A team compares identical compute resources across two AWS Regions and finds one is noticeably cheaper than the other, even though both meet its latency and legal needs. What explains the difference?

  1. Data sovereignty rules differ between the Regions
  2. Availability Zone count differs between the Regions
  3. Price varies by Region
  4. Edge location coverage differs between the Regions

Answer: C — Price varies by Region

Price is named directly as a reason to choose one Region over another, since it varies by Region.

9. An application must keep running even if a single data center loses power and networking. How should it be architected?

  1. Deploy across multiple Availability Zones
  2. Rely on a single data center with redundant power
  3. Deploy across multiple Regions
  4. Cache content at edge locations

Answer: A — Deploy across multiple Availability Zones

Availability Zones sit in separate facilities and share no single points of failure, so running across several survives the loss of any one data center.

10. A media company wants users around the world to load images and video with minimal delay, without needing any application logic to run near those users. Which part of the AWS global infrastructure should it use?

  1. Edge locations
  2. Availability Zones
  3. Additional Regions
  4. Additional data centers within one Availability Zone

Answer: A — Edge locations

Edge locations, used by Amazon CloudFront, cache content close to users for lower latency but are not where application code runs.

11. A company wants to keep operating even if an entire Region becomes unavailable, not just a single data center within it. What should it do?

  1. Spread the workload across more Availability Zones in that Region
  2. Deploy into a second Region
  3. Add more edge locations near the Region
  4. Add more data centers to the existing Availability Zones

Answer: B — Deploy into a second Region

Multiple Regions are what give disaster recovery; multiple Availability Zones only protect against a data-center-level failure inside one Region.

12. A company runs a web application that processes each request server-side, and wants low latency for users in both Europe and Asia. What should it do?

  1. Cache the application at edge locations near each user
  2. Deploy the application across more Availability Zones in one Region
  3. Deploy the application in Regions located in Europe and Asia
  4. Rely on a content delivery network to execute the application near users

Answer: C — Deploy the application in Regions located in Europe and Asia

Serving distant user populations with low latency for an application that runs logic (not just cached content) calls for Regions near those users, since edge locations only cache content and do not run applications.

13. A team needs an EC2 instance family suited to analytics workloads that hold a large dataset entirely in memory for fast querying. Which family fits?

  1. Compute optimized
  2. Memory optimized
  3. Storage optimized
  4. General purpose

Answer: B — Memory optimized

Memory optimized instances are built for large datasets held in memory, unlike the CPU, disk-throughput, or balanced-resource focus of the other families.

14. A development and test environment sits idle most of the day and is heavily used only during occasional test runs. The team does not want to pay for constant high CPU performance. Which EC2 instance option fits?

  1. General purpose instances
  2. Compute optimized instances
  3. Burstable performance (T family) instances
  4. Memory optimized instances

Answer: C — Burstable performance (T family) instances

Burstable performance instances provide a baseline CPU level with the ability to rise above it using earned CPU credits, which is what suits workloads that are quiet most of the time and occasionally busy, such as dev and test environments.

15. A company wants to lower its EC2 costs by committing to a consistent dollar amount of usage per hour for three years, without locking itself into a specific instance type or Region. Which purchasing option fits?

  1. Reserved Instances
  2. Savings Plans
  3. Capacity Reservations
  4. Dedicated Instances

Answer: B — Savings Plans

Savings Plans commit to a dollar amount of usage per hour, while Reserved Instances instead commit to a fixed configuration of instance type and Region.

16. A team wants to run its existing containerized application, using the container orchestration it has already set up, without provisioning or managing any servers. Which service fits?

  1. AWS Lambda
  2. AWS Fargate
  3. Amazon EKS
  4. AWS Batch

Answer: B — AWS Fargate

Fargate is serverless compute for containers and works with existing ECS or EKS orchestration, whereas Lambda runs code without containers at all and EKS on its own is managed Kubernetes rather than a way to avoid managing servers.

17. An application's incoming traffic needs to be spread across multiple running instances so that no single instance gets overloaded. Which capability provides this?

  1. A load balancer
  2. Auto Scaling
  3. A burstable performance instance
  4. A container image registry

Answer: A — A load balancer

A load balancer distributes incoming traffic across multiple targets, while Auto Scaling instead changes how many targets the traffic is spread across.

18. A team wants to deploy and run a web application without itself configuring the underlying servers or infrastructure. Which service fits?

  1. AWS Elastic Beanstalk
  2. Amazon Lightsail
  3. AWS Batch
  4. AWS Outposts

Answer: A — AWS Elastic Beanstalk

Elastic Beanstalk deploys and manages an application without requiring the user to configure the infrastructure, unlike Lightsail's simplified bundles, Batch's batch jobs, or Outposts' AWS hardware in your own data centre.

19. A company runs its database on an Amazon EC2 instance instead of a managed service. Which of the following remains the company's own responsibility?

  1. OS patching
  2. Server maintenance
  3. Hardware lifecycle
  4. Power, network, and cooling

Answer: A — OS patching

On EC2, AWS manages the underlying hardware, but the customer still installs and patches the operating system and everything above it.

20. A company wants its production database to fail over to a synchronously replicated standby in a different Availability Zone if the primary instance goes down. Which feature should it enable?

  1. A read replica
  2. A Multi-AZ deployment
  3. An Amazon ElastiCache cluster
  4. AWS Database Migration Service

Answer: B — A Multi-AZ deployment

A Multi-AZ deployment keeps a synchronously replicated standby in another Availability Zone specifically for availability and failover.

21. A company's database instance cannot keep up with the volume of queries coming from its reporting dashboards. The company wants to add extra copies of the database that can serve those queries directly. What should it add?

  1. A Multi-AZ standby replica
  2. A read replica
  3. An Amazon ElastiCache cluster
  4. AWS Database Migration Service

Answer: B — A read replica

Read replicas are the dedicated feature for scaling read traffic; a Multi-AZ standby exists for failover and does not serve reads in an instance deployment.

22. An application needs to store and query highly connected data, such as relationships in a social network or the links in a fraud ring. Which service fits this need?

  1. Amazon Neptune
  2. Amazon DocumentDB
  3. Amazon DynamoDB
  4. Amazon ElastiCache

Answer: A — Amazon Neptune

Neptune is the graph database, purpose-built for relationship-heavy data like social networks and fraud rings.

23. A company is migrating a database to a different database engine than the one it currently uses, and needs the schema converted before the data itself is moved, while keeping the source database usable throughout. Which approach should it use?

  1. AWS Schema Conversion Tool to convert the schema, then AWS Database Migration Service to move the data
  2. AWS Database Migration Service to convert the schema, then AWS Schema Conversion Tool to move the data
  3. AWS Database Migration Service alone
  4. AWS Schema Conversion Tool alone

Answer: A — AWS Schema Conversion Tool to convert the schema, then AWS Database Migration Service to move the data

When the target engine differs from the source, SCT converts the schema first and DMS then moves the data while keeping the source usable. DMS alone covers only a same-engine move.

24. A high-traffic application needs a serverless key-value database that responds in single-digit milliseconds. Which service fits?

  1. Amazon DynamoDB
  2. Amazon RDS
  3. Amazon Aurora
  4. Amazon DocumentDB

Answer: A — Amazon DynamoDB

DynamoDB is the serverless key-value and document database built for single-digit millisecond performance.

Take this test interactively

More AWS Cloud Practitioner practice tests