That description belongs to Amazon S3 because S3 is an object storage service for files and data blobs rather than mounted block devices; it is used to store durable objects and retrieve them by bucket and key, not to explain how Amazon S3 Inventory provides the function to generate lists of objects and associated metadata for analysis.

That description belongs to Amazon S3 Standard because S3 Standard is appropriate for data requiring regular access across multiple AZs; it is used to serve frequently accessed objects with low-latency retrieval, not to explain how Amazon S3 Inventory provides the function to generate lists of objects and associated metadata for analysis.

That description belongs to S3 bucket policy because bucket policies are resource-based JSON permissions for supported S3 actions; it is used to control allowed and denied access to an S3 bucket and its objects, not to explain how Amazon S3 Inventory provides the function to generate lists of objects and associated metadata for analysis.

That description belongs to Amazon Transcribe because transcribe provides managed automatic speech recognition; it is used to convert spoken audio into a text transcript, not to explain how Amazon Comprehend provides the function to extract sentiment entities and other insights from text.

That description belongs to Amazon Translate because translate is a machine-translation service used by multilingual applications; it is used to translate text between supported human languages, not to explain how Amazon Comprehend provides the function to extract sentiment entities and other insights from text.

That description belongs to Amazon Textract because textract reads forms tables and text from supported document inputs; it is used to extract printed and handwritten text and document structure, not to explain how Amazon Comprehend provides the function to extract sentiment entities and other insights from text.

That description belongs to Amazon RDS because RDS automates selected relational database operations including backup and patching; it is used to operate a managed relational database engine, not to explain how NoSQL access-pattern design provides the function to design a database around known key lookups and throughput patterns.

That description belongs to RDS Multi-AZ configuration because RDS Multi-AZ options support failover when the primary or supporting infrastructure fails; it is used to improve database availability using a standby in another Availability Zone, not to explain how NoSQL access-pattern design provides the function to design a database around known key lookups and throughput patterns.

That description belongs to Database on Amazon EC2 because self-managing a database on EC2 leaves patching, backups and high availability largely to the customer; it is used to retain direct control of database host OS and engine installation, not to explain how NoSQL access-pattern design provides the function to design a database around known key lookups and throughput patterns.

AWS Batch is intended to queue and execute batch computing jobs with managed scheduling because batch provisions and schedules compute capacity for job queues and compute environments; in this design the new outcome is to execute event-triggered code without managing virtual servers, which calls for AWS Lambda rather than AWS Batch or the earlier Amazon EC2.

Gateway Load Balancer is intended to insert compatible virtual network appliances into traffic paths because gateway Load Balancer distributes flows across appliance targets for inspection or similar tasks; in this design the new outcome is to execute event-triggered code without managing virtual servers, which calls for AWS Lambda rather than Gateway Load Balancer or the earlier Amazon EC2.

Amazon EC2 primarily supports the earlier need to operate virtual servers with explicit OS and instance management because EC2 provides configurable compute instances where customers manage guest OS and application software, but the new task is to execute event-triggered code without managing virtual servers; that is why AWS Lambda is preferable here.

AWS Direct Connect is intended to establish dedicated private network connectivity from a site to AWS because direct Connect provides a dedicated network connection through supported locations; in this design the new outcome is to improve global network paths and health-based endpoint selection using anycast IPs, which calls for AWS Global Accelerator rather than AWS Direct Connect or the earlier Route 53 latency routing.

Route 53 latency routing primarily supports the earlier need to answer DNS requests with an endpoint selected for lower network latency because latency-based policies choose among configured Region records using AWS latency observations, but the new task is to improve global network paths and health-based endpoint selection using anycast IPs; that is why AWS Global Accelerator is preferable here.

Amazon VPC is intended to create a logically isolated IP network for AWS resources because a VPC defines virtual networking boundaries and hosts configurable subnets and routing; in this design the new outcome is to improve global network paths and health-based endpoint selection using anycast IPs, which calls for AWS Global Accelerator rather than Amazon VPC or the earlier Route 53 latency routing.

AWS X-Ray is intended to trace distributed requests through supported application components because x-Ray provides distributed tracing to help find latency and failure sources; in this design the new outcome is to build and host web or mobile application frontends with integrated workflows, which calls for AWS Amplify rather than AWS X-Ray or the earlier Amazon API Gateway.

Amazon SES is intended to send transactional or bulk application email because sES provides managed email sending and receiving capabilities; in this design the new outcome is to build and host web or mobile application frontends with integrated workflows, which calls for AWS Amplify rather than Amazon SES or the earlier Amazon API Gateway.

Amazon API Gateway primarily supports the earlier need to create a managed front door for application APIs because aPI Gateway exposes managed API endpoints with routing and controls, but the new task is to build and host web or mobile application frontends with integrated workflows; that is why AWS Amplify is preferable here.

RDS read replica is intended to offload eligible read queries from a relational database primary because read replicas provide copies usable to serve read traffic with replication behavior; in this design the new outcome is to improve database availability using a standby in another Availability Zone, which calls for RDS Multi-AZ configuration rather than RDS read replica or the earlier Amazon RDS.

Database on Amazon EC2 is intended to retain direct control of database host OS and engine installation because self-managing a database on EC2 leaves patching, backups and high availability largely to the customer; in this design the new outcome is to improve database availability using a standby in another Availability Zone, which calls for RDS Multi-AZ configuration rather than Database on Amazon EC2 or the earlier Amazon RDS.

Amazon RDS primarily supports the earlier need to operate a managed relational database engine because RDS automates selected relational database operations including backup and patching, but the new task is to improve database availability using a standby in another Availability Zone; that is why RDS Multi-AZ configuration is preferable here.

S3 bucket policy is intended to control allowed and denied access to an S3 bucket and its objects because bucket policies are resource-based JSON permissions for supported S3 actions; in this design the new outcome is to upload very large objects as separately transmitted parts, which calls for S3 multi-part upload rather than S3 bucket policy or the earlier Amazon S3 Object Lock.

Amazon S3 Inventory is intended to generate lists of objects and associated metadata for analysis because S3 Inventory produces scheduled reports about bucket objects and properties; in this design the new outcome is to upload very large objects as separately transmitted parts, which calls for S3 multi-part upload rather than Amazon S3 Inventory or the earlier Amazon S3 Object Lock.

Amazon S3 Object Lock primarily supports the earlier need to enforce supported WORM retention against object deletion or modification because object Lock supports retention modes and legal holds on eligible S3 object versions, but the new task is to upload very large objects as separately transmitted parts; that is why S3 multi-part upload is preferable here.

VPC endpoint addresses a different objective, to connect supported VPC workloads privately to selected AWS services, because VPC endpoints provide service connectivity without requiring a public internet path; the checklist instead requires the distinct outcomes to reach supported services through private network interfaces in a subnet; distinguish outbound address translation from direct public routing; select whether connection return traffic needs independent explicit rules.

NAT gateway addresses a different objective, to let private subnet resources initiate eligible outbound internet connections, because a NAT gateway translates egress traffic when appropriate routes are configured; the checklist instead requires the distinct outcomes to reach supported services through private network interfaces in a subnet; distinguish outbound address translation from direct public routing; select whether connection return traffic needs independent explicit rules.

Route table addresses a different objective, to decide where traffic from a subnet should be directed, because VPC route tables contain destination-prefix routing entries for network traffic; the checklist instead requires the distinct outcomes to reach supported services through private network interfaces in a subnet; distinguish outbound address translation from direct public routing; select whether connection return traffic needs independent explicit rules.

Amazon EC2 addresses the separate need to operate virtual servers with explicit OS and instance management because EC2 provides configurable compute instances where customers manage guest OS and application software; it is not as direct as AWS Lambda for the actual requirement to execute event-triggered code without managing virtual servers.

Application Load Balancer addresses the separate need to route HTTP and HTTPS requests using application-layer rules because aLB supports layer 7 routing by host names or URL paths to target groups; it is not as direct as AWS Lambda for the actual requirement to execute event-triggered code without managing virtual servers.

AWS Fargate addresses the separate need to run container tasks without managing EC2 container hosts because fargate supplies serverless compute for supported ECS and EKS workloads; it is not as direct as AWS Lambda for the actual requirement to execute event-triggered code without managing virtual servers.

AWS CLI is intended to run scripted AWS API operations from shell commands because the AWS Command Line Interface exposes AWS service operations to terminals and automation; in this design the new outcome is to run AWS-managed infrastructure at a customer premises for local requirements, which calls for AWS Outposts rather than AWS CLI or the earlier AWS CloudFormation.

AWS CloudFormation primarily supports the earlier need to recreate multi-resource infrastructure from versioned templates because cloudFormation models resources as stacks and applies managed creation and update operations, but the new task is to run AWS-managed infrastructure at a customer premises for local requirements; that is why AWS Outposts is preferable here.

AWS Cloud Development Kit (AWS CDK) is intended to define AWS infrastructure using familiar programming languages because AWS CDK constructs synthesize infrastructure definitions for deployment through CloudFormation; in this design the new outcome is to run AWS-managed infrastructure at a customer premises for local requirements, which calls for AWS Outposts rather than AWS Cloud Development Kit (AWS CDK) or the earlier AWS CloudFormation.