A company is designing an API for a generative AI (GenAI) application that uses a foundation model (FM) that is hosted on a managed model service. The API must stream responses to reduce latency…

Question

Options

• AIntegrate an Amazon API Gateway HTTP API with an AWS Lambda function to invoke Amazon
• BConnect an Amazon API Gateway HTTP API directly to Amazon Bedrock. Simulate streaming by
• CConnect an Amazon API Gateway WebSocket API to an Amazon ECS service that hosts a
• DIntegrate an Amazon API Gateway REST API with an AWS Lambda function that invokes

• A
  76% (34)
• B
  7% (3)
• C
  2% (1)
• D
  16% (7)

Explanation

The requirements are streaming responses, token limit enforcement, and retry logic for timeouts/partial responses-all with the least operational overhead. An API Gateway HTTP API paired with a Lambda function invoking Amazon Bedrock is the right combination: Lambda can call Bedrock's streaming InvokeModelWithResponseStream API, enforce token limits via request parameters, and implement retry logic in code-all without managing any servers or containers. HTTP APIs on API Gateway have lower latency and cost than REST APIs and natively support Lambda streaming responses. Option B (direct API Gateway to Bedrock integration) cannot implement retry logic or complex token enforcement without a compute layer. Option C (ECS service) introduces container management overhead. Option D (REST API variant) is functionally similar to A but REST APIs have higher overhead than HTTP APIs for this use case, making A the preferred minimum-overhead solution.

No community discussion yet for this question.