A web app must show text as the model generates it. Which approach fits?
Why D is the answer
Streaming responses
Streaming lowers perceived latency.
In detail
- Streaming shows text as it is generated.
- Streaming does not make generation itself faster.
Use case
An illustrative scenario, not a real incident.
A writing tool shows text as it forms, so users can stop a poor draft early.
See it in different ways
flowchart LR A["Browser"] --> B["API Gateway WebSocket or streaming API"] B --> C["Lambda"] C --> D["Bedrock streaming response"] D --> A
mindmap
root((Inference))
On demand
Variable traffic
Provisioned Throughput
Steady and high
Batch
Offline and cheaper
Cross-Region
Spreads peaks
Streaming
Show text as it arrivesflowchart LR S0["Request"] --> S1["First chunks"] --> S2["More chunks"] --> S3["Done"]
Key point: Streaming lowers perceived latency.
Why the other options are wrong
- A. A nightly batch job that stores all answers in advance
- Stored answers are not live.
- B. Polling a database once a minute for the finished reply
- Polling once a minute is far too slow.
- C. Polling an S3 object until the full reply appears
- That shows nothing until the end.