An insurance data source delivers one complete extract at 02:00 every night, and nothing changes during the day. Analysts refresh their dashboards once each morning, and no downstream consumer needs intraday data. A stakeholder asks why the team is not building a streaming pipeline like a competitor did. What should drive the choice between streaming and batch processing here?

1. A. The freshness requirement: daily refresh satisfies every consumer, so scheduled batch is simpler and cheaper; streaming pays off only when data must arrive within minutes ✓ A
2. B. Industry fashion: streaming is always the recommended modern architecture for new data platforms, regardless of any actual latency requirement 💬 B
3. C. Engine unification: because Databricks runs batch and streaming on one unified engine, every ingestion must now be written as a continuous streaming job 💬 C
4. D. Data volume alone: the decision depends exclusively on how many bytes the source produces each night, never on how fresh consumers need the data 💬 D

Reveal answer

Correct answer: The freshness requirement: daily refresh satisfies every consumer, so scheduled batch is simpler and cheaper; streaming pays off only when data must arrive within minutes

Reference: https://docs.databricks.com/aws/en/data-engineering/batch-vs-streaming