Explain the difference between batch and streaming data processing in Data Fusion.

Batch processes bounded datasets in discrete runs; streaming processes unbounded data with continuous execution. Why the distinction matters: Batch has predictable cost (run N times, pay N × job cost); streaming has always-on cost and state management. Data Fusion context: Batch templates (e.g., JDBC to BigQuery) are scheduled; streaming uses Pub/Sub or Kafka sources. Scalability: Batch scales by partitioning and parallel jobs; streaming scales by partitions and backpressure handling. Cost implication: Streaming typically costs 2–5x more per record due to low latency, state storage, and smaller micro-batches. Architectural choice: Use batch for historical loads, daily aggregations, and backfills; use streaming only when latency SLA justifies the premium (e.g., fraud detection < 30s). Compromise: Micro-batch (e.g., 5-min windows) balances latency and cost. Best practice: Design pipelines so the same logic runs in batch (for backfill) and streaming (for real-time); avoid divergence.