Skip to main content

Table 1 Batch vs stream processing

From: Composing high-level stream processing pipelines

Parameters Batch processing Stream processing
Amount of data Big data Small + big data
Type of computation Complex (must access all data) Simple
Focus Latency Throughput
Processing time Collect and process later (Near-) real time
Used by Hadoop mapreduce, spark Spark streaming, flink
Use cases Daily total turnover Instant fraud detection