Continuously processing data records as they arrive in real time, rather than storing them first and processing in bulk. Like a moving conveyor belt that processes items one by one as they flow through, rather than waiting for a full batch. Core services include Apache Kafka (AWS MSK, Confluent), AWS Kinesis, GCP Dataflow, Azure Stream Analytics, and Apache Flink.
A ride-sharing app uses stream processing to analyze GPS coordinates from thousands of drivers in real time. Each location update is processed instantly to match nearby drivers with passengers, calculate ETAs, and detect surge pricing zones.
These services cover similar stream-processing building blocks: event ingestion, durable streaming transport, and real-time processing. AWS commonly combines Kinesis or MSK with Managed Service for Apache Flink. Azure often uses Event Hubs for ingestion and Stream Analytics for SQL-based real-time processing. GCP typically uses Pub/Sub for event intake and Dataflow for Apache Beam-based stream processing. OCI provides Streaming for event ingestion and Data Flow for Apache Spark-based processing, though OCI's managed options are less directly equivalent to Flink- or Stream Analytics-style continuous event processing.