An event streaming system persistently transports and stores high-volume events so independent applications can process them in real time or replay them later. Producers publish events to topics, while consumers read them at their own pace. section 15.1
Unlike traditional pub/sub queues, platforms such as Apache Kafka retain events on disk and preserve their arrival order. This decouples services, supports parallel processing, and lets consumers recover or process events independently. section 15.1 section 15.9
Common stream-processing applications include:
- Counting online ad views. section 15.1
- Detecting fraudulent transactions. section 15.1
- Analyzing production-line sensor data to predict equipment failures. section 15.1
- Processing clickstreams to personalize content. section 15.1
Producer-to-consumer flow
Applications or devices produce events, the platform stores them in a message log and routes them through topics, and subscribed consumers process selected events independently. If one consumer fails, other consumers can continue operating.
