Kafka Streams - keshavbaweja-git/guides GitHub Wiki

Stream processing

State of a real time stream processing application needs to be saved in a state store -

  • In memory
  • Local file-system
  • Remote database (network delay)

Performance & Fault tolerance are key requirements of a state store.

Stream processors are classified as

  • Stateless processor
  • Stateful processor (aggregation, join, window)