what is streaming
- data stream is an abstraction representing an unbounded dataset. infinite and ever growing; event streams are ordered, immutable data records, event streams are replayable.
(°0°)
kafka has stream processing
- kafka can support stream processing, like processing or transformaion capabilites; consumer, process and produce events without relying on an external processing framework. data bus/integration + process
refer
(°0°)
concepts
- time(event time, log append time, processing time)
- state(store middle stage info, like counting the number of events by type involve multiple events; local state, external state);
- Stream-table duality(contains current state and the a history of changes)
- time windows(size of the window; how often the window moves; how long the window remains updatable)
streams vs consumer
refer
(°0°)