A computing model that starts each job when its needed data arrives.
Think of kids painting one big wall. Each kid starts as soon as the color they need arrives. Nobody waits for the whole paint box.
The model starts ready tasks right away. It helps independent jobs run at the same time.
TensorFlow
TensorFlow uses dataflow graphs to show job dependencies.
MapReduce
MapReduce processes data in stages. It is a classic dataflow example.
Distributed Computing
The Dataflow Model finds jobs that can run at the same time.