Within each phase, a set of parallel workflows define the activities performed and the artifacts that are used, created, or modified by each activity.

Then MapReduce shuffles the map output by key and does another parallel computation on the redistributed map output, writing the results to the file system in the Reduce phase of the computation.

MapReduce breaks down a problem into millions of parallel computations in the Map phase, producing as its output a stream of key-value pairs.