Question

Under the MapReduce v1 programming model, which shows the proper order of the full set of MapReduce phases?

A. Map -> Combine -> Reduce -> Shuffle
B. Split -> Map -> Combine -> Reduce
C. Map -> Split -> Reduce -> Combine
D. Map -> Combine -> Shuffle -> Reduce

Answer 1

The correct order under the MapReduce v1 programming model is Split -> Map -> Combine -> Reduce. The Combine phase is optional and helps reduce network load, while the Shuffle occurs before the Reduce phase for data redistribution.

Step-by-step explanation:

The proper order of the full set of MapReduce phases under the MapReduce v1 programming model is: Split -> Map -> Combine -> Reduce. This means that the correct answer is option B. First, input data is split into chunks which are then processed by the Map tasks. Optionally, a Combine phase can be used locally on each node after the Map phase to reduce network load. After all Map tasks have finished, the framework performs the Shuffle and Sort phase (which is sometimes considered as part of Shuffle only), where data is redistributed across reducers. Finally, the Reduce tasks process the shuffled data to generate the final output.