The right level of parallelism for maps seems to be around . . . . . . . . maps per-node.

A. Generate intermediate key-value pairs B. Read and parse input data C. Aggregate values on the reducer side D. Execute MapReduce jobs

A. Task that processes the Map output B. Task that processes the Reduce output C. Task that processes the Map input D. Task that processes the Reduce input

A. Distribute data across multiple nodes B. Sort and shuffle data C. Combine intermediate key-value pairs on the mapper side D. Determine the reducer for each key

A. Manage computation resources B. Set job-level configurations C. Execute MapReduce jobs D. Read and parse input data