A system for scalable, fault-tolerant, distributed computation over large data collections.

Applications implement {@link org.apache.hadoop.mapred.Mapper} and {@link org.apache.hadoop.mapred.Reducer} interfaces. These are submitted as a {@link org.apache.hadoop.mapred.JobConf} and are applied to data stored in a {@link org.apache.hadoop.fs.FileSystem}.

See Google's original Map/Reduce paper for background information.