A programming model from Google for processing huge data sets on large clusters of servers. The "map" process transforms the input into key-value pairs, and the "reduce" procedure groups, sorts, filters and summarizes the data. See
Hadoop and
key-value pair.