Every time you run an SQL query on BigQuery, for example, you are executing those same fundamental map shuffle primitives on underlying data, it's just that the interface is very different.