Structured columnar data, analogous to pandas or R dataframes, but with a spark RDD basis, allowing massive, distributed SQL-like operation.