A recent HN posting referenced article http://www.facebook.com/note.php?note_id=89508453919 where Facebook engineers discuss Hive and Map-Reduce. The description there maps very closely to this 2001 patent from SenSage/Addamark: http://www.google.com/patents/about?id=_VaHAAAAEBAJ, in its discussion of:
* distributed compressed storage * columnar databases * temporal partitioning * distributed query planning and processing * ... many other aspects ...
The CEO of SenSage (formerly Addamark) points this out at http://www.sensage.com/blog/2009/06/11/mapreduce-made-easy-the-future-of-database-analytics/. Various other companies have also taken the Hive/Hadoop/Map-Reduce model, including Vertica (founded by Michael Stonebraker after his stint as tech advisor at SenSage). These commercial offerings are quite popular and have lots of uptake in the market. Probably a lot faster than standard Hadoop/Map-Reduce.