... en-vironments, and software license cost as Hadoop. Although theperformance of HadoopDB does not in general match the perfor-mance of parallel database systems, much of this was due to thefact ... a clear tradeoff betweenfault tolerance and performance, and parallel databases tend tochoose the performance extreme of these tradeoffs. For example,frequent check-pointing of completed sub-tasks ... the higher performance of all queries that process this data. For certain tasks, such as the Jointask, the factor of 10 load cost is immediately translated into afactor of 10 performance benefit.7....