A new computing paradigm could help us to overcome a key performance bottleneck to improve our ability to query large data bases.