They already were around, but I'm the only one who was trying to build some analytics engine on top of them. I think Cassandra uses bloom filters to decide whether a key could be stored on a certain node, and there is a paper from Smola that also uses PDSs.