Pretty much all these distributed databases at their core somehow handle the partitioning of data. Even if sometimes it's just throwing up your hands and saying

Pretty much all these distributed databases at their core somehow handle the partitioning of data. Even if sometimes it's just throwing up your hands and saying there's a user has to provide us a partitioning function. That's true in non-relational databases (Hbase, Casandra, Mongo) and in relational ones (Volt, Vetica, this db). It looks like here the data structure enabling partitioning is based on prior art distributed B+ Tree. From the paper it looks like it allows for full ACID transactions. At least to the physical extents that ACID is possible.