Thanks! Unlike a kd-tree or b-tree variants, HyperDex does not build an auxiliary data structure. It turns out that keeping aux data structures in sync with the data is very difficult if you want to provide strong consistency guarantees. Hyperspace hashing is purely a mapping trick, not a distributed data-structure trick.

Agreed with you fully that value-dependent chains are neat. They allow the system to replicate and relocate data, without any need for background processes. VDCs are the key to HyperDex's strong consistency guarantees.