One big advantage with this system is that the Indexing is totally independent of the storage, and we can wipe the Index and rebuilding them at any time. It will also allow us to distribute the Index to many hosts, to improve scalability. The main disadvantage is that for optimal performance, we are sacrificing Query consistency, and there are times when the data stored and the data queried are out of sync. Future versions of the Indexing subsystem will probably allow other 2 out 3 combinations of the CAP syndrome (Consistency, Availbility and Partitioning tolerance).