I guess the lesson here is that <Space x Time> is always a constant!

A B-Tree is already scalable, fast, and doesn't require 128x more work for a single query. THAT is your competition here, not some "fixed amount of processing". Kangaroo et all can run concurrently on an unlimited number of computing devices, and use the one and same central DB. Which may not even need to be large at all, only adding a few DPs every now and then. However, those special values represent trillions of EC operations EACH.

Faster = Larger
Slower = Smaller