Benchmarking for queries and indexes

[k88hudson-cfa]

Let's do some performance testing on simulations that (a) have lots of entities (maybe like 100 million?) and (b) have different configurations of properties across the population, particularly for multi-property queries.