Comment by MillstoneX

Can you give an example for a join between two different sharded document collections that can be executed efficiently on say 100 servers?