If I run a really dumb query that loads in a huge number of nodes into the expanded tree cache am I impacting other queries running on MarkLogic? In other words is the expanded tree cache a shared resource or is there portion allocated to my connection perhaps?