Reordering fields in structs and classes can result in smaller footprints if there are inefficiencies due to alignment holes. A common tool for identifying these "holes" is something called `pahole`. I ran it against the mongod executable and found that there is substantial room for improvement in many of our structs and classes, which could ultimately lead to better cache and memory utilization. What I did not investigate is whether or not this could actually result in faster performance in the server. There may be some low-hanging fruit here but we'd have to prototype it to know for sure.