I should add that it's the massive network switch that makes this all possible. You can scale disk I/O by adding disks, and you can scale in-memory sorting by

I should add that it's the massive network switch that makes this all possible. You can scale disk I/O by adding disks, and you can scale in-memory sorting by adding cores and memory. But it's awfully expensive to buy your way out of a bottleneck in your network when everyone needs to talk to each other at full speed.