If it fits on a single disk, it's not big data.

Me, circa 2007, when lovely hadoop guy trying to tell me how good it was for 10m rows of data... It was much slower than Oracle free version as I recall.