Your data fits in RAM.

If this is a test of how well the process would work for big data sets, then I'm not sure it's even big enough to be able to extrapolate from.