How would you add a very large volume of data to an existing production database without disrupting the system?