We are using two-tier backup schemes during data collection and again for the primary data repository.