Modern scientific research produces data at an unprecedented rate.

The Large Hadron Collider at CERN is an extreme example, producing roughly 1 petabyte (1015 bytes) of raw data per second from particle collisions. But even more standard facilities produce copious amounts of data, such as the beamlines at different synchrotron facilities. For example, the Advanced Photon Source at Argonne National Laboratory can generate data up to 20 gigabytes (10¹⁰ bytes) per second.¹