That's great. It's still over-engineered given processing this data in-process is more than fast enough at a scale far greater than theirs.