Thanks, that's actually very helpful. It's probably because they are big, but they look like they have issues very frequently. Maybe I'm off, but shouldn't a well planned architecture have less downtime than this? Everyone will experience a switch of server failure once in a while, but you're...