I had "dedicated" AWS instances that were rebooted. A dedicated instance means there is only one guest per box, right? So I'm curious why those had to be rebooted if there is no network-facing vector to this vulnerability. I guess because we could have read from the hypervisor's memory?