I suppose it can be seen as twofold: 1. an AeroFS-like part, i.e your files reside only on your machines and are synced via a P2P protocol, so not-yet-synced f

I suppose it can be seen as twofold: 1. an AeroFS-like part, i.e your files reside only on your machines and are synced via a P2P protocol, so not-yet-synced files are unavailable when your source is offline (unless like AeroFS, they are opt-in stored in the cloud). 2. a content-hashing part, which means files on your machines that happen to have an identical file elsewhere can be obtained from that elsewhere, even if your machines are offline. Still, no source => no data, but that's a given. Technically, 1. is of course a special case of 2. It's just that AeroFS restricts peer requesting to your machines (and — optionally — their cloud storage).