Partially it just depends on how the service fog is implementing
appears to users. I think if all of the requests are made via the
same endpoint/API then it is probably fine for all of the fog code to
appear within the same service. For instance on EC2, volumes/disks
are provided by the compute API. I think that is fine and sensible.
The other thing is that it is probably easiest to do this for now, and
we can always update that and move things around later if we find a
better division.
Thanks!
wes