Has anyone here used this before? If so, do you know how this tolerates latency? Could I use Remus to replicate a VM to LAX from NY or similar distances?
That's unlikely to be very useful since, to do transparent network failover, both VM hosts need to be on the same layer 2 network. And half the point of having geographically diverse sites is so they aren't on the same network.
If you had your own AS number, couldn't you multihome your address space with BGP, and still update the IP on the other machine if the other machine fails?
