This is a pretty rosy view considering that AFAIK basic HA tasks like taking over an IP address from a failed machine is much slower in the cloud. And let's hope your HA software doesn't want to use multicast...
In addition - At the entrance of the cloud(external ip's of the balancer) - I would suggest dnsmadeasy that can do round robin dns along with dns monitoring with a low ttl of 5 minutes that in case of failure can change your A records.
So the service should be designed with not having to rely on ip re-assignments.