Hello Sharath,
Unfortunately NO !
We do not have the Archetecture to fullfill your demands at this point of time.
The clients will try to reach the same server where they were enrolled for the first time.
Yes, you can have clusters replicating amongst themselves with the data sync.
Usually it takes some time to transfer the user data in cluster environment.
Yes, again you can use the cluster member as a failover server but you need to route the clients to this server they are not smart enough to understand the load balancing Architecture.
I think this has answered all your queries.