If you were using a high availability platform, like a Scale HC3 (just as an example off of the top of my head) you could easily get by with just a single RDS instance and a single file server instance, which would potentially reduce the licensing needs for Windows (this depends on how you want to use it), and if RDS were to be on a node that failed it would be automatically migrated to a working node. And the same for the file server. So you get automatic, instant recovery for failure without needing a complex load balancing scenario or external high availability tools.
RDS and file servers are ideal roles for platform high availability like Scale HC3 provides.