Failover is at instance level, you can initiate failover manually through the GUI's or through powershell etc as normal practice for a single instance, just choose the instance you want to failover.

Some info below from our new WP EMC MULTISITE DISASTER RECOVERY FOR MICROSOFT SQL SERVER 2012

https://community.emc.com/docs/DOC-14614

     Table 1 Comparing FCI nodes and availability group replicas

Exactly

For each SQL Server Instance, a virtual Instance will be created (MSSQLSERVER1, MSSQLSERVER2, etc)

Each SQL Server Virtual Instance will have its own installation of binaries and also Cluster Group

   Each Cluster Group will have its own resource dependancies (Net name, IP, Storage, SQL Instance (SQL Service)

You would perform a Move-Group (Move this service or application to another node).

It is vitally important that there are no underlying dependancies across SQL Cluster Groups, which in the event of say SQL #1 Group failing over to Node B, that it could affect any other of the 2-5 Cluster Groups.

     An example of this is an underlying disk resource used by both cluster groups (but cluster resource logic should prevent this from being configured)

Thanks James & Michael for your helpful answers!

Could I also have your advise on the SQL server storage sizing in SQL cluster environment?  Any best practice consideration or recommendation?