The current definition of the Redundancy Visible Server Facet implies that at least one failover mode needs to be supported by the product which is fine. To make the requirement more explicit I propose to create additional (optional) CUs for each mode which replace (or extent) the single CU Redundancy Server. Since clients need to support a certain failover mode this proposal would also enables clear identification of the supported failover mode to end users who build a system.

Test cases exist for all failover modes and would just need to be moved from the CU Redundancy Server to the CU dedicated to the failover mode.