VIM or EIOC redundancy

How does VIM and EIOC redundancy work on EtherNet/IP networks? 

Let's assume the VIM and EIOC are originating both Class 1 and Class 2 EtherNet/IP communications.  I would think the same configuration is  automatically download or cross-loaded and running in both the primary and secondary EtherNet/IP (VIM/EIOC) scanner whereas only the data from the Primary EtherNet/IP scanner is being processed in the Controller.   

I would assume you can manually force the Controller to switch over to use the Secondary EtherNet/IP scanner.  Is there an automatic switchover option?  What are the conditions that would cause it to failover to the secondary?  Does is favor the primary and automatically switch back?

The VIM and EIOC redundancy has two EtherNet/IP interfaces.  I can think of three network architectures as follows:

  1. Two completely separate EtherNet/IP system of devices,  The Primary devices are connected to the primary network and the Secondary devices are connected to the secondary.  All the Primary devices are independently collecting the same data as the Secondary devices. You essentially have a separate EtherNet/IP scanner for the Primary and Secondary EtherNet/IP and the Controller is just using one EtherNet/IP system or the other one.
  2. One set of devices, but each device has two EtherNet/IP adapters, one on the primary and one on the secondary.  Same as above.
  3. One EtherNet/IP network with simplex slave devices... The only redundancy is the VIM or EIOC scanner device.... In this case, how does the VIM or EOIC redundant scanner handle ownership of the simplex slave? Does it force the connection closed on one scanner and then open the connection on the second scanner?  Is the data cross-loaded or synchronized between the two EtherNet/IP scanners? or is it possible for "old data", "stuck bits', or other issues to arise as a result of a comm loss on one of the EtherNet/IP channels.