I don't know anyone who has rolled out this implementation yet -- mostly because of the huge cost in terms of servers. A multi-terabyte server is going to require about 11 VMs in order to operate properly... 3 x AZK, 2 x DB2, 2 x NAS (for arscache), 2 x TSM, 2 x CMOD. Now, with Rob Russel's article on IBM CMOD Docker Containers, it might make some of this easier to deploy, but it's still a little crazy in terms of infrastructure for what is already a very reliable service.
I still recommend Independent-Redundant configs for large CMOD customers who have a requirement for greater than three-nines (99.9%) uptime.
-JD.