We currently have a 2 site federated deployment with stretched T0's and T1's (some local site based ones too). We have this 2 site model for a fairly typical Protected & DR site model and we use SRM for recovery. It is worth mentioning we do have active workloads in our DR site and thus run as active/active. DR works well and we're happy with this current model.
However. We are undertaking a migration to new kit and are looking at further enhancements to our VMware estate, specifically moving to a metro cluster setup where we have synchronous SAN replication etc etc. We're happy with how this model will play out from a compute and storage perspective but I'm not totally clear on how NSX-T will look and wonder if any of you run NSX-T in this mode
Essentially we will still have 2 sites with one being mainly used for DR but will need to run active workloads all the same. How does federation look in this model? I think I'm ok with host transport nodes where hosts from both sites sit in a single cluster for HA purposes - but what about edges?? I will need edges to obviously provide connectivity to the physical network but do they need to sit in separate clusters still? Would I still have 2 locations in NSX-T or would this appear as one location like in the vSphere world?
I'm just really struggling to visualise how NSX looks. If I remove NSX from the equation for arguments sake it's easy to visualise