I’ve had an almost similar problem a while ago and have been comparing/testing various hardware solutions including Tintri, Nutanix, Simplivity, Atlantis and so on. The idea was to replace our main production cluster in HQ along with some minor clusters in our branch offices. Preferably without (or at least as short as possible) downtime. As usual in such cases, the budget was crucial. And since we are a DELL shop, the solution had to be based on DELL servers.
After digging a lot we decided to go towards StarWind hyper-converged solution https://www.starwindsoftware.com/starwind-hyperconverged-appliance. One of the main benefits compared to other solutions like Nutanix is that Starwind does not do any checksumming/splitting of the data. It’s a simple mirror which means you need more storage but this overhead grants you a way better performance and data integrity and consistency. Of course, all the bells and whistles like remote management, web management, VVols, SMI-s, RDMA along with various caching types and automated storage tiering are present. It’s not superfluous to mention that their technical support guys helped us to migrate our existing infrastructure in the shortest time without downtime at all.
Almost a year passed already, very happy so far.