Well... sure. Just as long as you keep in mind that a "not really tested [something]" will very easily get replicated on multiple vendor devices. It's not like

Well... sure. Just as long as you keep in mind that a "not really tested [something]" will very easily get replicated on multiple vendor devices. It's not like they did it because they hate Vendor1 but will make sure to test all the corner cases with Vendor2. Some scenarios will by design ensure you can't have all of them fail because they are not susceptible to the same failure modes. But to compensate you introduce additional failure modes (especially human error) by using a heterogeneous environment. Not to mention the expense. I was actually hoping for an interesting human independent failure. A cosmic radiation caused bit flip that would occur on a particular device. :) Human error can overcome any setup.