Hey, it’s Bryan from Oxide back with another FAQ Friday. A question that we get is:
When I’ve got an Oxide rack and there is a next generation CPU, let’s say, for example, that you bought a half-populated Oxide rack with our Milan-based sleds. What happens when we have our Turin-based sled, as we do right here, our Cosmo-based sled? How am I able to add this to a rack? Do I replace my older sleds? How does that work?
Well, the great news is that we have designed the rack to accommodate heterogeneous compute. So your next gen sled slides right in alongside your older generation sleds. We spent a lot of time on the software work to make sure that VMs can recognize the kind of machine they’re on, kind of processor they’re on, to take advantage of the special processor capabilities.
But this is designed to operate in a holistic, unified system, so you can begin to move your workloads from those older components, older parts, older sleds, onto the newer ones. Maybe you wanna remove those older sleds and replace them with newer ones. Maybe you want those older sleds to keep on running, because these things can run for a really long time.
Either way, it’s your choice, and you get to run with heterogeneous compute in the same rack. It’s a core part of the Oxide value proposition. Thanks, and see you next time.