I’m particularly interested in the linkup with Pure Storage and Dell. So, what I gather is that in those situations where you use those as external storage, Nutanix becomes compute-only. So why, in that case, would people buy Nutanix at all when they could buy any compute?

First of all, I will just say it’s a little bit more than just compute-only. It’s compute plus networking, micro-segmentation and operations capability. So, the rest of the stack minus storage.

We didn’t do this in the past because most people have stayed on the VMware hypervisor, but now there is a lot of interest from customers that want an alternative to VMware. And if you think about what’s available out there, I think AHV [Nutanix’s Acropolis hypervisor] has emerged as one of the best options, and is why the storage providers are now interested in working with us.

And that’s why you heard from early access customers for the Dell option with PowerFlex. You saw Moody’s on stage yesterday. They’re interested in it, and they’ve been an early access customer. So, the landscape has changed around us.

What’s stopping you just opening Nutanix up to any external storage? In other words, what are the engineering hurdles?

When we did the Dell PowerFlex integration, that was more of a one-of-a-kind, because Dell PowerFlex is a unique scale-out array.

As we start doing Pure, there’s an opportunity to build a more standardised approach towards integrating third-party storage. I’d love for us to get to a point where we can just have a self-certification programme, through which we can onboard new storage platforms. We’re not quite there yet.

What, specifically, is the engineering involved?

We definitely looked at VMware. They had vVols as one of the ways of doing this. We don’t have an equivalent. We never focused on this in the past with our hypervisor, because it wasn’t a part of our platform. We weren’t doing third-party storage support.

So now, as we started doing it, Dell PowerFlex was kind of a unique thing, because that’s a unique array. That’s different from the rest. It’s scale out, not scale up, and it’s IP only. With the others, we want to take a more standardised approach, with Pure and then beyond Pure.

The link-up with Pure allows for quite an expansion in terms of scale, in terms of storage nodes, etc, so is this move in some ways an admission that Nutanix has been limited in terms of its ability to scale?

I don’t think this is an issue of scaling at all, because even yesterday we had Micron talk about how it has built fairly massive scale.

So, this is not about scaling, in as much as, if you look at the total market out there, we have been trying to eat into the external storage market for many years. And HCI across VMware and us – the two big players – has taken over maybe 20% of the market after all these years.

The pace of migration in these things is slow. So, there’s still 80% of the market out there on three-tier, not necessarily because HCI can’t scale, but because there’s a lot of inertia in the system. People just don’t move.

So, to me, the reason for doing this was more that other 80%, some of which we will continue to convert to HCI. For the rest, I think we can get in there with the compute-only platform, and then at some point, once we are in there, maybe someday they’ll start converting portions of it to HCI.