Nutanix's Mirani: The Real Scaling Will Happen Through Partners

Lorna Garey

Since going public in August, Nutanix has topped the 6,000-customer mark, including 57 of the Forbes Global 100, and is closing in on $1 billion in billing. As a reference point, when SVP of engineering Rajiv Mirani joined the company four years ago, he says it was driving, on an annualized basis, about $5 million in business. At the recent Nutanix NEXT conference, we sat down with Mirani to discuss how the company plans to increase adoption of its Acropolis hypervisor, as well as how it’s empowering CSPs and customers to move to an end-to-end cloud architecture.

The NEXT conference itself has grown during his tenure, from 900 at the first event in Miami to 2,200 last year in Vegas (we were there) to just about 4,000 last month. The conference was heavy on product and partnership announcements, most of them well-received by attendees. Highlights include an alliance with Google Cloud, announced by SVP Diane Greene, to enable a unified, full-stack cloud infrastructure; new Calm automation and life-cycle management software, based on Nutanix’s 2016 acquisition of Calm.io; Xi Cloud Services, which the company says will deliver a turnkey cloud service built from the same infrastructure stack with the same tooling and SLAs as the core Nutanix Enterprise Cloud Platform; and expanded partnerships with Veeam, HPE and other technology partners.

Nutanix’s Rijav Mirani

Of course, it’s easy to lay out an aggressive road map. Delivering is another matter.

“We are big believers in strategy — of course you need to have a strategy, and it needs to be coherent, and it needs to make sense,” said Mirani. “But then it’s going to all come down to execution, right? The world, at the end, is not an intellectual exercise.”

Here’s our Q&A with Mirani, edited for length and clarity.

Channel Partners: Out of all the announcements, which one excites you the most?

Rijav Mirani: There are some things that are definitely more strategic to the company, that from a customer standpoint, it gives them reassurance that when they look at Nutanix, it’s not a company whose technology they use just as point products; it’s a company they can really use to re-platform and rethink their infrastructure.

We have a lot of performance and availability improvements — these are things that customers will use literally tomorrow, as soon as the release goes out.

If you look at Calm, that’s a big piece, because they’ve been saying, “Look, you’ve simplified the infrastructure, but I need to get that last cloud infrastructure layer for automation and provisioning and then life-cycle management.”

We expect a lot of customers very quickly to start playing with it and then, depending on what they think, actually deploy it in production. And then when you look at …

… Xi, it will take us a while to get the service rolled out, and then once it comes out, it will depend on the quality of the service itself.

CP: One of the key things I’m picking up is that you’re pushing to make it easier for customers to use the Acropolis hypervisor.

RM: Yes. If you look at it from a customer standpoint, there are three things they worry about when they think, “How can I get onto AHV?”

First, it should be easy to migrate. That’s why you saw our extract product. We’re just making it simple. If you’re running on ESX we’ll very quickly be able to move those VMs over to Nutanix.

Now once you are on AHV, customers worry about things like, “How do I back up the environment?” And then it’s an ecosystem play. And with Rubrik, Comtrade and Commvault already in there, it’s a pretty healthy, mature ecosystem. Similarly, on the switching and on the [enterprise data center] side, we have our partnership with Citrix on NetScaler.

AHV has gotten to the point that, the core platform is, of course, mature, but more importantly, the onramp onto it, as well as the ecosystem around it, is equally mature.

CP: Most recent numbers are 23 percent AHV adoption as a percentage of nodes sold.

RM: Because of VMware’s success, and because of how successful their technology was, the hypervisor is the stickiest bit. It’s not easy for our customers to consider a move. Look at Microsoft with Hyper-V — after like 10 years, they haven’t made significant inroads.

I’ve yet to meet a customer who goes, “Look, ESX, vSphere, I love, but it’s kind of expensive. Hyper-V, I can sort of make do with it.”

With Nutanix and AHV, most of our customers say, “It’s better, I actually love the experience.” And we don’t charge anything separately for it.

CP: Microsoft also bundles Hyper-V. Why do you think AHV will win where it didn’t?

RM: I think our advantage is twofold. One is that we are a much younger company so we don’t have the burden of history on our shoulders that essentially says, “Look, we should not be doing this,” in some sense. In plain language, we are less afraid because there’s less to lose. And that means a lot.

The second thing is, in terms of the core technology and the fact that we have come up from the bottom of the stack, that helps in creating a seamless experience.

CP: Once you get upstream, Nutanix seems very agnostic in terms of public cloud.

RM: Yes — part of the reason there was a meeting of the minds with Google was that they have a pretty open philosophy, and we do as well. Our view is that we will build a full stack of infrastructure, but the choice of which portions of technology you use at different layers in the stack is completely on the customer.

So you can say, “I’m going to use Nutanix for this but something else here.” That’s fine. “I’m going to use Nutanix for these two and do something else over here?” That’s fine. “I’m going to use the full stack of Nutanix” – that’s fine as well. We are not going to say, “Here is this one right way and everyone must …

… jump onto the train exactly at the same time and at the same station.” That’s just not practical.

CP: How are you selling to the cloud service and colocation provider market?

RM: That’s the point we’re trying to make with Xi. But first we have to win the market’s respect with a truly reimagined hybrid service, and that we believe we can do only if we don’t outsource the hard problems to someone else. That’s the conviction and realization we have had in the eight years that the company has been around. It’s almost a showcasing of what is possible.

The real scaling is going to happen through partners, whether it’s folks like Google who will give us the global reach or our multiple MSP partners who are trying to morph into CSPs; a lot of them are using Nutanix today.

That’s why we’re taking great pains to make sure that whatever’s running within our data centers, we can package the software and make it available to our cloud service partners as well. Then they can run whatever Nutanix they are running today and then at some point, if they want some of the operational efficiencies that can come from standing up their version of Xi as a service, they can do that as well.

CP: What problems do you see MSPs have when they try to become CSPs?

RM: If you look at AWS, there’s a difference between what does ASW have and what does a cloud regional CSP have. The CSP has the customer relationships — they know their customers, they know their kits, all of that. AWS is this abstract thing, behemoth, sitting somewhere.

What does AWS have? The services, they’re experts.

If you look at the CSPs, what technology do they have? They’re trying to cobble together the usual, maybe they have an EMC array, are running VMware. They’re trying to make do with whatever they can.

In our view, when the service gap is small and manageable, then those other things win out — the fact that, “Hey, I know the customer” and the customer feels, “Hey, this team really understands my specific problem.”

If the service gap is huge, it’s just insurmountable.

So right now, the problem that the CSPs are having is they know their customers, and their customers are telling them, “I want to go to the cloud.” The customers are also telling them, “I want it to be you.”

Because look, if I’m somewhere in Texas, I’m like the millionth AWS customer, and if something goes wrong … well. Whereas if I’m your 100th customer, that’s a better spot. But right now the hard truth is, if they go to AWS, the service experience is, boom, you can get going in like 3 minutes.

For the CSP, there’s a big gap.

What we are hoping is to level the technology. We’ll go to the CSPs and say, “Here’s our stack, take it,” and now the gap between AWS and the CSP has been erased. And now again, their real advantage – the relationships that they have, their understanding of …

… the local customer and their problems – that will again become supreme and dominant.

CP: There’s also a financial hurdle. Can you talk about PayGo and ELAs?

RM: Let me start with the [enterprise licensing agreements]. It’s all about choice. There are some customers who are buying five nodes, 10 nodes, 20 nodes — for them, there is no value in bringing a do-it-yourself approach; you’d much rather get an appliance.

In some other cases, it’s like, “I’m going to be deploying Nutanix on 600 nodes, and I want to let the hardware vendors bid it out.’ You want two players in there. [Ed. note: Nutanix software runs on Dell, Lenovo, IBM, HPE and Cisco hardware.] That’s why there’s a lot of focus on having a software-only form factor. And as part of that, you’re going to have ELAs. Typically, when we sell software only, it’s sold as a subscription license on a node, but there are larger customers who say, ‘Look, I’m going to be buying 500 nodes over three years. I don’t know exactly when, but I just want to cut down on the number of negotiations and the number of transactions that I’m doing with Nutanix.’

We will support that. You can say, ‘Here’s your pricing schedule that’s locked down.” We negotiate and it’s there, and then you keep consuming towards the ELA over the next three years.

On the PayGo side, we have this program called Nutanix Go, which essentially, on an appliance form factor, brings cloud consumption onto on-premises infrastructure. Say you need infrastructure worth 16 nodes of Nutanix, but for some reason, there is uncertainty around the project — maybe you’re starting a new business initiative. You can say, “I will rent those 16 nodes for a year.” At the end of one year, if, for whatever reason, your business didn’t work out, you get to return them to Nutanix. Or at the end of one year, if you’re growing like gangbusters, you can add 20 more.

So it’s kind of like renting things out, much like you do in the cloud, for on-premises infrastructure.

Choice is paramount for the customer. Whether it’s choice of form factor or choice of consumption model, we’re making sure that customers can run on-prem or cloud. Will you run appliance or software? Whether you pay the traditional CapEx fee or you want a pay-by-the-drink kind of model, that’s your choice.

CP: The micro-egmentation capability is a key point for CSPs as well, something that could cost more.

[Editor’s Note: The Acropolis Microsegmentation Services inspect, monitor and govern communication flows between individual workloads. CSPs hosting multiple customers can define policies in the Nutanix Prism management system.]

RM: I go and talk to a lot of customers, and some of them are looking into VMware’s NSX product, and it’s a good product. Perhaps, as a lot of things that come out in the Valley, it’s a bit more over-engineered than it needs to be, but it’s a good product. One of the big gripes I hear from customers is that it’s also quite expensive.

In the end, if you’re on the full stack of Nutanix infrastructure, it’s not like for each little brick that you put up in the building, there’s a $10 tag associated with it. That is important because …

… incremental added costs force our customers to make suboptimal tradeoffs.

CP: Let’s go back to other announcements. There was lots of excitement around the ability to offer automatic failover as well as shorter RPOs.

RM: We had customers saying, “For me to move my mission-critical applications, there needs to be a shorter RPO.” This was probably the most asked-for feature, if you will, for a long time. So I’m glad we are able to bring it down, not just to 15 minutes but to 1 minute RPOs. Right now, it’s not just the classic physical catastrophe like an earthquake that takes your data center out. Much, much more likely is going to be some sort of attack that takes your data center or at least some portion of it, applications, down.

CP: Yes, ransomware is so prevalent.

RM: That’s right. I’m digressing, but security as a technology is unique, it always keeps evolving. Security is the only domain where there are good guys who are trying to use the technology in its given state to protect, and there are bad guys who have access to the same technology and they’re trying to get in. That’s the only piece where, no matter what the technology is, you know it’s always going to be a cat-and-mouse game.

CP: We’re short on time, but can you touch on X-Ray?

[Editor’s Note: Nutanix X-Ray is an automated benchmarking tool that enables IT or a CSP to continually evaluate the performance and availability of applications and infrastructure under real-world conditions.]

RM: X-Ray is our effort to make sure we are an honest company. This is something that’s a personal pet peeve of mine. With a lot of technology companies, the marketing is so over the top you can’t tell who does what any more. You go to the web site and everyone starts with, “We are making a better world.”

I get it, but can I get something below that?

X-Ray is a tool that lets customers benchmark Nutanix against other vendors. It’s not just the classic speeds and feeds. This is what I think is the ultimate head fake: How many customers’ genuine problem is that ‘I’m at 400 microseconds or 453 microseconds of latency?’ That’s not their problem. Their problem is much, much, much bigger, about larger trends that are happening.

And if you go to the vendor, the benchmarking tools are ‘Oh, you’re at 452, I came in at 420.’

It’s an attempt to refocus attention. Yes, performance is important, but it’s one dimension of several that are equally important.

CP: X-Ray is a good selling tool for your partners.

RM: Yes. As in any sort of technology shift, how do you separate the signal from the noise?

It’s an attempt to tell the customer, “Look, we want to be very open, much like we have been with our technology with the Nutanix label. Everything’s open, we’re going to tell you exactly what the tool does, and we’ll let you be the ultimate arbiter.”