Exclusive - Nutanix CEO Dheeraj Pandey on post-legacy IT: "Never buy big stuff again"
- Summary:
- After a meaty keynote that covered a range of news announcements - including the Nutanix Xtreme Computing Platform (XCP) and their own hypervisor, I had a sit down with Nutanix CEO Dheeraj Pandey to get his take on how Nutanix's rapid evolution fits in with the priorities of today's CIO.
Nutanix customer stories - 75 hours to go-live?
Jon Reed: One problem I see for Nutanix is that your customer stories sound too good to be true. For example, when I first heard that Computer Services Corporation completed a Nutanix project in 75 minutes, I assumed that was a pilot project, or an empty storage farm yet to be populated. But on stage today, Karen Zetes [Director of infrastructure for CSC] explained that it was a major project - they moved about 275 VDIs and 25TBs of storage in 75 minutes.
Dheeraj Pandey: Yes, and it was up and running in 64 hrs, from the first phone call to the initial go-live. And she's taking it further. She thinks eventually the entire IT infrastructure can now be invisible. And by the way, that was only one block, she bought 7 or 8 more blocks soon after, and those were live in a few weeks.
On Nutanix going public - a firm date has not been set
Reed: During your interview with my diginomica colleague Den Howlett about a year and a half ago, you said something about a 12-18 month IPO time frame. I realize those timelines are moving targets. Are are we getting close, or are we still a ways off?
Pandey: We are not what I would call a ways off, but we don't really have a firm date. The process takes time. If you think of Box, three years before they went public, they were doing rehearsals. I think it is advisable for anybody going public not to put a date. But at the same time, we know that getting deeper in the enterprise requires being a healthy public company, so we are constantly thinking about it and planning for that day.
Why the new hypervisor wasn't the lead story at .NEXT
Reed: When you talked with Den a year and a half ago, you were almost waxing philosophically about SAP running on Nutanix someday, and that's a reality today. It's no longer about what can or can't run. It's really more about helping folks to understand the implications of such an agnostic framework, where they can run the VMware or Microsoft hypervisor within your platform - you don't care. Some folks were surprised you didn't go into verbal attack mode on VMware when you announced your own hypervisor today, but it seems like you are trying to change the infrastructure conversation entirely.
Pandey: Less than half a percent of people tweeted saying "You guys are going negative on VMware." But most of the people said what you did, that, "Hey, they didn’t say anything about VMware."
Reed: Coming into the .NEXT event, the industry buzz was, "They are going to announce their new hypervisor." The way I interpret your show is: you're glad to have your hypervisor in the market, but that's not the point of the show. You want to change how people think about the administration of infrastructure.
Pandey: Yes - because otherwise we fall into the same trap, which is sort of a non-sequitor. What if, all of a sudden, we say the hypervisor is less important, but our hypervisor is the most important piece of software that we have? You can't have both these things together.
Resolving the agnostic contradiction - why Nutanix provides their own hardware
Reed: The only thing that appears a bit contradictory: your strength is your agnosticism, but you've held on to the hardware piece pretty strong, although you do offer Dell as a hardware partner option.
Pandey: We do some soft relays, but at the end of the day, we do what the market asks us to do. Most of the market is actually pretty happy with appliances, and we would be foolish to go against what the market is asking. The good thing is that we are ready for any form factor. We are ready for an Amazon form factor, and an Azure form factor.
Reed: So you don't have customers saying "We want to buy/use our own hardware, and just use your software."
Pandey: Very few. But we’re very wary of talking in that language. As opposed to saying, "Let's change the language." Over time, change is the only constant. We have to keep changing the way we look at consumption models.
Nutanix and the CIO of the new stack: "Never buy big stuff again"
Reed: There's something deeper afoot though, because your products are speak to a fundamental change in IT. It seems to me that you're trying to address the business relevance of IT, because if folks aren't squandering their time managing storage and virtual machines, then something is going to give. Some companies might downsize their IT teams. But I think more likely, they are going to figure out how to apply their resources to new, business-relevant initiatives, like building new apps.
Pandey: Somebody was asking me recently, "If you guys abstract so much, will IT people have less to do?" The answer is no. It's like saying Ruby developers who build a shopping cart in one line of code have less to do. True, they don't have much coding to do as a C developer, who would write a commerce app in 50 times more code than the Python or Ruby developer. But they're all working just as hard - it's just that you're working in different abstractions. And the fanout is different. All of a sudden, you make specialists into generalists. Because they start doing more things across different boundaries, and the boundaries will actually fuse.
Reed: So what would you say to a CIO who is struggling with managing the complexities of their environment on the one hand - which is probably a mix of legacy and cloud - and on the other hand juggling the pressure from the lines of business to serve up relevant projects and applications?
Pandey: There's two or three different pain points that I talk about. One is: get speculation out of your procurement behavior. Never buy big stuff again. Always buy small, and then experiment more. The earlier you fail, the better it is for you. So you need to do more experiments and fail early. This is the reason why you can't commit to hardware too early - especially big hardware too early. Buy small, and then let things prove out and add more. Pilot is production in this new world.
It’s not like Amazon pilot is different than Amazon production - it’s the same thing. The other one is that you have too many islands of people that do specialist work, when you should really have generalists. We need to bring everyone together, and figure out what it means to solve problems that can be upleveled by using analytics and visualization and scripting and automation. It's not just being unified on the technology and products side - the people side is important as well.
Reed: Your view of the ongoing relevance of on-premise data centers, combined with your plans to incorporate Amazon, Azure, and other public cloud instances into your infrastructure, is an interesting twist on the public cloud message. You're saying that when it comes to infrastructure, companies are not all-in on the public cloud. It's not the "cloud first" message some might expect.
Pandey: Yes, it's more nuanced than that. Look - even the NoSQL companies like Cloudera and Cassandra - all these paradigms finally embraced their enemy, SQL. You need to make it easy for legacy to come along with you, instead of scoffing at legacy. The pendulum always rests in the middle. If you look at the 90s outsourcing fad prior to Y2K, the pendulum swung to one side, but it rested in the middle.
Image credits: All photos by Jon Reed.
Disclosure: Nutanix covered the bulk of my travel expenses for my attendance of the .NEXT show.