What is a Hyperconverged?
Building an abstraction layer to separate the storage substrate from the hypervisor and application will be Nirvana for users – but Nirvana is a difficult place to reach
Nutanix in numbers:
Nutanix is on an annualised $300m run rate based on fiscal Q2 figures ended January 31 2015.
The firm says it has 1200 customers including 50 who have purchased north of $1m in products and services.
In five years it has raised $320m from five rounds of funding which started with VCs such as Lightspeed Venture Partners.
It employs 850 people.
Ambrose McNevin spoke with Dheerav Pandey, founder and CEO in London just before the latest figures were released to discuss the finances, the technology and the vision of the business.
Nutanix was founded in September 2009 by three individuals who wished to build a system that melded web scale engineering and consumer design.
"We set out to build web scale engineering which is all about running on x86 commodity servers but which would also be consumable by the masses, sellable by the channel, and manageable by college drop-out administrators as opposed to PhD qualified engineers, says Dheeraj Pandey, Founder and CEO"
Content from our partners
The company’s initial investors were VC Lightspeed Venture Partners, and then it graduated to investors such as Goldman Sachs and Morgan Stanley and strategic partners like SAP Ventures. In the last round it raised money from institutional investors such as Fidelity and Wellington Management.
CBR: What are the investor expectations? Is there an IPO end game?
DP: When Fidelity invests and when Wellington invests you know that they manage trillions of dollars so they are not in for something small.
Wellington manages 1 trillion and they are primarily public investors and only recently in the last three years have they shifted towards private companies – to the ones they believe will become market makers. We have a small chance of becoming an iconic infrastructure company, so we’ll see.
CBR: What scale you are operating at currently?
DP: A little over six months ago we released numbers which were real numbers and not just percentages and we were doing over $200m of annual run rate. [On Feb 4th Nutanix issued the figures quoted above of a $300m annualised run rate.] And we were growing at least 25% quarter over quarter. So things have been good. We haven’t released the latest numbers yet but we are way bigger than $200m."
CBR: What is the addressable market and hyper-converged market – is it about taking market share?
[In December IDC measured the hyperconverged market and said Nutanix had 52% market share]
DP: "Most of the time the addressable market and the TAM is a lagging indicator. Go back to 2004 and no-one would have guessed that VMware would have technology installed in 90% of machines installed in large enterprises. Same thing in 1999, you couldn’t believe that NetAPp would be in 70% of the enterprise storage market. Or back in 1996 that MS would be 40% of the data centre market.
It is a time and not a leading indicator. Look at SAAS in 2005. The incumbents would scoff at SAAS. Oracle and IBM would and say what cloud, we’ve been doing cloud for 20 years and its called on-demand business. So everything Salesforce says they do we can do also.
So, what makes a large TAM is the ambition of the vendors and innovators and pushing the envelope and use cases and workloads in vertical. Building solutions and talking a different language to the late majority other than the tech speak we speak to the early adopter geeks.
That shift will happen. But hyper-convergence could flatter to deceive if companies like us don’t keep pushing the envelope in workload terms and use cases and solutions.
It is a horizontal play. As long as the workloads are usable. We’re everywhere. As long as workloads are horizontal, we make a case for ease of use and time to market.
Virtualisation provided enough pain points that organisations had to learn the hard way that SANs are difficult and painful because we didn’t build infrastructure for virtualised workloads from the ground up. We tried to retrofit a SAN which was designed before the internet and use it for an extremely demanding and dynamic compute side of things because VMware made it that way.
See Next Page for details on growth plans
CBR: What is your relationship with the web scale players?
DP: We have 350 developers, 80% in San Jose, with others in Bangalore and Seattle. But what have we built? We are bringing web scale lessons learnt in Google, Facebook and Amazon and applying that to the enterprise.
Now, why does that require a lot of software engineering? You have to make sure that the apps of the enterprise don’t change. You have to make sure that VMware, Microsoft, Splunk, Avaya doesn’t change.
To build a software abstraction that is scale out – which is what the web scale guys do – to use a bunch of machines and to make it look like a single machine is a very difficult thing to do. To do that with the facade of a single system so for the rest of the system, VMware, HyperV etc is basically what we’ve done.
We have built all the goodness of EMC and Netapp into our software – all of it, plus more.
Any enterprise grade feature that you see in EMC or Netapp highest end arrays, we’ve brought them into pure software and made them work on multiple machines. So as opposed to a scale up Netapp or a scale up EMC VMAX. We have a scale out system. That has the same feature sets but the software makes it look like one system.
Now what’s the value. You can start small. You don’t have to speculate on capacity planning, you can pay as you grow. And the fabric keeps extending itself so you can realise cloud economics without having to buy expensive systems.
CBR: What are you doing in UK?
DP: We have a sales force in the UK and a lot of channel partners – Kelway is a big one for us and we have an OEM deal with Dell.
What are your growth expectations for 2015?
Everyone wants to look at the way Oracle or Salesforce grew – 100% YOY for five or six years. That’s what success really means. What we’ve done in three years of selling is also unprecendented. To keep our own bar will be difficult but it can be done.
We’re in act II of Nutanix. We built the data fabric and storage software. I think we have to build a platform out of Nutanix. That is not easy but we have to make the hypervisor the next commodity. Where our stuff runs on top of anything, HyperV, Vmware, ESX, KVM, and Amazon and Azure.
CBR: What’s your price model?
DP: If technology is to become mainstream it must commoditise. It is good enough it will be become the enemy of the great and that’s a good thing and then you dont’ have to evangelise as much. So commoditisation is actually a good thing. You have to educate less. But the downside is price erosion. We have a straight line to a billion dollars. But we have to be prepared for it.
