Converged infrastructure is attracting a lot of attention in industry circles as the key means to reducing IT complexity in the data centre. But how exactly does infrastructure ‘converge’, how do various converged offerings differentiate and to what purpose? To answer these questions and more, InsightaaS spoke with Anton Granic, director of Nutanix’s Canadian business operations and 17 year IT veteran with a background that spans networking and security, and now data centre and virtualization. Granic is a practiced hand at bringing disruptive technology to market: a half decade ago, he introduced next generation firewalls from Palo Alto Networks to Canada, and is now working on developing the Nutanix profile north of the border.
According to Granic, like Palo Alto Networks, Nutanix is experiencing “tremendous success” because the solution “solves a problem” – it provides hyper converged infrastructure built on the integration of server and storage resources which allows companies to scale. If this description of the Nutanix value proposition appears impossibly simple, it is – but the company’s success is not questionable. Based on the growing popularity of its web-scale solution, last August the California-based Nutanix announced an additional funding round of $140 million to bring company value to $2 billion, approximately five years post launch. Nutanix now boasts 800 customers in the mid-range and enterprise segments, has built out its channel program and established an OEM relationship with Dell for the integration of its Virtual Computing Platform into the Dell XC appliance, and this March introduced a new technology certification program that supports growing interest in hyper converged infrastructure.
In the edited version of a conversation with InsightaaS managing editor Mary Allen which follows below, Granic provides a keen description of converged infrastructure benefits and Nutanix disruption, as well as an update on the company’s advance into Canada.
Anton Granic, director of Nutanix’s Canadian business operations: When Nutanix approached expanding into Canada just over two years ago, I was asked to come on board and help build Canada literally from scratch. You can envision Canada at that time as a very, very traditional, in some cases legacy market. With all of the incumbents – Cisco and HP and IBM and Lenovo and all the server vendors, and all the traditional storage vendors like EMC and NetApp and HP – the market itself had become fairly stagnant in terms of innovation.
When I came to Canada with my systems engineering director Michael Burnett, our task was to build up Nutanix Canada from scratch. At that time we didn’t have any customers; we didn’t have any partners; we didn’t have any brand awareness. All we had was a story, and a very unique, innovative, disruptive technology. Since then, we’ve managed to build out a business that now spans across the country. We have a national presence, with teams in BC, Alberta, Ontario and Quebec. We’re fully staffed with sales and marketing, and what we’ve managed to do, in a very short period of time, is bring on board over 130 enterprise customers, all of whom are now able to leverage the goodness of what we call web scale hyperconvergence.
Mary Allen, managing editor, InsightaaS: Can you provide a bit more detail on “national presence”? How many people do you have?
Granic: We’ve got over 20 people dedicated to Canada – some of them are located in Canada, some are corporate-based out of the US, but in terms of the actual team, it’s close to 20 people. And in parallel, we have a national channel strategy, made up of what we would call national channel partners, regional channel partners and local channel partners. We have spent a lot of time educating the reseller community in Canada, and the technical consultants and evangelists, on how this next generation platform can help modernise the data centres. So we have our Nutanix team and our extended team, which would be our go-to channel partners.
Allen: I was assuming that these 20 would largely perform a sales function as a lot of Canadian operations of US-based firms are basically sales offices. But you are working 100 percent through channels?
Granic: We are. In fact, it’s one of the very unique things that we bring to market. We are 100 percent channel, and we have dedicated sales and technical resources who work with our channel partners to enable them to bring the technology to the end user.
Allen: Can you give me an example of what the ideal channel partner for disruptive technology in a more conservative, or traditional market would be? What did you look for when you were building those relationships?
Granic: I think too many times in this industry people abuse the term value-added reseller, or partner – there are a lot of organisations out there that will pretty much resell anything and everything. From our perspective, what we were looking for specifically were resellers that first of all had a very strong technical understanding and presence in the data centre. So resources that were well-versed in servers, storage, and virtualization, and not only from a sales perspective and a technical perspective, but also from a support perspective.
I think Canadian customers very much want to work with local resources and contacts who they can trust on a day-to-day basis to be there when they need them. For us, it really came down to focus on organisations that see themselves as either innovative or disruptive or as a trusted advisor. I hesitate to use the term ‘trusted advisor’ because everybody uses it, and it gets somewhat cheapened by the repetition, but I would say that the partners that are ideal for us as the ones that are thought leaders within their customer base. That’s the best way to describe it.
Allen: I see you now have a certification program. Would you require your partners to be certified on your solution?
Granic: We do, yes. In fact, we have different levels of certification, including sales and administrative certifications. And we now have a certification that is considered one of the most demanding in the industry, which is known as the NPX, which we’re looking to make a very, very rigorous process for people that are very well versed in all aspects of hyper convergence. Certifications for our partners are crucial as it’s a very technical sale. This is something that we feel is a unique offering in the market; certifying our partners ensures that our customers get the best resources when it comes time to implement this solution.
Allen: Where have you been finding your best success in Canada? What kinds of customers, in what verticals, or what kinds of companies?
Granic: Here’s a real testament to the beauty of the solution. We don’t do better in any specific vertical than another. For us, every vertical, every customer, small or large, benefits from our webscale platform. What I mean by this is that today, everybody is trying to leverage the benefits of virtualization, even though ten years ago when people were shown the prospects of being able to virtualise servers or applications, they were very hesitant. It was a very foreign concept, they were sceptical, and it took a sort of gradual easing into a comfort level where the organization could run production workloads virtually.
With Nutanix, we’ve seen some very similar discussions happening, where this is something that’s so radically different from what or been doing for the last ten years, that businesses typically want to start with something small. And similar in that we are a purpose built virtual computing platform that can run any virtual workload, so every organisation in Canada that’s running some type of virtualization is a prime customer for us.
We’ve had tremendous success in the financials, in education, in healthcare, in retail and in manufacturing. There’s really been no single vertical that has done better than others because they all face the same challenges around the data centre – it has become too complex, too costly, and there’s absolutely no innovation that’s allowing people to maximize the benefits of virtualization. The majority of workloads or infrastructure that people are running today weren’t actually designed for virtualization. Nutanix has had an amazing explosion into the market because we’ve identified that the legacy approach doesn’t work, and we’ve basically changed the game altogether. We’re very fortunate that all verticals and customers of all sizes can benefit from our hyperconverged platform.
Allen: Do you have different offerings for an SMB, for example, and an enterprise customer?
Granic: We do. And that also ties into your question around which verticals do better. What we see is that every customer wants to do more with less, and ultimately that’s what Nutanix provides. But we have multiple offerings that range from very small workloads to medium workloads to large and very demanding mission critical workloads, and all of those different platforms that we have can actually interoperate within the same ecosystem. So it’s a very flexible solution that allows people to basically run any sized workload on our platform.
Allen: What exactly are the key principles behind the hyper converged platform you offer? Are you a storage platform or a server virtualization platform? Where you sit in the stack might be a more helpful way to look at it.
Granic: That is the multi-billion dollar question. So if you look at the traditional approach, and how the stack is set up today, it’s very siloed. In most organizations, you typically have a server team, a networking team, and a storage team, and in some instances, even a separate virtualization team. In this architecture, if you can visualize this three tier architecture of servers, networking and storage, it’s a very complex mess of connectivity, process and change control – everything that you would imagine and see on a daily basis in the enterprise.
What we’ve done is we have completely blown that model, or that architecture, apart. And we have collapsed that three tier architecture into a 2U purpose built virtual computing platform. What that means is, we’ve actually put compute and storage together in a single application, and all of our networking goes to a ten gig top of rack switch, as opposed to having to traverse across an entire network with multiple hops.
So what does that mean? First of all, all of our principles, everything we do, has its origins and DNA in the big web scale companies. This is very important because multiple players that are now starting to come to market are saying “we can do what Nutanix does, we’re hyperconverged too.” But the reality is, technically, there’s nobody that can do what we do, because our origins come from the likes of Google and Facebook and Amazon.
When, for example, you go into a data centre at Google, you don’t see Cisco, HP, IBM, Dell, Lenovo, EMC, NetApp, Hitachi, or any of those vendors. The reason is, they’re very complex, they don’t perform well, and they can’t scale linearly. There’s an architectural flaw. On top of that, it’s incredibly expensive. So Google actually blew away the concept of a SAN, or a Storage Area Network, and they brought the storage local to the computer; they basically created their own data centre with X86 servers, basically commodity, enterprise-grade X86 servers, direct attached storage, and special software known as the Google File System, to create a virtual storage pool.
In essence, that storage pool replaced the SAN. At our core, this is what we’re doing in the data centre. The team that designed the Google File System actually left Google and founded Nutanix. Their reasoning was that if Google has a challenge around performance and scale and rack space and power and cooling, there’s going to come a point where every organisation, small, medium or large, has that same challenge. You can’t continue to throw more hardware at the problem, you have to create something that’s much more efficient – something that in essence is software-defined. And that is how Nutanix was created, and that is how the Nutanix Distributed File System was created, and that distributed file system, which is our proprietary IT, that’s our software, runs on a 2U appliance that has direct attached storage with computing built in.
So the efficiencies that we bring, the time to value that we bring, the performance that we bring, the scale that we bring are key, and most importantly, if you think about it operationally, we’ve drastically simplified the data centre. Because today, organisations need one, two or three different management consults to manage all the different pieces of that stack. With Nutanix you have a single dashboard managing the compute, the storage, and the virtualization layer. And that has dramatically changed the way people look at building up their data centres.
Allen: Okay, so you have software that is your IP, which you build into an appliance that has storage and the compute capacity. What happens when you go into customer environments that have installed hardware, and it’s not end-of-life for this hardware? How do you deal with those heterogeneous environments, which certainly in Canada I would think are most common?
Granic: This is another great question. There are a couple of ways that we deal with that. First of all, what we’re finding in the majority organizations that we’re engaged with, is that there are multiple different vendor solutions running in their data centres. Over the years, they’ve collected a hodgepodge collection of different server vendors or storage vendors and management consoles, all of which have become very difficult to manage.
In those instances, we find that customers are looking to try and simplify and eliminate the multiple touch points or choke points or support contacts that they need to basically run their infrastructure. Where organisations have completely refreshed their hardware, and they’re looking at not making any changes for the next three to five years, the conversations we typically have are around the new workloads or the new data moving forward.
What ends up happening is a couple of things. Every day we walk into companies across Canada who have a specific project, which may be to virtualize desktops, SharePoint, SQL, Exchange, Oracle, Big Data, or virtualize a server. Any one of those six or seven different use cases are all applicable to Nutanix strengths. What we say in those scenarios is, “give us a chance to show you how our web scale hyperconverged platform can get up and running significantly faster than your existing infrastructure, and at that point simplify your operations, with the single dashboard.” In the cases where there’s a specific project, we actually come in and ask for the opportunity, and customers that have grasped this potential are able to say “we’ve heard great things about this, let’s see if this is an easier way of doing things.” And in those cases, we apply very laser-like focus on the specific project, or whatever needs to be delivered.
From a delivery perspective, this is very important. Traditional infrastructure can take weeks or months to deploy due to the challenge of having to coordinate this hodgepodge of different solutions: you have to talk to the server team, you have to talk to the storage team, you’ve got to get the virtualization team involved, and you have to commission the change request. This can take anywhere from three, four weeks to three, four months.
With Nutanix, we’re on site in ten business days, and up and running in a matter of hours. That time to value is actually a web scale principle – that time to value is incredibly important for organizations that are trying to provision workloads which in essence drive their business. So if we can reduce the time for a project from six months to ten days, our value is very tangible. When it’s project specific, it’s very easy for us.
If we walk into the scenario where it’s a full blown environment with a specific vendor and that’s all they deal with, typically what we do is augment that existing infrastructure. We’re very flexible, and able to insert ourselves into existing environments as an augmentation or add on, and conversely, we’re able to deploy very specifically to a project or a workload to ensure that the customer can get this up and running without any impact to business timelines, or to performance or the end user experience.
Allen: How does your software work with an operating system that might be on the server? Also, when some people talk about converged infrastructure, they include the networking component too, but you are simply storage and compute. What kind of challenges are there for you, or for your customers, when they need to introduce your appliances into their networking infrastructure?
Granic: One of the things that we really pride ours on is the ability to give the customer choice and flexibility. The networking piece is a part of that, as is the hypervisor component. I’ll get to the operating system in a second, but we are the only hyper converged platform that supports three different hypervisors – VMware, HyperV from Microsoft and KVM, and we absolutely believe that the customer should not be locked in to a hypervisor. This allows the customer to make technical decisions based on choice, and based on what’s best for their business.
The same principle applies to the networking component in that we are very accommodating to any networking fabric that is out there, whether it’s Juniper or Cisco or Brocade or Arista. We understand that most customers have already heavily invested in their networking infrastructure and it would be presumptuous of us to say here is the networking component that you have to use for this hyper converged platform. We preach simplicity and flexibility.
From a networking perspective, we offer the customer choice of any networking platform that allows them to run, for example, a ten gig Ethernet network. We’ve found is that this is actually being very well received by our customers. They always ask us for recommendations, and we have multiple technology partners that we work with that are also disrupting the space. There’s a lot of focus on software defined networking, and for us, being able to connect to the different offerings out there is important since we’re a software defined storage solution. We actually have multiple patents within the software defined storage space that revolve around our IP and data locality.
On the operating system side of the house, we absolutely own the entire stack. The Nutanix operating system is the core engine that runs the hardware that we provision. Everything is X86 architecture, and we have our Nutanix operating system, which allows the customer to basically get a data centre in a box. This system is quality assured, it’s hardened, and it comes with all the enterprise storage features that you would get today from your existing legacy storage providers, but with all the added benefits of improved performance, scalability, ease of use, ease of management as well.
For us, this was a strategic decision when we went to market. If we went out as a software vendor only, and we tried to get our software certified on all the different hardware platforms out there, we would have actually been out of business before we got out of the gate, because there’s way too many variables and configurations that are out there.
Some of our competitors actually tried to do that, and they walked into a world of complications. Since we are able to support the entire stack, when you call 1 800 Nutanix, our level three engineers can troubleshoot the operating system, the hardware, the storage, the compute, as well as the virtualization aspects of that stack. This is very critical for customers who want a single throat to choke – and has led to our having the highest net promoter score in the industry.
If you look at all the players that are in this space, whether it’s on the server side, or the storage side, or even the virtualization players, Nutanix has by far the highest net promoter score. It’s actual 88, a testament to our approach, and our diligence with engineering and Q&A.
Allen: The converged infrastructure space itself, I think, has a lot of variation. As I see it, there’s a scale of providers, where on one end you incumbents like HP, which has been talking about converged infrastructure for a very long time, and on the other you have a company like SimpliVity, which seems to be doing something a little more like what you’re doing. So if a customer asked you, “how are you different from HP? How are you different from SimpliVity?” What would you say to them?
Granic: I think you have to look at convergence in two lights. This is a very interesting time because marketing terms are starting to dominate the discussion. Convergence from the traditional players, whether it’s HP, Cisco, IBM, EMC or whoever, is not true convergence in the sense that they’re re-architecting their solutions or building something new. All they’ve done is taken that three tiered architecture that’s been around for 10 to 15 years, and repackaged and pre-packaged it so they can present it to the customer in a single chassis. Architecturally speaking, they still have the same challenges that a traditional three tier architecture had.
The best analogy I can give you to show how Nutanix is different from the traditional converged players, is for you to think of what HP and others do as ‘Convergence 1.0’. It was a very good theory: vendors said, “Hey Mr. Customer, instead of buying all of these components separately, I'm going to pre-package it for you. I'm going to pre-engineer it for you in my lab or in my factory, and I'm going to send it to you in one big chassis, and all you have to do is call 1 800 whoever.” The problem with that is that under the covers, you still typically had one or two or three different vendors comprising that stack. So support became an issue.
In the case of the vendors who did both the servers and the storage, the same architectural flaws apply, which is they’re running an infrastructure that isn’t software defined, and they’re crippled by what’s known as a dual controller architecture. If you think about that, it’s kind of like having a two lane highway on the QEW and you keep adding more cars onto the highway. If one of those lanes shuts down, architecturally speaking, you’re going to have a traffic jam. This is the traditional issue with scale and performance and SANs in Convergence 1.0. When you take into account the fact that it’s pre-configured, (and there is very little flexibility to the type of configuration you can do), and that it takes anywhere from 45 to 60 days to deliver, the value that was touted upfront really doesn’t shine through.
So in Convergence 1.0, all the traditional players are taking their ten year old Motorola flip phone, their ten year old camera, their ten year old pager, their ten year old compass, their ten year old calculators, spreading them out on the kitchen table and putting them all together in a nice little pile, and duck taping them together to say, “here’s your smart phone.” What Nutanix has done is brought the iPhone to the data centre. Everything is done within software, under the covers, transparently to the end user. Fundamentally, that is the difference between Nutanix and everybody else that’s out there. There are other players – you asked about SimpliVity and others – and the big difference is that the majority of these players are dependent on hardware. Being a software-based solution with a distributed file system means that Nutanix is able to participate in all verticals and work with all workloads.
Many of these players, who are now coming to the hyper converged space, are experiencing major challenges with scale and performance because they’re dependent on hardware. But since Nutanix is able to scale linearly – basically to infinity – with our distributed file system, we maintain a huge advantage over all the newcomers in this space.
The reality is, companies like Simplivity are actually older than Nutanix, but they’re one fifth the size, or one tenth the size. The reason is they’re much more focused on the SMB market and they’re dependent on hardware. It doesn’t matter who the different players are: the traditional players are challenged by their dependency on hardware and legacy architectures, while the new players lack a truly software defined solution or a distributed file system, or the ability to support multiple hypervisors, and this makes them very limited in what they can achieve.
Allen: I’m thinking about scale and the high levels of efficiency in operations that you would bring, and about the potential for your appliances to introduce better space utilization in the data centre. These are attributes that are often associated with cloud computing. How would you describe the Nutanix solution in that context? Is it the building block for a private cloud, or would you have some public cloud customers?
Granic: If you take a look why customers are thinking about going to the cloud, it’s very simple. What is appealing about the cloud? Speed, performance, agility, and theoretically, lower cost of ownership, right?
The user says, “I can’t wait for my server team or my SAN team to provision their stuff. I'm going to take out my credit card, swipe it, and I’m just going to put this workload up in Amazon. Amazon is easy, it’s quick, it’s scale.” Now, in theory, that’s wonderful. In reality, Canadian customers, at least from a basic compliance and security perspective, can’t afford to simply send their data to the cloud without really understanding what that means. Where does that data reside? Where is the data centre located? Who has access to that information?
All of the goodness of the cloud – which, by the way, isn’t as cost effective as everybody thinks – is very appealing because traditional architectures have crippled their ability to innovate or deliver services to their internal resources or their customers.
By taking the principles of web scale, what Nutanix has done is bring the goodness of Google and Facebook and Amazon to your data centre. You hit the nail on the head: we are a private cloud-in-a-box. We bring the cloud ease of use, agility, speed, scale, and lower cost of ownership to the customer’s data centre floor, where the organization can actually control the data and manage where its data resides.
The beauty of the cloud is that there definitely are instances where putting data up in the cloud makes sense. Tier three applications, test/dev, cold storage, all of those different scenarios have their benefits, and what Nutanix has recognized is that it’s not going to be all cloud, and we know that it’s not going to be all on premises. There will be a mix of customers who are going to manage their own private cloud because we’ve made it easy for them, and those same customers will want to burst to the cloud to try and save money or space.
What Nutanix has done is created a mechanism for these customers to connect to cloud providers such as Microsoft Azure or Amazon S3 so they get access to this hybrid cloud best of both worlds experience. Again, this is all managed through a single dashboard that resides within the data centre. We firmly believe that the cloud is absolutely the direction that everybody is going to be looking at. By being able to bring the cloud to the customer, we think that we’ve really eliminated a lot of the uncertainty around where that data resides, the performance, the compliance, and obviously the cost issues. So we’re very excited about being able to leverage the best of both worlds.
Going back to our mantra of doing things uncompromisingly simple, we really believe that giving the customer the choice to say this is how we want to roll out our private cloud, this is the hypervisor we want to use, this is the cloud provider we want to connect to, and this is the workload that we want to run –is very critical, and I think it’s going to be one of our biggest differentiators moving forward. And as we continue to evolve and innovate, what we have coming out in the next three or four months is probably going to be just as impactful on the industry as we were four years ago when we first came to market and started shipping.