Nutanix SVP: Edge Computing, Not Public Cloud, ‘The Center Of Gravity’

‘We need to make it equally, if not more convenient, for people to create applications on the edge than in the cloud,’ says Satyam Vaghani, SVP and GM of Nutanix’s IoT and AI business in an interview with CRN.

Nutanix AI and IoT Leader On ‘Pulling Away’ From Edge Competitors

Nutanix’s innovation engine at the edge is roaring as edge computing begins to position itself as the primary point of presence for enterprises, not the public cloud, says the company’s artificial intelligence and Internet of Things leader Satyam Vaghani.

“Public cloud vendors think that the public cloud is the center of gravity. Our thinking is the reverse,” said Vaghani, senior vice president and general manager for Nutanix’s IoT and AI business in an interview with CRN. “We think the edge is going to be so important that you need to treat it as a first-class engineering problem, not just as a secondary thing you do among many other things. … We need to make it equally, if not more convenient, for people to create applications on the edge than in the cloud.”

The San Jose, Calif.-based hyper-converged infrastructure pioneer has transformed itself over the past two years into a software and application specialist with a heavy focus of R&D at the edge, which has sprouting innovations such its Xi IoT platform. On Tuesday, Nutanix unveiled a new technology partnership with consulting and software firm Hardis Group to provide a new solution to boost customers cognitive services, edge computing, IoT and machine learning capabilities.

Vaghani talks to CRN about Nutanix’s market differentiation at the edge, its new partnership with Hardis Group and how IoT will enable channel partners to “become the center of the universe” to their customers.

How does Nutanix’s edge strategy stack up against competitors like Dell and HPE as well as public cloud providers?

There are many vendors whose edge computing model is providing new types of hardware form factors for the edge. We think that needs to happen, but just like how it’s been proven in the private and public cloud, the key differentiation is going to be on the software side. That answers to a big bucket of people who are focusing on hardware innovation at the edge. We are different from the edge. We are a layer above. The second big bucket like hyper-converged infrastructure (HCI) vendors, including Nutanix, you could make a point that you can run HCI at the edge and that makes it all very simple. You don’t need to buy SANs on the edge and switches and all that nonsense -- that’s true -- but unfortunately the edge is not just virtual machines. It is also a whole new generation of micro-services-based applications. So applications that are on AI, containers and so on. You need to answer for both. You need to answer with HCI, that’s a starting point, but you also need to have a story on the other part. So that’s where I see us pulling away. The distance, in my biased opinion, is pretty big, because we answer to both sides. We have an HCI play on the edge but then we throw Xi Edge and Xi IoT. Now we have this incredible edge PaaS (Platform as-a-Service) on the edge as well.

What is another major technology differentiation for Nutanix that other edge vendors don’t have?

For the edge, you can’t just say, ‘Create this like another data center.’ You just can’t do that. You can’t deploy IT professionals here. We’ve done a lot of innovation on the management side of things and security side of things. Here are two examples. One is we created Prism Central which is literally from one central place, you can manage all these HCI clusters spread across 100 oil rigs. In fact, some of our customers do exactly that. Or you can manage all these HCI clusters spread across 100 retail stores. That’s where the differences become even more pronounced, because the manageability of the edge is a very big problem.

Security is another thing. If you drill down into the Xi IoT technology stack, the Xi Edge appliance that is powering some of these applications -- including the Hardis Group application -- is running on an absolutely lock-down appliance. The appliance is so locked-down that there is no administrator password. In fact, even for administrators, you are not allowed to login. Because the idea is that if you are going to do an honest job at edge computing, you need to treat the enterprise edge just like how we use a Chromecast. There is no SSH [Secure Shell] into a Chromecast, but it still works. So why would you expect anything different on the enterprise edge? That is the level of security we think about. You can’t say, ‘We’re secure because we do passwords. Then we do HTTPS for the password.’ How about not doing password in the first place?

What about in terms of public cloud vendors?

So the last bucket of edge players are the many public cloud vendors, without going into names, who have realized that edge is going to be a big point of presence and they need a story around it. They have one, but it reflects their thinking that the public cloud is the center of gravity. Our thinking is the reverse. We think the edge is going to be so important that you need to treat it as a first-class engineering problem, not just as a secondary thing you do among many other things. That is reflective in the depth of services we provide on the edge and the way we treat security, user management, and multi-tenancy at the edge.

Cloud computing is very popular as we’re seeing the rise of things like SaaS and PaaS type of software models in the cloud. It makes it very convenient for people to create applications in the cloud. So we need to make it equally, if not more convenient, for people to create applications on the edge than in the cloud. That will require a very good implementation of edge computing stack. From a quality point of view, it should be equal or better quality than the PaaS-type stacks that are already available in cloud. So we’ve been creating an extremely high quality, very rich set of PaaS services on the edge.

Can you describe the innovation around Xi IoT?

Xi IoT is an edge computing stack. It has three specific things from a product point of view. We have an edge computing stack which consists of a rich set of PaaS services concept that you can [have] on the edge. You can [have it] on a device as small as an Intel Atom-based box -- very small boxes, power efficient boxes -- all the way to Xeon-class boxes like a Xeon D processor or a core Xeon processor. It’s an edge computing stack with a wide variety of hardware. It can be a bare-metal install at the edge or a virtual machine that runs alongside other point of sale systems or ERP systems on the edge.

The second part is this planet-scale management console that is offered as a service. The idea is you get into that service to manage all these edges, which are physically dispersed all around the planet, where there’s one point of interface. So both operators and developers can login and do their stuff. The operators can lifecycle manage the edge and all the applications that are running across all those edges. Developers can inventory their applications and can build new applications using all the platform services we provide.

The third part is we think no IoT solution is an edge-only solution. IoT data tends to move between the edge and the cloud. That use to be the developers problem. Instead, now we provide this concept called a Data Pipeline which stretches all the way from the edge into the customer choice of a cloud. It could be one of the three big public cloud providers, or it could be a Nutanix-based cloud or a VMware-based private cloud. So that Data Pipeline abstracts away the whole concept of, ‘How do you move data between the edge and the cloud, and how do you make secure?’

Internet of things - IOT via communication network service on mobile apps and smartphone and tablet technology for people in digital 4.0 lifestyle

Internet of things - IOT via communication network service on mobile apps and smartphone and tablet technology for people in digital 4.0 lifestyle

Where does Hardis Group fit into Nutanix’s strategy?

The problem we tried to solve this year was, ‘How do we make the delivery of IoT applications as simple as it is to install a Chrome browser on your laptop?’ If you think about it, IoT tends to almost frighten the user from getting into such projects because there are so many moving parts -- you have to deal with AI concepts, you have to deal with clouds, with edge computing, sensors – you don’t even know where to start. To do that, you can partner with people like how we did with Hardis Group. What we did this year was, ‘Can we actually make technology which makes the delivery of IoT applications extremely easy to the whole world?’ Think about it as Netflix. Movies have always existed, but the big achievement Netflix did was distribution, not the creation. It made it so easy that now you can binge watch 12 movies at a time. So can you make the distribution of ridiculously complicated IoT applications very easy? It required the invention of all our edge computing around the centralized planet-scale management, but it also required us to work with some of these very highly vertically focused ISVs and SIs. Hardis Group is a great example of one of those providers.

What problem did Hardis Group solve?

Hardis Group solved a very specific problem very well – supply chain automation. They digitized a whole bunch of things that happen in a warehouse by using machine vision. As soon as you digitize those things, you are now able to make those processes more efficient. They used to achieve it one small install at a time, now with something like Xi IoT, they can spread it across the planet. It makes it easy for them to scale it out to many customers. They can add value to every customer because now application development is much more easier. If you look at Hardis Group, they are essentially sourcing video information from all the surveillance cameras that are deployed at the warehouse. Then they are using AI to understand what’s going on in the scene, like understanding how goods and people are moving around the warehouse, whether they are compliant or on time, etc. To do that, you have to run AI on the edge.

What is your advice to channel partners around their AI and IoT strategy?

It’s one of these very few chances in computing where channel partners have a chance to become the center of the universe. For example, you can be a technology provider who provides a technology, like a database, and run away from the battle field. With things like IoT, you can provide that technology only if you care to understand the exact high-level problem that is being solved. That’s what we’ve proved with the Hardis Group. We have a unbelievably different level of understanding about what happens in a warehouse than we did a year ago. Where our channel partner comes in is that there are so many of these services that are now going to be created. It can be very simple so people don’t need to feel discouraged in starting. You can start very simple. In every vertical, like manufacturing for example, the first thing for a channel partner is establish a great practice around factory automation. That’s a moderate bar to meet. Then that gives channel partners a stepping stone to take that same practice like factory automation and expand your scope to industry 4.0-type of use cases.

Nutanix has spent the majority of this year creating this concept of the Xi IoT App Library. The idea is, if you’re a partner who has created this vertical-specific application and now you’re looking for a means to dramatically increase distribution of that great value-add that you have -- we are the guys. We are the Netflix for you.