Experts call this the year of hyperconverged infrastructure (HCI), a software-defined solution that combines compute, networking and storage on standardized hardware in modular units under unified management. We agree. We have seen a remarkable rise in interest from our customers in hyperconverged infrastructure and clouds.
Other industry experts have also observed the strength of hyperconverged infrastructure. IDC reports that hyperconverged infrastructure is the primary growth driver in the converged systems market, which grew at a rate of 19% year over year during the first quarter of 2019.
For this post, we interviewed Ben Moore, a systems engineer and HCI expert at Nutanix. He explains how hyperconverged infrastructure and 10G networking are transforming the modern data center. He notes that hyperconverged infrastructure is helping solve the need for certifiable, compatible and easy-to-operate multicloud IT environments that deliver for businesses.
As a warmup to our BraveIT panel on Hyperconverged Infrastructure in a Hyperscale World, we appreciate of the opportunity to talk with Ben. Companies around the world use Nutanix Enterprise Cloud OS software to bring one-click application management and mobility across public, private and distributed edge clouds, including those delivered by service providers like TierPoint, so businesses can run any application at any scale with a much lower total cost of ownership.
The Evolution of Infrastructure Technology
Interviewer: How has infrastructure evolved from traditional three-tier infrastructure to hyperconverged?
Ben: I was an infrastructure architect at a major health system in North Carolina for 11 years before I came to work at Nutanix, so I’ve experienced firsthand the evolution of infrastructure from three-tier to hyperconverged infrastructure. There have been a couple of turning points.
There came a point where the complexity of three-tier architecture had gotten out of hand and was preventing businesses from achieving the goals of their infrastructure, limiting their flexibility and agility. To solve that complexity, Converged Infrastructure was introduced as a turnkey package of servers, storage switching, software – all managed as part of a compatibility matrix to make it supportable. Some examples of converged solutions are the VCE products like the vBlock, and FlexPods from NetApp. These products are basically certified reference architectures.
As a customer, I looked at converged products when they came out. On the surface, they simplified the task of keeping things in an operable state and making sure they delivered the outcome that you implemented them for. That was the bright promise of it. But folks that deployed them share a common experience that it is very difficult long-term to keep something qualified inside a compatibility matrix. The first custom need breaks the compatibility of the system, which puts you on a more individualized support plan – in which case you paid a lot of money for a certified infrastructure that’s no longer certified.
At the same point, infrastructure technology took a big jump thanks to improvements in data center networking. Hyperconverged technology at scale wouldn’t be possible without 10G networks and above, which were becoming more cost effective at that time. In the market, I saw hyperconverged infrastructure then leapfrogging converged infrastructure, because people realized they could do things in a better way with hyperconverged infrastructure, inherently. So, all the goals of having certified, compatible, and easy-to-operate environments became real with hyperconverged infrastructure – while also delivering on the promise of business outcomes.
Why go with Hyperconverged Infrastructure
Interviewer: Why go with hyperconverged infrastructure over other options?
Ben: Like LEGOs, hyperconverged is designed to be simple to build with and starts at a small unit of scale. If you snap two LEGO pieces together, you have an object, not just two pieces. Hyperconverged infrastructure usually starts with three nodes but can be less. So, in terms of an easy chunk to bite off, hyperconverged starts at a scale that’s well below what three-tier or converged infrastructure was able to do. As a result, it became more manageable for people to tackle projects with hyperconverged, because it starts cheaper in comparison to our old three-tier building blocks and smaller, more fit.
Another point, with a solution that’s built in hyperconvergence, each solution can be designed for the exact purpose you want to accomplish. HCI is not a one-size-fits-all solution that you have to conform to. You’re not ordering “a rack of converged infrastructure.” Instead, your hyperconverged infrastructure solution can be designed to do precisely what you want it to do.
Hyperconverged Infrastructure or Hyperscale?
Interviewer: We’ll be talking about the benefits of hyperconverged infrastructure in relation to hyperscale at BraveIT 2019. What are your thoughts on the two?
Ben: Nutanix is a company that you might not think at first is public cloud-friendly, hyperscale-friendly, but we are. We believe most customers inherently know where their workloads belong. For example, they know if they have developed a great 12-factor app that’s natively written for services that AWS and Microsoft provide in their public clouds, it probably belongs there. But most companies also have experience with public cloud now, so they understand some of the trade-offs like cost and security that come with it. We believe that in the future, most companies will consume hybrid cloud or multicloud computing – selecting the public cloud for workloads that belong there, Infrastructure as a Service such as a private cloud provided by TierPoint for traditional workloads, and edge clouds to locate compute and storage really close to where you are doing business or where your customers are.
The future from a cloud perspective is that there is no one-size-fits-all prescription for most companies for cloud. We’ll all choose from different clouds services depending on the workload. Our goal at Nutanix is to make each customer’s multicloud experience seamless.
Choosing the Right Hyperconverged Solution
Interviewer: How do you go about choosing the right hyperconverged option?
Ben: On the surface, because it’s so easy to start with hyperconvergence, people might not look at how their HCI solution will grow with them. Just like any other IT decision, it’s still very important for customers to think about how their choice today gives them flexibility and avoids future lock-in. Maybe five years down the road they will decide to change hypervisors, or re-platform to a different hardware manufacturer; unfortunately, not all hyperconverged solutions are created equally when it comes to customer choice. At Nutanix, we currently support our software running on over 20 brands of hardware and on all mainstream hypervisors. Ask these questions as you’re making your decision – remember, one of Henry Ford’s famous quotes about the Model T was, “Any customer can have a car painted any color that he wants, so long as it is black.”
With TierPoint Hosted Private Cloud Powered by Nutanix, for example, our mutual customers can choose AHV, which is the hypervisor included with Nutanix, or VMware vSphere. Today, over 40% of all the clusters that we’ve sold are running AHV. TierPoint is addressing a substantial market and supporting customer choice by supporting AHV. TierPoint’s offering allows customers to try AHV in a disaster recovery (DR) scenario, for example, or a managed private cloud solution, without taking on the risk to run AHV themselves. That’s empowering. It also serves to reduce the customer’s cost by eliminating hypervisor licensing from their stack.
At Nutanix, we also believe that storage architecture matters! When we were founded in 2009 our core principal concepts were Data Locality, staying away from proprietary hardware offload cards, and sticking to our user space CVM storage controller technology – these have all been the harder route to take for development. What matters, even if this is more difficult for us, is that it results in a better outcome for our customers. And we hear that routinely – our product lives up to (and exceeds) our customer’s expectations.
Coming Hyperconverged Infrastructure Trends
Interviewer: What trends do you see in the future of HCI?
Ben: First is the IT strategy most customers have or will execute: multicloud. That is, really thinking about the right destination for each workload. This really grows the need to have mature toolsets in place - tools that allow you to see the cost, manage your deployment method, and make informed decisions on where workloads belong. Mature tools will allow you to provision from a single control plane into a public cloud service, on-prem into a Nutanix cluster, or into legacy architecture – all while showing you real-time cost and compliance data.
Second is the continued commoditization of the hypervisor. Customers are taking the power back on hypervisor choice and are looking to eliminate costs when they aren’t getting value out of it from their current partners. This is a huge current trend that we owe to the Public Cloud – who sees hypervisor as an included part of the services they provide their customers.
Third is managed services, which relates to TierPoint as a provider for hyperconverged technology as a managed service. Because making it simple, making it work and delivering on an outcome, which is why customers are choosing hyperconverged infrastructure, is often not enough. We see from our customers that, even though we’ve achieved all those goals, there is a need for someone to take on the day-to-day responsibilities. The service provider vertical for Nutanix is providing our mutual customers with the ability to have the solution in Nutanix that they know and love, with the option to have someone else run it, and to pay in an OpEX fashion. The managed service frees our customers to focus on business deliverables and outcomes, instead of day-to-day tasks.
Learn more at BraveIT 2019
Nutanix will be part of the hyperconverged infrastructure panel at TierPoint’s BraveIT conference on September 19 in New York City. We look forward to seeing you there.