Enterprises today are innovating in ways they’ve never dreamed about before. At the same time, competitive pressure is higher than ever. Every company has to innovate, or they fall behind. One of the main ways that enterprises are driving innovation is to intelligently leverage cloud technologies—and thereby transform their IT architecture and the fundamentals of the business itself.
In this interview, Ingo Fuchs—Chief Technologist for Cloud and DevOps at NetApp—discusses the transformative nature of cloud adoption, the challenges associated with evolving to a hybrid multi-cloud environment, and steps to take to prepare your business for taking a smart leap into the cloud.
SDxCentral: Cloud has been a hot topic for years now. What makes cloud adoption so transformative for an enterprise?
Ingo Fuchs: As I see it, the cloud offers two great benefits. The first is that it lets you scale very quickly, so you can create new services, roll them out, and bring them online immediately. So you’re able to address special circumstances like Black Friday in a way that lets you quickly bring additional resources online, scale the resources to serve that event, and then turn them off when you’re done. The sheer ability to scale when you need to—and unscale when you’re done—is a key value of the public cloud.
The other key benefit involves the best-of-breed services that service providers deliver. When we talk about specific services—for example, facial recognition, voice recognition, specific analytics tools that might be useful for web traffic analysis, areas like that—we’re seeing a lot of innovation from public cloud companies that customers want to adopt. Many customers are finding value in what we call the hybrid public cloud, in which you consume some services from various public cloud providers, and other best-of-breed services from providers that specialize in a certain area, and still other services that you prefer to have on premises.
SDxCentral: How do you expect to see that enterprise transformation evolve moving forward?
It goes back to that idea of customers saying, “I want to have the flexibility of the cloud, and I want to have compliance and cost models that work on premises, so how can I build a cohesive and transformative IT infrastructure architecture that meets the needs of my developers, my application operators, my financial officer, and all the other stakeholders in my organization? What’s the best mix between on-premises clouds and public clouds in this kind of environment?”
You need to be able to measure things like performance, cost, compliance, regulations, and all that needs to be reflected in the future enterprise IT infrastructure design.
SDxCentral: Many enterprises haven’t yet embraced cloud computing, or they’re just getting started. What challenges do they face in moving to a hybrid or multi-cloud environment?
There are a lot of challenges! First of all, you need to be very clear about the outcomes you’re trying to achieve. Do your goals involve scalability, innovation, optimizing for cost, reducing risk—what are the key outcomes you’re trying to achieve? Out of that analysis will come certain needs.
But one of the big challenges I see customers trying to overcome is the concept of walled gardens, where you have applications and services running in specific environments, either on premises or in the cloud, and you can’t get the data and the application out of these gardens. So you might be committed to one public cloud, but how do you get the applications and services and data into another cloud or put them on premises? A key challenge is to eliminate these walled gardens.
Another challenge is that every cloud is different. Every cloud uses different tools, every cloud requires different training, every cloud uses different terminology and different architectures. We need to introduce abstraction layers that make it easy to use multiple clouds without requiring in-depth training and a building up of expertise across the stack.
Finally, we need to make it as simple it possible to move data and applications wherever they need to be in the hybrid multi-cloud, and that requires a full stack of solutions—across both the public cloud and the on-premises cloud—that are built with the clear design and architectural goal to make this movement of data and applications very easy. Overcoming this particular challenge is important for our customers because they’re just now starting to encounter these problems.
Obviously there are a number of criteria to think about when you’re building your infrastructure. Looking at NetApp IT as an example, we have a very strict process where we look closely at all the applications in our environment. We classify all those applications according to business needs and the necessity to innovate them. You start by sorting your applications into different buckets, whether these are applications that are core to your business and are driving new innovation, or they’re systems of record, or they’re applications that you’re simply tolerating and they’re fine where they are.
When you come out of this analysis, you discover that you have some applications that you really want to drive innovation with, and you probably want to lift those into a cloud architecture, whether that’s on premises or in the public cloud. There will be other applications that you’re fine running in a traditional or virtualized IT environment that isn’t fully cloudified. You end up with a spectrum of applications where you’re mostly concerned with cost and reduction of complexity and other applications where cost isn’t the problem, it’s more about performance and scale. You want those applications to be as fast as possible because you know that every millisecond of slowdown will cost you customers.
SDxCentral: Where do you recommend companies start when they’re building cloud infrastructure?
The clear trend is that customers are moving to hybrid multi-cloud. This goes hand in hand with another major trend we’re seeing, which is the importance of the DevOps team.: It’s becoming more and more important for IT departments to serve their developer and application operations groups because the public cloud has become almost a competitor to your internal IT organization. It’s turned into a fight for survival or relevancy within the organization now that developers can simply go to the cloud to develop applications, and operators can go to the cloud to run applications—especially once they’re containerized.
The question is: What can IT do to deliver a better experience for the developers and application operators on premises so that it’s just as easy as the cloud?
SDxCentral: Getting a hybrid cloud up and running is just the starting point. What do enterprises need to change to effectively use a hybrid (or multi) cloud?
A difficult thing to do in this world is to hold all of your providers accountable, and that includes your on-premises IT and your public cloud providers. This stresses the importance of infrastructure analytics and infrastructure management tools that can investigate this hybrid multi-cloud and tell the decision makers, stakeholders, and IT department very clearly what performance you’ll achieve, what capacity you’ll get, what cost will come back (whether you want to charge back or not), what it will cost to achieve a certain level of performance? And also to predict the resulting cost and performance of moving a workload somewhere else?
SDxCentral: Any final thoughts?
When I look at this hybrid multi-cloud world, it seems very similar to the move from mainframe to client-server infrastructure. That was a dramatic shift not only in the technology that was being used but in how firms re-identified how they leveraged IT to run their business, how they organized their teams around these new infrastructure options and models. This change to hybrid multi-cloud is at least that big—it’s a very exciting time, a big opportunity for everybody to show what they can do, for every IT department to reinvent themselves and reinvent the contribution of IT to their business.