Opening Up About the Cloud
SPONSORED BY NUTANIX
A Q&A with Open Source champion and enterprise cloud pioneer, Brian Stevens
Prior to joining the Nutanix Board of Directors in June 2019, Brian Stevens was the CTO of Google Cloud where he helped the technology juggernaut tailor its public cloud services to the needs of enterprise customers. Before that, he spent 12 years at Red Hat, where he evangelized the then-revolutionary idea of Open Source in the corporate datacenter.
NEXT Magazine recently caught up with Brian to learn more about his journey, where he acquired his “open” point of view, and to share his thoughts on the current and future state of cloud technology.
NEXT: You began your career as a software developer. How did you getstarted in tech and what were those early days like?
Brian: I grew up at a time when computer technology was just starting to be accessible to high school students. So, while I thought programming was a lot of fun, I had no idea you could build a career on it. Fortunately for me, I had a wonderful guidance counselor that said I could, so I switched gears from wanting to be a carpenter to wanting to be a software developer.
Back then, technology wasn’t as open and readily accessible as it is today. To gain access to technology, you had to work for one of the big tech companies. I was lucky to get a job at the Digital Equipment Corporation (DEC), so I suddenly found myself in this great environment where I had all of this technology to learn from and experiment with right at my fingertips.
NEXT: Open Source played a pivotal role in your career. When were you introduced to the Linux environment and what part did you play in its further adoption?
Brian: I left Digital to join Red Hat in 2001. I think I was chasing disruption. As a technologist, I’ve felt the need to be connected to the outcome—the software I was working on had to have a massive impact on the user. Red Hat provided that opportunity. They built this great trusted brand that developers loved, but they weren’t being used in enterprises. Back then, it was the IBMs, HPs and Sun Microsystems of the world inside the corporate datacenter.
My focus was always on enterprise customers, and it was pretty evident to me that there was an opportunity to revolutionize enterprise computing if it were more open. You could see this consumerization of technology happening at the desktop with Intel, but the R&D cycle of investment ultimately created service at great cost. So, the disruptive opportunity was to deliver great value to customers by bringing “untrusted” Open Source Linux on Intel into the enterprise infrastructure. We built a team to develop an enterprise version of Linux and launched the industry’s first software subscription service to create a tighter relationship with our customers.
NEXT: When were you first introduced to cloud computing?
Brian: One of our biggest customers at Red Hat was Amazon. We helped them build the first instance of what became Amazon Web Services (AWS) in 2006. When I saw it launch, it was such an epiphany for me. Just when you think that Open Source and Intel in the enterprise is the end state, you suddenly realize that this public cloud infrastructure-as-a-service is just another step in that same journey. This is the future. It will make the world so much better for developers in IT. When Google approached me in 2014 to run product for Google Cloud, I was really excited to get directly involved shaping the future of cloud. Google was doing a great job focusing on their current users, which were all the startups and native-cloud companies, but they weren't focused on incumbent enterprises. So, just as with Red Hat, while Google had many innovative technologies, they were missing key areas needed for enterprise integration and adoption. And of course, with enterprise, there's a lot more to it than just technology. Of course, we had to focus on security and integration and such, but there were also business concerns to address: privacy agreements, service level agreements (SLAs), what processors you use to operate the cloud. As much as it was building the technology, it was very much about building the business and the controls such that enterprises would be comfortable with our service.
NEXT: What’s your cloud prediction for 2020 and beyond?
Brian: The industry has been through a world of consolidation. Now, we just talk about the big three (AWS, Microsoft Azure, Google Cloud). The organizations behind each of these companies are huge—tens of thousands of software engineers. Every day, there’s more code being written, and more complexity as each company strives to deliver a more capable cloud. But as each cloud becomes more advanced and feature-rich, they also become more disparate, more isolate, proprietary, and harder to use. Just like the datacenter incumbents a decade earlier, the big three cloud providers had been fighting open standards and commoditization. But you’re starting to see Open Source influence cloud services and architectures. Kubernetes and Docker and API gateways, all of those primitives that are really valuable to end-users are starting to be adopted in common across all the public clouds. I think that's a great thing. We're now just starting to see more cloud portability at the application level between clouds. While that sounds like a small thing, it'll have a huge outcome for many years to come.
Looking out over the next five years, you’ll see cloud providers not only federating artificial intelligence (AI) capabilities to their customers but using it themselves. You’ll begin to see clouds that are better at predicting and making decisions and offloading that from IT and developers. All that work IT has to do around resiliency or availability, scaling up or scaling down—you’re going to see the clouds using AI to automate every aspect, from security to availability and resilience. All of those outages that took out whole regions will also become a thing of the past. With AI’s ability to monitor vast amounts of data and predict outcomes, the public cloud will be more like dialtone service with regard to uptime and reliability.
NEXT: You’ve called hybrid cloud “a game-changer” for the enterprise. Why’s that?
Brian: Enterprises are much more comfortable using public clouds, but they’re still going to want to run applications on-premises. So, the ideal state for them is to be able touse technology in their datacenters, technology in their branch office, and technology in multiple clouds—with no restrictions. Unfortunately, you have this world where the public cloud providers are only focused on their own public deployments and the on-premise providers are focused on their side. When you look at an enterprise customer trying to use both, it's incredibly difficult and costly in people, developer and IT skills, because the environments are so dissimilar. The companies that can make this super simple with a great user experience—make it easy for developers and IT to deploy and manage applications on-premise and publicly, without a degradation in capabilities—are going to provide a tremendous amount of value for enterprises.
NEXT: In the meantime, what advice would you give enterprises dealing with these challenges?
Brian: I would tell them to build a software factory. I see too many companies that go all-in on one set of technology stacks and then they get left behind as a new capability comes about somewhere else. If you continue to thickly provision and develop for a specific platform, it becomes way more difficult to adopt new platforms, operating systems, or new environments. It’s very costly.
In a software factory, you think about how you can isolate the code from the deployment environment. If you have a release engineering sort of pipeline for everything that the CIO office and the lines of business do, and they all use standardized software factories for production, you can now leverage a set of similar processes that allow you to take advantage of disparate deployment environments at the backend.
NEXT: It’s always been a challenge to find and keep good IT talent. What do you look for in an individual and what’s your advice for nurturing high-performing teams?
Brian: The IT environment is constantly changing. You don’t just hand someone a spec and say, “go build this thing.” So, when looking at talent, it’s really around what I call aptitude and velocity. Does a young software developer have the aptitude and are they curious and self motivated to learn new things? You’re looking for people that understand the choices and are brave enough to go out there and figure out the answers. To me, it is less about the university someone went to and more around the individual’s intellectual curiosity. Do they like to explore, play, learn, and break things? That’s what you want.
As far as team building, the reality is your team members are not necessarily going to be in Mountain View or Boston or London anymore. Cloud and Open Source have federated technology so that far more people have access and can go out and play with different frameworks and things. You have to build a model that enables people to collaborate wherever they are if you want to scale your business. And you need to pick team players. When I’m interviewing individuals, I always ask, “Tell me about the last team-based project you worked on. What did you do? What did they do? Did you ever shift roles? What did you do to help them, and vice versa?” You get far more interesting conversations when you start with this approach.
Most importantly, you want a team that is outcome-oriented instead of being task-driven without any sense of what success looks like. When they know how to measure success, a lot of things fall away because they’re not necessary to the mission. It’s liberating for the entire team. It enables them to focus and reassess daily whether they’re pointing in the right direction.
NEXT: Thanks so much for your time, Brian. Before we sign off, could you tell us which cloud applications excite you the most?
Brian: What excites me is the continued federation of technology to deliver access and services to people and environments previously cut off or underserved. When you combine cloud services with AI and all of these APIs to connect things and devices, the possibilities are endless. There are thousands of examples of tech doing good but one of my favorites is a diabetic retinopathy test developed by Google. Diabetics can go blind when this condition goes undetected. This doesn’t happen often to people sitting here in Silicon Valley or anywhere else with easy access to doctors. Now, anyone in the world can snap a picture of their eye with a mobile phone and upload it to the cloud where an AI algorithm will analyze their risk of blindness. Before, you had to have labs, doctors, and be physically present in certain regions. Now, the hundreds of millions of dollars that went into developing AI capabilities for some other reason are being deployed to benefit people thousands of miles away from Silicon Valley. That’s exciting.