Transcript (unedited)
Brent Schroeder: So this whole notion of service architectures and applications interconnected with a number of services, whether it's a physical machine, whether it's a VM, whether it's a container, you know, containers have actually been around for three decades. Now, the question is, why didn't these explode in adoption step back and think about what's different now automation of the orchestration of containers. Because if you think about a monolithic application or these large-scale applications, I might have five, 10 moving parts. And I got a little bit of automation or a little bit of scripting. It might be a little fragile, but I got a person over here that knows this stuff inside now. And if something's going wrong, he can jump in and fix it. But if I've got a micro-services app, that's thousands of services that are changing on a daily basis, it is humanly impossible to keep up and manage that and scale it up and down to balance the scaling, what services are running on which systems it requires orchestrations. It requires automation.
Jason Lopez: This is the Tech Barometer podcast from The Forecast. That's Brent Schroeder, the CTO of SUSE, which was founded in 1992, the first company to market Linux to business customers. Today one of their roles is to help businesses make transformations to the cloud. And one of the things we recognize at the forecast is that beyond the marketing talking points, technologists working in the industry have expertise and insights about how solutions and processes work. They’re involved in complexities that require partnerships such as Nutanix’s's partnership with SUSE so in this interview we pick Brent's brain on transformation, cloud native computing AI, and rapid change.
Brent Schroeder: To get automation everything has to be co-defined. That drives the ability to scale in all manners of that application, scaling my velocity scale and my number of user scale and my number of locations. That's the secret sauce of the era we're in. Anytime something's not codified requires a manual step. So I could write the nicest CICB pipeline. That's taking the code, the instant, the developer does it, and it's compiling and is building a testing and everything right up to the end. And then you say, wait a second. I got to do a security test and manually wire together the security settings to ensure that all the firewalls, all the boundaries, all the identity management's done. So, all of that automation, all of that velocity agility comes to a full stop while that manual step occurs. So, what did I gain? It's these manual points that have been the inhibitor to this velocity and Kubernetes was probably the snowball that started. It had containers everywhere. Developers could go write containers all day long, but it wasn't being deployed in mass. I couldn't manage it. As soon as I got Kubernetes, this little snowball of Kubernetes and started rolling that down the hill. I mean, it's just blowing up.
Jason Lopez: I want to ask you more about automation, but I wonder if you could give your view on how the COVID lockdown has affected these transformations. Just off the top of my head recalling the headlines, lots of companies that committed to cloud and remote working and such did well. But I can't help thinking COVID is shaking out the late comers. What is your sense of that?
Brent Schroeder: No, I, I think that yeah, I think that's spot on COVID where it's going to be the final wake-up call opportunity for many companies. If they didn't realize that they need to be able to react real time to things and survive through that, even if they survived by accident, everybody needs to change and recognize how fast the world evolves. And if you're not capable of keeping up with the pace of change, then, you know, I think that demise of companies is going to definitely accelerate for any of those that are not there. I think the point is, what are you waiting for? It may not be perfect, but the fundamental foundational capabilities to digitize the business, to make the business more agile, to be able to respond, you know, at the speed of the economies today is all there. Don't wait for COVID 21, get yourself prepared and get prepared, now.
Jason Lopez: Well, we're going to hear the word transformation a lot in this interview. It's like we'll be overusing it ourselves. As you told me before the interview that one of the things you wanted to do was get away from using vague buzzwords. So let's talk about what transformation actually means today.
Brent Schroeder: Yes, this isn't unique to this transformation, but people often want to say, why don't you take this and lead this transformation effort and go over in the corner and pick two people to help you then come back when you've transformed us. And a whole bunch of people get left out. I think this is one of those that it's, everybody needs to own some responsibility to be part of the solution, the business stakeholders, application owners in the process for how applications are developed. The infrastructure team has a role in participating in putting in the automation and making their investments. And so from the very people that are defining the requirements of the application from the development teams that are responsible for building it from the infrastructure teams that are responsible for having some place to run it, whether it's on-prem or in the cloud, wherever it is, then each has a set of responsibilities. And this absolutely requires new learnings and new skills. Building the applications and managing the infrastructure will be different. The tools are different, the development methodology is different. And so the company has got to buy into training, into hiring the skills that are necessary to take advantage of the set of capabilities. And so that's really what I mean by, it's everyone's responsibility. Because if you want to compartmentalize this and just put it off in a corner, and then, I think it will fall short.
Jason Lopez: Well when I think of new learnings I think of one that‘s going to be integrated in all of this and that’s AI right?.
Brent Schroeder: I think AI is now going to start to enter the fray because as the complexity grows, as I scale even larger, I get more and more occurring. You almost have to then have automation powered by AI. And so that's where I think the AI ops starts to come into play. You know, we were looking at it, in an example of containerized storage. If I've got a storage file system, that's distributed over hundreds of systems and it's got thousands of microservices in there running. If I get hotspots as an administrator and move it, I'm going to be so far behind the curve that it's not going to be possible to keep up. And that's where I think AI starts to enter, whether it's something like infrastructure with storage or it's order processing with, you know, global seven by 24 vulnerability following the sun and making sure that I get minimal response times to my customers, AI has to come into play to maximize it, to take it all the way to the end across the goal line.
Jason Lopez: And what's the role of AI, is it to manage the process and or is it basically to help look at the data and help IT understand it?
Brent Schroeder: No, I think it, yeah, I think it's that I think AI is identifying what needs to be done. It's the what, cause it can go through all of that information to say, okay, here's the triggering event here's what's occurred. This process needs to run. I need to move something from X to Y. So I think it's a participant in there with other orchestration tools working in conjunction with it.
Jason Lopez: Do you know when we were talking earlier and he mentioned some real world examples which will get to one of the moments here but those examples almost always seemed invariably come down to old software that's still on the factory floor helping to run things if you know what I mean. What is it about those kinds of applications that have a target on his back so to speak?
Brent Schroeder: It's where you're delivering your business value, right? If there's a monolithic application or this large application, that's the core of the business value since the business is always changing, you're wanting to extend it. So it could be manufacturing. It could be, you know, your point of presence from, interacting with customers, maybe some monolithic banking ATM system that's just running, you know, off of this monolithic mainframe application where you don't ever want to touch it because
Jason Lopez: I've heard that with point-of-sale apps where it has to be done in a very mission critical kind of attitude.
Brent Schroeder: Yeah. So I think it's those that, versus the things that are more in the kind of in the back, you know, it's hidden away tucked away someplace and it's just turning the wheels, but it's not the business flow where you're always doing that optimization. We're interfacing with key stakeholders and customers.
Jason Lopez: So, I know you came prepared with an example of this, the idea of a monolithic application that gets moved to the cloud.
Brent Schroeder: Well that was the one I've written down is that one of our customers is a global manufacturer, 45 sites around the world. The challenge that they were having is monolithic applications in the manufacturing centers and it needed to be on the factory floor due to latency, but the whole application, you know, required these large systems to be sitting in the factories. So they went through a study of what the application looked like? How do you break it down? And if they went to a containerized model that actually could shrink the on floor footprint significantly in the 70 to 80% range by using microservices architecture, then they brought a big portion of the application up into the cloud, left the latency sensitive sides of the application in the factory. So now they've got a more centralized, easy to manage core processing in the cloud and a much smaller footprint, lower cost footprint in the factory that gives them the same degree of latency. That's there. Then from an application standpoint, by decomposing it and having a microservices architecture that can continually deliver new capabilities to the factory without bringing down the factory. So now they have lower downtime and can bring new applications faster and have a lower operating cost in the factories because of switching from a monolithic architecture to a cloud native microservice architecture.
Jason Lopez: We talked about the speeds and feeds but you mentioned earlier that it's more than just about the hardware and software, it's in the beliefs people act on… the culture. And in many ways it's not really so counterintuitive, is it? How businesses have to respond faster, there are new opportunities and new threats. What's your thinking on how to respond to these things?
Brent Schroeder: Yeah, there's the technology adoption, but the business goals aren't necessarily about the technical transformation, what they want is a business transformation. They want to, to be able to transform the business faster. I think this last year provides a great example, at least for us in the United States, right? On March 11th, you know, I just mapped out my 2020 plans and everything was great on March 13th, everything was different for almost every business. Now, many of them were like, there's no way to adapt to some of the things that they were being asked to adapt to where other businesses had the processes in place were nimble enough to be able to adapt, to be able to quickly deliver new capabilities or new ways of meeting with the customer. Almost overnight gave them a competitive advantage. And so the business goal I think, is to have that degree of ability to respond or take advantage of things that are happening around them, new entrance to the market, geopolitical events, you know, whatever the catalyst may be. but people want to be able to respond faster because just doing business the way we did last decade or two decades ago, et cetera, things just move quicker than that.
Jason Lopez: Business goal are not necessarily what you're talking about, well let me put it this way, what you're talking about is that business goals have to do with the company’s entire culture which includes its technological capabilities.
Brent Schroeder: Yeah. Let me break it into two or three steps. Because the first step is kind of the lift and shift. I've just got some existing things I'm going to just lift and shift that to either traditional cloud VMs in one word, or they want to lift and shift it from a VM to a container, whatever it is just that generic term. And some return might be found in that, but it's clearly not going to maximize the return. Then there's the, do I transform my applications and which ones do I transform? I recommend companies to actually not start there. If it's running, start with something new, do I transform existing or do I start with a new project? Because I've got to go through some technological transitions, I've got to go through some process and skillset transitions, et cetera. Companies are I think more successful quicker, but finding a project Greenfield opportunity where I can set it off, get the infrastructure set up, get a team that comes up to speed on it, gets trained on it. They're not blind, it's blind. It's probably too strong of the words. They're not distracted about how it's currently implemented and how to shoe horn that into a new way of doing it. They can start with a clean sheet of paper and say, this is how it built an application. They start to build muscle memory. They start to build a skillset of recognizing and, and they can go through some fast fails and it has no or less implications on the rest of the business. If I want to fast fail in a transformation, that's much more difficult because I've got an existing application running and when I move it over, I've got some FLAS and that I expect when I move it over, failing fast with that is, is a tricky proposition. And so my recommendation is identify something to start with and go through a couple of iterations, learn, build the team, build the skillset, be able to recognize what works and what doesn't work and do that first. And then come back and look at what needs to be transformed in your second generation. And there will be applications we'll, you'll see a, a monolithic application that would logically break down and be more efficient. Something has got to routinely be updated and lifecycle managed or new capabilities have to be put in. You know, we've got some of our case studies are this, is that a company having a manufacturing system that's built off of monolithic applications. And so every iteration is this big, long planned out, endeavor with ultimately needing to stop manufacturing for a short period of time while they do this light switch from the old application to the new application, and then restart being able to put that into a cloud native has a big return, you know, because now I can start incrementally adding capabilities, et cetera. On the flip side of the coin, if I've got a monolithic application that I've not touched for a year and I periodically got a security patch it, or make sure that basic care and feeding done, but I'm not delivering new capabilities through it, et cetera, just leave it as is. Not everything has to be microserviced. Traditional monolithic apps can exist, right with new cloud native applications. And there's, there's ways to get them to integrate and work together. You know, if an application is working just fine, it's been there for five years. It's been there for 10 years and I'm not extending and adding value to it on a frequent basis. Then my recommendation is just leave it and let it run its course and add new capabilities outside of that context.
Jason Lopez: While I'm listening to this Brent it kind of goes back for me to some conversations I've had with CIOs in the past few months just about the buy and that's needed because it's a people thing where people need to get on the same page with the vision it goes beyond the coding and the actual technologies but what are people going to learn what are they going to do with those learnings and how does the company expect to integrate all this it sounds like the whole thing that you're talking about.
Brent Schroeder: Yes, exactly. And being on the same page, you know, I was talking to a company not too long ago and this particular company, fairly large company in healthcare, the CIO is responsible for the infrastructure, but not the applications and the business units have the application developers and they are fundamentally disconnected about how to achieve and what they want to achieve. And so that's a perfect case. Example of, we all have to kind of work this as a team, across the organizations. We have one page on what the strategy is. There's not one right recipe, but everybody needs to be working from the same recipe.
Jason Lopez: Well maybe you just answered this question but I wanted to ask her anyway if there was one thing that you wish businesses could hear about transformation, the transformations were talking about what would it be.
Brent Schroeder: The one thing that customers really need to hear the businesses really need to hear is that there is a tremendous opportunity to gain efficiencies in the delivery of new capabilities and in the operations of those capabilities. And that there is more to it than the buzzwords. You know, what we're seeing is the opportunity and the reality is very, very high, but it does come with some asterisks, study it before you enter it. And the opportunity is there.
Jason Lopez: I’m thinking one of the asterisks is it’s not cheap.
Brent Schroeder: One of the notes I had wrote down, you must invest $2 to get $3 back. You've got to invest to get the return back, whether it's in technology, whether it's in people. And it's not just a physical cost or a capital cost, it's not a specific cost. If you go in with that understanding, then you know the returns are great.
Jason Lopez: Brent Schroeder is the CTO of SUSE. This is Tech Barometer, the podcast from The Forecast, an online magazine on technology and the people in tech. I’m Jason Lopez, thanks for listening. If you enjoyed this interview check out our other podcasts and print stories at www.theforecastbynutanix.com.
