>> Announcer: Live from Las Vegas, it's theCube, covering IBM Think 2018. Brought to you by IBM. >> Welcome back to Las Vegas, everybody. We're here at the Mandalay Bay. This is theCube. And we have two days, sorry, three days of live wall-to-wall coverage of IBM Think 2018. My name is Dave Vellante. I'm here with my co-host Peter Burris. Steve Robinson is here; he's the general manager of client technical engagement for IBM, and he's joined by Bala Rajaramen, who's an IBM fellow, expert in Hybrid Cloud. (microphone feedback) Gentlemen, welcome to theCube. >> Oh, good, thanks for having us. Always a pleasure. >> Thank you. >> You're very welcome. So, Steve, let's start with you. >> Sure. >> We were talking off camera about some of the work that we've been doing in what we call true private cloud. You talked about some of the work you've done with BCG, your own internal work. What are you seeing in terms of private cloud and the resurgence of private cloud? >> You know, it's kind of fascinating. You know, over the past, probably two years, we started to see this kind of next definition of private cloud coming about, where most firms had spent a lot of effort on virtualizing their data center, building up these beautiful VMware firms, etc., and then this next level is, how can I start to do more cloud capability back behind the firewall? This notion of CaaS, container as a service, started showing up in RFPs. People wanted to say, hey, can Kubernetes come back as well, could I use private cloud as a parking lot for certain workloads, and could it possibly be the basis for doing true multi-cloud down the road where some of these environments may start landing both on private, multinode private, and even on public as well. So it's been a real resurgence from our side. >> So we made the observation several years ago with our research team that CIOs are realizing they couldn't just put their business into the public cloud. >> Right, right. >> Rather, they wanted the cloud experience and they wanted to bring that experience to their data wherever that lives. >> Exactly. >> So, Bala, what technical challenges does that bring and how do you guys solve them? >> Yeah, it's interesting because, I mean, when you look at cloud, it's about what makes something a cloud. And I think the two things that CIOs are struggling with, which is why public cloud was an attractor initially, was that easy self-service. I can get to things quickly from the business perspective, and I can manage it very consistently because everything works the same. And I think what Steve alluded to was when you bring the cloud to you, it's not just bringing the capability, but it's bringing the experience. And can people get to it easily? Can businesses be competitive in that environment? Can the operations guy manage that environment like they would manage something at a cloud scale? And that, essentially, was the challenges we had to solve, not just in moving things, but in moving all the right pieces around it so that it was a cloud, yeah. >> So, we talked with you about the whole concept is the cloud experience where the data warrants it. And as you said, it's not just bringing technology in, it's also bringing the entire operating model-- >> That's right. >> Of how the cloud works. IBM is a big company, has always been its first customer. What has IBM been learning as you become more of a cloud-first company, or a cloud-oriented company, and how are you bringing that to your customers? >> Well, definitely I think the key thing we've been doing has been in a, you know, spirit of transformation for the past three years, as well. One of the things we picked up critically when we started the private cloud effort is there's a dimension of having to fit in with what an enterprise has already. They've got a strong system management process in place, they've got ticketing, they've got their plasmas up on the hall showing the up time of their applications. The biggest challenge was when they were moving to public cloud, they were kind of giving that to the public cloud vendors and they were losing visibility in that as well. So part of this, we had to respect them to be able to allow them to see their applications, to be able to fit into their existing environments, and be able to fit into the process. We can't leave that system management team behind. >> And just to add to that, I think when you look at the evolution of things like microservices, you're breaking something that was intrinsically a whole and manageable as a whole, into a bunch of individual pieces. That challenge has always existed when you move from mainframes to distributed, because the management challenge more than anything else. You could build applications quickly, but it's really hard to manage them with microservices across multiple clouds, it's a fascinating exercise. So I think our learnings, to your point, was we have to think about it in a different way. Think about from an app-centric way, not from an infrastructure-centric way. And I think that's critical. >> I want to build on that for a second because Judy talked this morning, and we certainly would agree with the concept of your data as an asset. Are we really thinking apps-centric longer term, or data-centric longer term, and apps-centric is more how do we affect the transition because that's where the value proposition is today? >> Right, and that's where your assets are, right? >> Right. >> And your data becomes an integral part, an entangled part of it. As you split your applications, you're also looking at splitting your data, and how do you manage that? How do you manage where the data is placed? Manage where applications are placed? I think the true cloud value, going back to your question, is how does this multi-cloud universe around placement of data, placement of applications, security models, availability models, how does it all come together? And I think that's the biggest challenge, and I think we are doing some interesting work to address this. >> We almost view it always as kind of two planes at the same time. Where do we optimize the application based off of the performance characteristics, you know, how much compute do we need around it, you know if it's a very sophisticated investment banking, let's get that closer. We've even been running private cloud back on the mainframe, Kubernetes clusters back on the mainframe. But then the whole data story now with regulatory, with GDRP, etc., gives you another layer of complexity. So we almost have to look at what's the app doing, and then what's the data doing at the same time? >> You've kind of called it cloud your way. >> Yeah, right. (laughs) >> You used that statement a while back. And so we could define cloud a lot of different ways. We're talking about our data, you talked about some of your studies, and you show them, actually, the private cloud and the public cloud infrastructure's comparable in size. >> It's pretty close. Pretty close. >> We show private cloud smaller but growing twice as fast, so, okay. >> But we also call it two-prong cloud. >> Yeah, so we maybe have a different definition, but let's talk about the customer definition. >> Of course, yeah. >> A cloud is in the eye of the beholder-- >> Beholder, right, right, right. >> Beholder's the customer. So to me, it's about the business impact. Are they seeing an impact on agility? Is it changing their operating model? Because if it is, then it's going to have a bottom line impact, and if it's not, it's just a lift and shift on prem. What are you seeing in terms of the customers? >> Well, I think it's interesting, though, you used the term lift and shift. That's one of these, I call it an urban myth of cloud. Nothing is a lift and shift. >> Dave: Right. >> I think part of the challenge for us is could we bring some cloud attributes back behind, and what would that do for you? I know Bala mentioned self service. We, you know, some degree of horizontal scalability. We'll never have ultimate scalability like we have in the public cloud, but we can spin up multiple instances and start to manage pieces in a different way. The other area that we looked at that we had never thought about when we did our Bluemix local product, etc., could this be a path also for their middleware coming forward at the same time? Could we take this opportunity to start to containerize WebSphere, MQ, DB2, so that more workloads could move towards the cloud without having to have them be fully replaced and change up all the dependency chains, etc. So that's been a key thing, to pull the gravity of that middleware forward, while you kind of have it back on premise, as well. >> Yeah, absolutely. I think, going back to the lift and shift point, right, I mean, I think the traditional disadvantage of a lift and shift was you're moving your bad with your good, right? >> Right, right. >> And I think what this gives us in approach is how do you actually decouple that? Your applications are your crown jewels. You have invested a lot of effort over many years. What held you back was the processes you put around it that slowed you down. So being able to, to Steve's point, when you bring WebSphere, for example, onto a cloud platform, you minimize the risk, you enhance the value of building your application or moving your applications as is. That's a valuable lift and shift. But what you're not lifting and shifting is all of your processes, all the bureaucracy, all of the more traditional ways of doing things, and that combination, I think, is really the, to pick on your definition, is a true private cloud because it brings a customer-perceived value of, and a customer-perceived values risk. It is cost, and how do you optimize that? You're minimizing the risk, you're giving them a new operating model, a new self-service model, that takes away the bad and keeps the good. And I think that is, to me personally, I think that's a very exciting thing. >> Well one of the things that people always talk about when they talk about cloud is they talk about elasticity. >> Steve: Right. >> Great idea. But we like to talk about plasticity. >> Steve: Yes. >> Which is a different definition. Elasticity is same workload and scale, plasticity is the ability to consume, bring up, new workloads, do a better job of patches and updates. >> Steve: You got it. >> What do you think about that notion? At what point in time does the industry start to focus more on the fact that you can use cloud to fit your business differently? To snap your business into place differently, as a consequence of these services? >> That's a great insight, and it's one that I think most people just don't realize out of the gate that even bringing some of these cloud capabilities and also some of these more advanced container orchestration capabilities to all of their workloads gives them a lot more flexibility. We use a term pet versus cattle. You know, where in the old days, I would stand up, middleware stack, etc., and I would do everything to make sure that thing stood up, it was never impacted, etc. With some of the orchestration that we find in Kubernetes, I can stand up six versions of those. If one ends up getting knocked down, who cares? I can just automatically launch another one right back up. It changes the way how I manage that environment. It gives you more flexibility. It gives you more dynamic capability as to where I actually put individual pieces, even with my own infrastructure as well. So I think this could open a whole new era of how I manage. The plasticity; I like that idea as well. >> Yeah, that's a great word because I think when we started this discussion, I did not define cloud as being elastic for very much the same reason because from a business perspective, elasticity is a lower down function, or more of a second-order function. Being able to consume it easily, not be worried about how it's deployed-- >> It's a value proposition with a cloud guy's term. >> That's right, that's right. >> Exactly. And so plasticity's a much better word because that is a business impacting statement, which is, all to the point, which is I can deploy it. I can remove it. I'm not locked into particular things. I can evolve it very quickly. I think you're absolutely right and I think it's different. >> So speaking of the cloud guys, I got to ask you. So if the cloud guys were here, the public cloud pure plays wheel, they would say, oh, IBM, and we get this all the time with our true private cloud, that's old-guard thinking. >> Sure. >> Okay, so what we're doing is changing the world, what they're doing is trying to put a little, you know, lipstick on virtualization. How would you respond? >> If you look at the workloads that a typical enterprise, now, trust me, if I was building greenfield applications or doing a brand-new start-up with my BC money, etc., boom, if I had the chance, I would put it on public and run right away. A lot of flexibility, etc., with that. But the enterprises that we've worked with, I tend to say that most of the ones where we come in and we evaluate large number of workloads, you know, we just did one with a bank. We evaluated 900 different workloads. 15% met their regulatory and their risk policy and could move to the public cloud. That leaves 85% that are either going to stay in their legacy state, or are not going to start taking advantage of some of the cloud concepts we have. So, yeah, you've got to come back behind. And I think if you look at the public vendors, they're trying desperately to either send hard drives down or send appliances down. They understand they're going to have to extend down so that they can bring more workloads the right direction here. >> Of course. >> Now, we're advising our clients to focus on what's their value proposition, what activities are most important, what data's required to perform those activities. >> Steve: We say right cloud for the right workload. >> Yeah, and the question with data is latency, regulatory, and IP protection. Does that resonate with you guys? >> Yeah, that resonates very well. And I think, to me, we are trying to impose a strategy on a current state of the universe. So I think we are arguing whether public cloud is the right answer, or private cloud is the right answer, based on how we perceive private and public today. I think it's, in the next 10 years, you're not going to be able to tell the distinction. I mean, it's going to be more like a central office model where you have the core switches, you're going to have distributed switches. That is the cloud. And who manages what, how you delegate it, multiple providers, cross-provider billing, it's going to become a fabric. Then I can't tell the difference between what's public and what's private. I mean, I have the boundaries well-defined. And so, I think I view that as the eventual strategy, and I think we are now predicting a future that we are just guessing. >> Does that suggest, Bala, then, that the capabilities of the on-prem services are going to be substantially similar to what you see in the public? Do you guys benchmark yourselves against your IBM cloud brethren and have a little healthy internal competition, or? >> No, it is contextual. So if you take something very complex like weather, where it is gathering data from a whole bunch of sources, it makes almost no sense to have something that's local. But if you look at some of the other services, even things like machine learning and so on and so forth, there are some that make perfect sense on a cloud. There's things that make sense on closer to the data, on premise. But what's going to be more interesting is how they work together. And over time, you're going to see the programming model evolve to eliminate the distinction between what is private and what's public. And you're going to see an operational model evolve with the right delegation and controls that wipes out the distinction. In 10 years, I think we are not going to be having this discussion of private versus public. It is going to be a cloud with private components, with public components, and the ability for, from a business perspective, for a client to manage it in the right way. >> So things like, sorry, Peter, things like functional programming models will be pervasive. >> That's correct. >> And it will be up to the client to choose which, where their data is, essentially, is going to dictate what they use and what-- >> Well, and I think-- >> The business requires. >> We envision today where it's almost done on a dynamic basis, you know, where you're really to the point where I may have a load that's based on CPU, etc., running predominantly in the private cloud. Then we have a bursting scenario, actually be able to pick that container up and dynamically move it up to public as need be. As my risk and compliance rules begin to change, I could dynamically say, the same application, these three we're running here today, let me do a distributed, distribution of those as well. Not heavy lifting. >> Really quickly, so we're going to focus more on how you get value out of your data where the infrastructure's not the issue, and even the applications are less of an issue. One quick question though. >> Steve: Yeah? >> So we talk about, we talk about self-service, we talk about rolling updates, we talk about new maintenance styles, all associated with the cloud. What about metering? What about pay as you go? At what point in time does pay as you go start to really hit private cloud options? >> Sure, sure. >> I think it'll hit it sooner than later, but I think what's going to be interesting is the economics of it. >> Yeah. >> I think there's a supposition that pay as you go is a better model from an economic perspective. Not always. It depends on the duty cycle of your workloads. We are seeing movement where, when the workload is variable, that pay as you go model is, is a better fit. When things get where you can actually understand the application, optimize the application, optimize the infrastructure behind the application, a different model which is-- >> But doesn't it make sense to give the customer options? >> Yes, it does. >> Of course you do. But I think we always talk about clouded option and cloud transformation. There's both a technical piece and there's a cultural piece as well. I had Forrester on stage with me yesterday, and I said, "What is the one thing that "enterprises have to get right with cloud in 2018?" He said, "Procurement." And I can recall a CIO asking me one time if I could sell him compute by the nanosecond. I said, "Can you buy compute by the nanosecond?" And he said, "Touche." (all laughing) They are used to buying in big blocks in certain circumstances. They're used to the enterprise license in certain instances. So that's going to have to show as much change as, you know, we could do fine-grain billing today. Does it match, and does it fit the need? >> All right, we've got to go, but Steve, I want to give you the last word. We really didn't talk much about Cloud Private, which is your sort of branding and your offering. Maybe you could give us a little commercial on that? >> Sure. Yeah, we launched this last year, early November. We did IBM Cloud Private. So what we did is we took a core Kubernetes base, we extended it with some other compute models, you'll see Cloud Foundry in there, you'll see VM's in there as well. We took our middleware, we did a full containerization of it so you'll see a lot of rich stack of our middlewares, and then you see this automation layer's on top of it, our processes, etc., to kind of help you manage that overall environment. It's gone gangbusters. In just two months we had over 150 of our large enterprise clients. We got some of the great ones here with Hertz, MRN, etc., and getting great value out of it already. So we're very positive. Getting a lot of great press off of it, and we got a sales team extremely excited about it as well. >> Okay, Steve, Bala, great discussion as always. Really appreciate you guys coming on theCube. >> Oh, always great. >> Have a good rest of Think. >> Well, thank you again. >> Thank you, guys. >> We appreciate theCube. >> All right, keep it right there, everybody. We'll be back with our next guest right after this short break. You're watching theCube live from IBM Think 2018. Be right back. (electronic music)