 Live from San Francisco, it's theCUBE, covering DevNet Create 2017, brought to you by Cisco. Okay, welcome back everyone. We are here live in San Francisco for theCUBE's exclusive two-day coverage of Cisco Systems' new inaugural event called DevNet Create, an extension and an augmentation of DevNet, their classic Cisco developer program. Again, augmenting and creating some intelligence, that's AI, augmented intelligence. I'm John Furrier, Peter Burris, bringing a lot of intelligence here with Craig McLucky, who's the founder and CEO of Heptio, CUBE alumni, been on many times. Guru in the cloud, great to see you Craig. Thanks for coming on. Thanks so much for having me back on. Thanks for coming in. More importantly about this event is really the community extension for Cisco. Cisco ingratiating into the community of open source and developers in a big way, but not like brute force. It's very humble event, small event in San Francisco. Really you see the connection of app dynamics with the networking and again, developers want DevOps, they want infrastructure as code, they want the scale of the internet. That's the purpose, just thoughts on this event. Yeah, so far I've been very impressed. It feels like a pretty authentic developer-oriented event. The sessions so far have been quite accessible and generally pretty well thought through. I think Cisco's doing a really great job of actually doing exactly what you say, which is creating an event that's relevant to the developer that isn't necessarily tied to Cisco's interest and establish themselves in terms of thought leadership and actually creating a narrative that works with me. Yeah, one of the thoughts that we're talking about and conversations that we're saying on theCUBE, especially this year, is really the two waves that are clearly emerging in the digital transformation is cloud and data. And cloud being public cloud, private cloud, hybrid, essentially large scale resource data, creating value and then application developers really taking advantage of that. And you can't look further than containers and Kubernetes as a key thread to bridging these two worlds of pretty much unlimited capacity in terms of compute, obviously pricing, business models and operational models are different by vendor, but the emergence of multi-cloud points to me, the future. I personally don't think it's ready for prime time, but certainly I would say directionally correct. But hybrid cloud is reality. So developers are going to have on-prem and off-prem, but how do they connect it all? How do you orchestrate it all? This is a core challenge. I know you're working on Heptio and it's near and dear to your heart. You're thought of the state of the industry with respect to Kubernetes containers and how app developers can get that freedom without being a networking guru, really truly getting infrastructure as code without a lot of, well, I got to run this with that. I just want seamless cross-connection between applications. Yeah, I think we're definitely, you know, a couple of things are true. One is it's been spectacular to see the amount of progress we've made as a community in the last three years, going from a situation where we were just seeing the spark of awareness around containers as a framework to package and deploy applications, you know, into an environment to become really something that most every organization out there is looking at to solve both difficult, long-standing challenges in the outer space, but also to open up this world of multi-cloud to create opportunities for people to move their compute around in this increasingly heterogeneous world. And it's been interesting in the last year to see a growing awareness of the importance of multi-cloud. I think there's two things that've been really motivating that from my side. You know, one has been an understanding that it really isn't a one-horse race anymore. We're really starting to see a surge in effort from both Microsoft and Google. And that's generating a lot of relevance for folks who want to run in this multi-cloud world. And the second thing we started to see is is a legitimate interest in this edge computing phenomenon. As organizations are aware of the implications of increasing volumes of data showing up in their networks, showing up on premises, showing up these environments, having the flexibility to move compute into those ecosystems is huge. And so, you know, obviously I'm a little bit biased. I think Kubernetes and containers are an amazing platform that really, you know, tap into both of these growing sort of points of awareness. One is being able to create a natural compute fabric that decouples your applications and services from the cloud provider. That allows you to look at the cloud provider as both an infrastructure offering, but also to judiciously pull through services that are special that you might need. But then also having the flexibility to offer you up your own services and then move that around from environment to environment. It's been quite wonderful to watch that. So, take care. Craig, you've also got some good insight into architecture and you actually know the cloud game. You worked at Google and Google's got some great stuff. Got TensorFlow coming out. You're seeing kind of that going on its track. I would agree with you. It's not a one horse racing anymore, absolutely. However, there's a lot of pressure on the businesses. The customer impact to deploying this digital transformation is, pressure cooker's hotter than ever. I mean, Ford just fired their CEO, stops down 39%. He's two years into his transformation. How the hell does a CXO transform their business if they got a gun to their head? What's your advice to the guys out there that don't want to be the next Mark Fields who have to essentially run as fast as they can without disrupting the operations and also trying to perform top line revenue, which is a busy apps. I mean, it's a tough spot, your thoughts. No, it is really interesting. You know, I tend to think about IT as this kind of really interesting optimization problem. The thing you're trying to sort of maximize is velocity. You have to be able to use technology to solve core organizational problems. You have to be able to point technology at your business and you have to be able to move from this world where technology was being delivered in a traditional product fashion to a world where technology's being delivered as a service. If you look at Tesla, Tesla's no longer offering a car that's just a static car. It's actually almost a service, like the technology that's being deployed into Tesla evolves day to day. The car's becoming better and richer and more amazing. And so CIOs have to start looking at this as an optimization problem where you want to optimize for velocity. You have to maintain an effective posture around risk management. And then inside that, you want to achieve an acceptable SLA. This is the really interesting thing is that a lot of folks are looking for like four nines, five nines, six nines, whatever crazy availability you're looking at. Except that the higher you're pushing availability, the further back you have to pull on your velocity. And so for me, the most exciting thing in terms of, as I had these conversations with the CIOs that are looking to make this transform into this new world, is helping them understand this balance between code velocity and availability, reliability of the underlying systems. Understand the role that some of these modern, automated orchestration systems are playing as a way to drive up your ability to move fast without necessarily driving down your general services availability. And then looking at ways to transform the organization itself from being a technology organization that is throwing tech over the fence to a much more nimble sort of sort of smaller teams that are delivering up technology as a sort of services. It's an amazing time to be a part of this transformation. So we think, Wikibon, we've done a lot of research on this. See if I can find your statements in there and just get a sense of how you think about these things. Is that at the end of the day, a business institutionalizes work around the assets that are core to its mission. And John and I were talking about this beforehand. In the old days, the biggest asset, the most important asset was the hardware. And so you institutionalize the work around the hardware and then it became your application portfolio, whether it was SAP or something else, and you institutionalize work around that. Today our observation, here's the test, is that the asset that's most important is your data. And you're going to re-institutionalize work around the data and how you use that data and apply that data to a lot of different business activities. What do you think about that? Is data becoming that kind of central asset around which IT and ultimately even digital business gets re-institutionalized along the lines of what you're saying? Yes, absolutely. I think it's really important to understand that there's really two components to this. IT is information technology. It is literally just the process of making sense of data and information. And presenting it in a way that you can make effective business decisions. So we're going back to our roots in many respects. DP, data processing. But it's also about experience. And it's also about creating an experience for the customers. So I think at the heart of it, this IT transformation is around two things. It's allowing modern businesses to generate a better understanding of their customer through the leverage use of data. Based on customer created data. Based on customer created data and observable information about the customers. And then it's based on experience. It's using that to create and craft a richer, more satisfying, better experience for the underlying customer. And obviously data is central to both of those, but the experiential side of it has a lot more than you can't look at that purely as a data processing thing. There's a set of mechanisms that you need to do to create those wonderful experiences. And you can start leaning into things like artificial intelligence as a way to drive and improve experiences. You can lean into new form factors and sort of new ways to connect customers with their businesses as a way to drive that experience. And the products themselves are becoming increasingly evolved. Like I've certainly seen recently in talking to auto manufacturers this dawning awareness that the car itself has to be creating an ongoing and sort of richer and sort of more interesting experience. It has to be more intrinsically tied to the customers. So software and data are connecting. So software and data are coming together and software is allowing businesses to gain insight and then the data is allowing the software to create more relevant experiences and you can't really separate those two things. Well, software is data. The end of the day, software is data. Go back to Turing, that's what he said many years ago, but bring us back to Heptio for a second. So if you take a look at Kubernetes and we agree, Kubernetes in the last three months the cubes had what, 50,000 shows or something like that. And we've had a lot of very, very bright guests on it. And we've had, and Kubernetes has been a consistent thing. Containers are important. These technologies for managing and orchestrating these containers is going to be especially important. And Kubernetes is right in the mix. But Kubernetes kind of looks like an infrastructure almost, I don't want to say a nerd, but you know what I mean. It's just not, Heptio is taking it is how you use that a little bit better. And it's introducing some of these concepts of design, which is the one thing you didn't mention when you started talking about this stuff. How does design, experience plus design come into play here especially through a tool set like Heptio? No, it's interesting. At the end of the day, I think there's two components of design, there's designing for the users. Honestly, I want Kubernetes to be the most boring thing in the world. At the end of the day, I want a business to not think about the infrastructure. Like it just needs to fade into the background and become this invisible substrate around which they live. The water in which they spoke. I've said for years that the value of infrastructure is inversely proportional to the degree to which anybody knows anything about it. Absolutely. And so for me, my interest in Kubernetes, I don't want to necessarily show up and fade into the full brain of developers that are using the technology. I want to fade into the background. I want them to be focused on the design activities that are helping them do their work and succeed and create great outcomes. And so if I have one complaint about Kubernetes right now, it's too interesting. Too interesting. It's too interesting. We need to make it. Boring. Boring. We need to like. Ubiquitous. Well, no, I think a lot of people are working on it because I think they've identified it as an opportunity to connect things and make it easier. So there's work being done. People are funding companies and. Absolutely. So exciting. So no, there's a lot of work to be done. All right, what's the biggest challenge is technically Kubernetes has. If it's going to be boring, what has to happen to make it boring? So it's interesting. You know, I think there's a number of things that need to be done. You know, one of which is that when we were building Kubernetes, we created this configuration syntax, which is sort of YAML. It's effectively a simple serialization of the underlying Kubernetes APIs. And for new users of the system, the first thing they encounter is this what we call a wall of YAML. Or a yard of YAML to the head back. It's a very daunting experience. And so we're thinking hard about ways to change that. So create a much more elegant experience, have much better tooling, have the experience of editing that sort of fade into the background of the developer's punches. And then to your earlier point about design, making it really easy to use some additional concepts that other people have put forth. It's creating higher degrees of usability and discoverability for other pieces. Through projects like what Microsoft's been doing with the Helm project is really important as well. And so you can expect to see us make a significant down payment on trying to, you know, help to make a significant down payment on trying to address it. Well, certainly going to be following you with a lot of the Linux foundations the CNC apps got out there. Final question for you is thoughts on multi-cloud. What's your definition of multi-cloud? What does it mean? Obviously we kind of commented earlier because certainly there's not a winner take all cloud game. There's going to be multiple cloud players. There might be even specialty clouds. As things get boring in that abstraction layer, get simplified with developer friendly interfaces. Clouds will emerge as resource pools, but what does multi-cloud mean for you? What's the customer? How should they look at what multi-cloud is and what is the path to multi-cloud? So, you know, multi-cloud, it starts with heterogeneity. It starts with the ability to run your workloads in a variety of environments. You know, so it's heterogeneity first of all around the physical infrastructure provider, not being tied into a single provider model. No one necessarily wants to move back to the world of IBM circa 1985, where you're locked into a single provider and hopefully no one gets fired to buy that provider. But the problem with it is it really softens the amount of innovation we're going to see around that. You know, you have these single points of innovation. The second thing that I think about is heterogeneity in terms of locality. The ability to create something that runs both at the network edge, you know, all the fog computing sort of realm as Cisco has pointed to term. In a data center location that is, you know, potentially customer owned, so sort of in one of the localities, or in a public cloud, it's going to be about multi-regional support. Being able to build an application that you can run in the US geography and then in other regions that have regulated requirements around data mobility. So we have to hit all of those things. And inside that, I think there's kind of three key attributes of evolving sophistication that people need to think about. The first is your cloud is just solving an infrastructure outsourcing problem. And that's the most sort of simplistic way of looking at this. The second thing about cloud is it's a way to consume a broad array of interesting technologies as a service. Right, so it could be, you know, a simple VM, but it could also be a database or something else. It's moving the world from this situation where that thing that you're consuming as a service is being provisioned by a ticket, where there's an operator at the other end of it, to a world where it's being provisioned by an API. And then the final piece of it is being able to move your own infrastructure to that services realm, your own technologies, the things that are running your business, and deliver them as a service into your own world. And so for me, multi-cloud means hitting that level of heterogeneity, and then being able to provision arbitrary services at the end of an API, and then deliver your own services to the same fabric. Greg McLucky, founder and CEO of Heptio, former Google Cloud leader, certainly subject matter expert. Thanks for coming on. Great to see you again. Appreciate it. I appreciate it. Keep alumni. Laying down the epic knowledge here inside theCUBE, bringing you the inaugural coverage of Cisco's DevNet Create Developer Conference as they go out into the open source community with the full force of Cisco. Of course theCUBE's here. I'm John Furrier with Peter Burris. Stay with us for more coverage after this short break. Hi, I'm April Mitchell, and I'm the senior director of strategy and planning for Cisco DevNet. Hi, I'm...