 Live from the Mission Bay Conference Center in San Francisco, California, it's The Cube at Google Cloud Platform Live. Here are your hosts, John Furrier and Jeff Frick. Okay, welcome back, everyone. We are live in San Francisco. This is The Cube, extracting the signal from the noise. I'm John Furrier, the founder of SiliconANGLE, joining my coast. Jeff Frick, our next guest, is Brian Stevens, vice president of Google's Cloud Platform, the former Cube alumni, legend at Red Hat for many, many years, industry legend. Brian, welcome to The Cube again. Thank you very much. That's a good intro. My kids will be proud. Yeah, they'll be proud. So we just talked with Craig, like he's the product manager named Kubernetes, just totally geek talk. But the big thing we talked about was we're kind of seeing the old centralized mainframe cloud similarities where systems program and operating systems, the revolution in Unix, Linux, kernels, the old major innovation engine of systems, operating systems, in the client server, in these large systems, now becoming completely decentralized and in an empowering way. So you've been there, you worked at DEC, probably had the biggest OS, and I think the biggest network at the time with DECNet, Red Hat, obviously, made an industry. You guys have a really interesting view here. So are we going into this next generation of a completely federated, organically growing operating system? I think it's a good way to put it. I mean, I said it is the operating system of the cloud, right? But I think it's the roots of building reliable enterprise-grade distributed systems, which we all knew was important. But it was a different model when you had to hand a pile of software to somebody. And so now, all of a sudden, it's making everybody rethink it when you're actually running infrastructure service. So it is absolutely a new day. And I think to date, we haven't even scratched the surface of necessarily even understanding what that should look like, less so around how so many should use it. I published last night a market share study by TBR that Google's got 1%. And hey, you can double market share next year by going to 2%. Kind of like that. So I mean, relatively small relative to Amazon and the public cloud. But there's a bigger game going on here. So I want you to tease out for the folks watching about the Google opportunity in the cloud to understand you guys had a clean sheet of paper. You have a clean sheet of paper vis-a-vis, say, Amazon, who has legacy now, certainly winning with market share. And so what is the purpose of Google? We see them, we can move. Is there a strategy that you can articulate and help people understand the play, the Google angle on the market? I think a big part about it is today it has largely been about capex. And so the public cloud is a great service in its current form. But it's largely about, where do I put that virtual machine? So that's what we talk about. I mean, the number one service that we think about is for hosting virtual machines. And I think that that's, well, that's interesting. That's sort of the race to the bottom. Like on how cheap is hardware, the managing hardware. The real value that we've seen is just really getting started. How do you actually allow a developer to build an application that can store data, analyze data, add a rate in machine learning, as an example, and pull insight back at cloud scale? And I think when you can actually start to do that, then it actually starts to change the types of applications that are being built. That really changes the industry. Right now, we're really in around where we run workload. And to me, that's probably a slight improvement on hosters. But that only gets what we've started to design. And essentially now the model is not just a big chunk of software that you hand the customer. Also, it's not about one vendor anymore. You've got open source is now a leveler in terms of the power. So I got to ask you, are the platform strategies and execution different? The old day was lock in, win the platform, and monetize the hell out of it. Now it seems to be the tooling is the way to go. And if you look at Amazon, how they handle their retail business, almost a direct mirror to how they handle their cloud, which is give away, burn the village, win the platform by pure zero cost, and then tool up in packaging. So you're seeing that similar trend. Are you guys looking at that way? Is tooling important? Absolutely. Well, I think it's the mantra that you believe in that's really important. When you think about enterprise IT are smart. They've been trapped before, right? And they know that the investments that they make, they want to be able to move forward, maybe revisit those down the road. And I think that if you're not really fully committed to an all-in open platform, then I think you don't have a very strong value proposition in winning the next generation workload. That's why you've seen us publish a lot of APIs and whatnot as open source. The differentiation is in the service, right? It's not in the code itself. So I think you'll continue to see us to realize that open frameworks, open source is really where it's at from a developer perspective. We're just going to embrace that really hard. Yeah, a lot of people didn't predict Red Hat from being a viable business model. You were there for many, many years, over a decade. So Google has that same approach, not directly, you know, overlaid to Red Hat. But similar, right? Be open, but build value. What Red Hat's done, it's interesting, is they built over a decade SLA into the software. It's open source. I mean, customers like that. I mean, that's not, that's value we have. So how do you guys see Google Cloud being the same way where you push stuff out there, make the tooling free, add value on top? Well, you know, I mean that the Red Hat model was, you know, just such a huge value from the current environment of IT. The current environment was super expensive. Didn't work that well. Didn't perform that well. Because when you think about it, vendors, once they win it in transposition, it's all about growth. So it's what's the next thing they can do. And they don't usually reinvest back in the old thing that they did that you're already trapped on. So Red Hat absolutely took advantage of that by presenting a different model. I think the corollary to that, or the parallel to that, rather, is that we too are looking at, like, not how you can do the same thing differently or run it somewhere else. We're looking at it and saying, wow, look at the scale of the Google Cloud. Look at the network that we've laid down that spans where it grew over. If we can actually expose that and allow customers to actually use that infrastructure and use those services, what can they do on top of that platform that they've never done before? And so that's sort of the model. And reliability, too. You have to nail the reliability issue. So what I found interesting was the interconnect announcement today, which is very interesting. We were calling it on the intro, the Netflix problem. Everyone knows Netflix. And the service providers can throttle. They can certainly do deep packet inspection. Oh, it's so good Google, but there's a trust there. If you can connect peering relationships with companies, they can avoid potentially that risk. Is that... Well, I mean, we are even doing our 2015 planning right now. And the two top requirements that we have for all the Cloud Platform are reliability and user experience. So when you think about it, it's not about like, bring out the next feature, and how quickly you get the next feature out. It's really, reliability is the bedrock. And so we see that, I mean, you see that from search. It's in the core of the company. If Google search is down, a lot of the world is down in terms of capability. And so we see, too, from a Cloud Platform perspective, that increasingly you're going to be running critical workloads on there. Someone else's business. Someone else's business. And so it's your brand. So it's, you know, the day starts and stops reliably. And everything else on it is crazy. So it's interesting, Brian, you say reliability and user experience, but you guys are beholden now that everything is not mobile. Come along, but mobile first, really, to the carrier networks and stuff that's really outside of your control that as we had an earlier guess, you got a 75 millisecond pump just to get from the phone to the tower. So how are you guys addressing that? Is that part of your strategy into those two priorities? I'm sure there's probably other parts of Google that are working on and how to get better mobile connectivity because they too suffer even in the consumer lines as well. And where the mobile network is. So I don't think, you know, that necessarily, you know, those, that's sort of a level playing field for everybody. I think that part of it, you know, but what it is, it's a realization is on our part is that today if an application or a service is built and it doesn't have a mobile way of accessing it, then it almost doesn't matter that it exists. And I think that that's sort of the, that's sort of a new model for Enterprise IT when you think about it. In the consumer space, they got that. Right. But in Enterprise IT, in many cases, they're still building their old internal services and applications without, you know, a handheld. Right. And so part of what we're doing is actually just trying to provide, extend the reach, you know, of mobile applications to, you know, segments that traditionally are still writing web scale apps. The other thing I want to follow up on your keynote, which I thought was interesting, you know, you outlined some major shift points, some major trends, Act 86, Linux, virtualization, and then public cloud. And then you basically stated that now this containerization is as big as the shift as those other five. So I don't know if people really understand that that is adding a whole other layer to what they, a lot of people probably still perceive as really just kind of the cloud piece. Yeah. You know, what's funny is though, each one of those shifts that started 15 years ago, 15 years ago, they all build on each other. So when we look at, like, sort of our cloud platform, for example, it's building on Act 86 and Linux. Right. Next, well, I took advantage of virtualization, built a public cloud, and now today we announced that we're embracing containers. So in many ways, disruptions aren't things that, like, you know, take all existing technology and move them out, but rather if you don't take advantage of disruption and you're probably not going to be relevant or offering, you know, the opportunity that you could be. And so the container paradigm is interesting because it's, for the first time, we're looking at models where, you know, we can actually deploy lightweight applications that are portable. And that's been the bane of IT's existence for years. If you ever really spend your time with, in an IT department, you know, self as a vendor, I've never had to live that on a daily basis. But I spend a lot of time with them, you know, that, you know, they live in a world of pain, right, that apes out of the time they're doing the unglorious things, you know, where they're just rolling out new versions of software for no incremental value. It takes them six months to do it, you know, and it's hard for them to really get to the real meeting for the business. So containers is probably, you know, cloud is one that's going to help them a lot, but then the container aspect of that is really going to change that process from a testability and reliability perspective. And seeing them embrace it, they're actually enterprises before they're even products around or embracing the whole container technology movement and they're already retooling all their own internal apps with that demand. Yeah, so there's a lot of commentary on the containers. Want to just get your thoughts on this. Then I want to pivot into how you guys can get into the enterprise. Some are saying it's going to be a tough haul for you guys to get in there without winning the developers first, which it's pretty clear you're already going doing that. You're going to have to, are going to the developers to bring them the Google goodness, goody basket from your cloud, as well as solving the scale problem. The question here is, timing of the containers is perfect for this marketplace. Eliminates a lot of the needs for cross-vendor cloud slash infrastructure with comment management capabilities layered in. Do you agree with that statement? Yeah, yeah. That it's a cross-vendor cloud. I do. I think that, you know, all underlying platforms aren't created equal, and every application that you find inside of a container, you know, doesn't necessarily get the, you know, good housekeeping seal of approval on it, right? So there's still, you know, certification and value you have to do to understand the two parts. But the reality is that, yeah, the application and the opportunities have been far too married together for too long, and that's been really, part of it has been the ban of IT has managed that. So containers is the first time that there's been technology that's helping with an operational problem in quite some time. So, yes, the other comment from Mark Thiel who says containers also offer much reduced overhead improving server CPU utilization and performance. Right. Is that a small nuance? No, that is true, and that's when you look at sort of like when we talk about that Google runs a container architecture internally for years, that's a non-virtualized environment. So every app that's provisioned on any piece of infrastructure happens dynamically in a container without virtualization. Super likely. But this doesn't make virtualization dead, it means virtualizations are part of the architecture. Built in. It's built in. So like our cloud platform is based on virtualization underneath containers, and that's really valuable because what it does is it allows us to service the underlying infrastructure. So, you know, if we need to service hardware or if we need to service the underlying operating system, we just live migrate our customers' workload across. We see the virtualization being the adaptive fabric. It can be, it's a huge part of the maintainability of the underlying platform. Okay, so IT, you guys, what's your plans with IT? I mean, obviously, Google's pretty well thought out, pretty, you know, they don't just go shotgun, they pretty much take a target and approach to winning, and I want to try to nail that with you here. Is containers your way to backdoor into the enterprise, or, hey, that's bad, I should say, you know, because it's a heavy lift to win the enterprise, to get in there and compete. You got a lot of leverage, you got a lot of muscle, certainly, you guys have been doing containers for a long time. What's the strategy? Well, there's a lot of, when we're already there, right? So the enterprise in Google and enterprise in Cloud Platform aren't a new relationship. But certainly, you know, the brand that is Google has a strong affinity with developers. And sometimes when you have that strong affinity, right, the people think, you know, that's all you're talking about. On the front end, you guys have that lock-in. Yeah, so the, but the reality is when you start talking to enterprises, right, and they're here today, they're not necessarily looking at and saying, you know, I've got 30,000 cores and these are my applications on it. I want to lift it up and put it over there, right? They call IBM to run their data center. That's what they want, right? What, you know, what's happening is, they're seeing that they too are development organizations. So they're not developers over here, enterprise IT over here. They have strong development organizations that actually want to invest in the business, create new capability for the enterprise to allow the enterprise to compete. And usually that means that applications that new applications happening on the cloud doing something very different. Dave Elanth and I always talk about this because we're old enough to remember the 80s kind of, you know, stuck in the 80s opt-up time machine to the 80s. And back then, you had robust development staff, certainly in financial services, but still today. But then over time, the outsourcing client service brought in the management consultants. Right. And then you had a slow kind of like, you know, really a gutting of IT. Yeah, yeah. I mean, you had, you know, basically drone storage admins doing their job of Oracle DVAs who wasn't a lot of love in IT. And all of a sudden, now, an inflection point and a shift is happening at the same time. So I have to build on a development organization like overnight. Do you agree with that trend and you see that same thing? Somewhat. I mean, it landed in the land of package apps, right? So you ended up here. Here's the guys that run the infrastructure. Here's the package apps. And now I think the first wave of that was, all of a sudden, they're moving this off as a service. And that provides a compelling alternative to some of the package apps that have been running inside of my enterprise. I think in some ways, they didn't realize it, but that was a wake-up call that you can build better apps on the cloud, you know, with really good integration that has better operational performance. So what we're seeing now, though, is that, and then that together with some of the new WebScale apps that have been out there. Look at what, you know, Netflix has done and drop and other people have done. And so that, I think, has, in some ways, woken the sleeping genre. Shadow IT. You know? And so, but it has, it's now all of a sudden, but the imagination is happening. So, every enterprise comes in. Well, an empowerment. They think, now I could drive my business and actually not be a cost center. And they can. And they can. And they can. And they can. Because just the methodology of building apps where you used to be spec it, build it, ship it, sit on it, collect your 15% forever and ever and ever. Now it's like, ship it, fix it, ship it, update it. So we all agree that developers are coming into the enterprise. So the question is, what inning are they in? Are they, like, bottom of the first inning? Because a lot of, a lot of practitioners, which is, you know, we want developers in-house and we haven't seen this kind of movement since the mainframe days, right? Like in-house, quote, everyone's got in-house developers now. It seems that's the big trend. Yeah. Well, and you have to, like if you, you know, look at the FSI vertical, right? Where a lot of technology starts from adoption in FSI, right? And you look at, can you think of any other places where data analytics matters more than FSI? Whether that's for security, fraud, derivatives processing. All of those are, there's not enough cores in the world to really build that business. And so that their natural applications, their data scientists, their lines of business. So before Jeff gets to his question, he wants to ask a question. I want to get your take on the keynote and what's going on here. Analysts in Wall Street now, in particular, like to put things in boxes, but it's really hard to put people into a, pigeonhole into a corner because you've got integrated stacks. I mean, it'd be nice to put Google in a nice bucket called infrastructure as a service. But if you look at cloud paths, you're seeing kind of the verticalization of these stacks and different use cases. So there's no one stack. It is the Lego blocks model. So what's happening in the announcements today? What's really going on with Google Cloud? Is it the integration and the mashup of infrastructure services with paths, software connected? What is the big theme around? What's hanging the story? I think a big part of it was, was, you know, our realization that the audience is very nuanced, right? In some cases, somebody's going to come to us for data streaming services, and that's what's going to be the interaction. In other cases, they're going to come to us because I want to run, scale out global mobile applications, and I want to have, you know, geo-replication of all my data across the world. In other cases, they come because, you know, the Snapchats that already have built on platforms as a service, others want raw VMs. And then what we said today with Container Engine is that, you know what, we can build that life cycle engine, that manufacturing engine that builds the production line for anybody that's manufacturing software, and depends on that, but there's ridiculous for them to do that on-premise, and as they know, they can just, the testing resources take to build a developed line. Doing that in the cloud is the natural step. Instead of them having to figure out how to like, manage that with Docker and Kubernetes on-premise, here it is, it's running as a managed service, and it's going to be compatible with anything you do on-premise. So you guys don't necessarily want to get locked into one use case, because there are so many. It's kind of like a Google search. People can type whatever keyword they want, to get a search result. Yeah, I think that otherwise it's too narrow, right? I mean, it's back to us, which I think is kind of comical. It's enterprise or developer, right? I mean, enterprise, have you ever said that to a real enterprise customer? They were laughing, right? Because they just thought, they have... I know, the world has to change. Certainly the venture capital community wants to put people in buckets. It's like, I'm doing DevOps, I'm doing digital convergence, I'm doing all this stuff. Yeah. It's really a whole new world. Where do you put self-driving cars? Right? All over the place. But the other thing I just thought was interesting, Brian, is you talked about user experience being one of the strategic competitors for 2015. And Google's in a unique position that we talk a lot about the consumerization of IT. Not so, you know, in the context of people's expectations of the way applications behave are really driven a lot by your guys' applications as much as Facebook and Amazon and some of the other ones. So you do have a unique perspective in terms of frontline UI interface, interaction with users. And then to be able to take that into the enterprise, I don't think most traditional enterprise applications come at it for long user reactions. No, they don't. And we're not where we need to be either on that. But I think it's just sort of like you absolutely take a page out of some of the new designs done by consumer companies. And they get it. And I think that that's, you know, as we roll forward, usability is a top all for us, right? It doesn't make any sense to build the fastest, craziest service if it's hard to use. And that means infrastructure needs to be, you know, not only just sort of the time to value has to be short, but the service has to be understandable. You have to build a monitor. You have to build a deploy on it very easily. And so for us, I think that's, it's probably a culture change on the infrastructure side of the business. But it's one we're already well along to make. Yeah, it's really interesting. We did the Infor show with them. You know, they have a lot of little, specialty applications. And, you know, they've got a dedicated team just working on New York. Sure they do. And look and feel. You know, they kind of figure it out because that's what people expect. That's what they want. They go to the Apple Store. They want to see pretty things. It's true. It's true. And so I think that that's taught us a lot in notes. That's why I think it, I never thought about saying, you know, consumerization of IT could lend itself to UXD. But I think in this case, that's a really good example. Well, your example on the screen is very simple in terms of the atomic fishing guys found out their course, right? It's a very complete UI. Exactly. There wasn't a lot of choices. Yeah. You had like two dropdowns and they went and hit go. Yeah. Yeah. And I think we redid that and that was part of the showing people what container engine looks like. And I think we went through like three or four rounds of iteration on that just saying, we understand the technology too well. So sometimes it's hard to put yourself on the other side and say, how do I present that out? Right. And so that was an iterative process to say, let's make this easier to conceptualize. Yeah, and it was. It was good. Brian, so talk about your role. You're new to Google. What's your objective? You're getting your feet wet. I mean, what's it like in the Googleplex? Obviously you have experience with these guys working with them. In other capacities, it's certainly a red hat. But now that you're on the other side, what's it like? Yeah, it's been a blast. You know, I mean, it's a life change being by Coastal for now until we finish the move out to the West Coast. And then. Think about Palo Alto? Yeah, Palo Alto, South Coast, kind of area. Somewhere close to work. Palo Alto is very expensive. Yeah, that's right here. That's right here. But yeah, it's been fun. I mean, it's like, you know, I think Google's always been a, you know, you know, public company about a lot of how they do things. Like, I mean, if you saw the video inside the data centers, say during the keynote, I mean, I was as riveted to that as anybody was, right? And I worked there, but. It's a geek factor. It's just crazy. And so there's that much going on in every corner of the company from a technology perspective. So I'd have to say that. What sold you when coming over? Did you have to take a test, by the way? I didn't have to take a test. I did. Thank God I didn't have to take a test. This is the kind of stuff we want to know in the queue of personal questions. Did you take the free lunch? No, we didn't spend a lot of time on Google headquarters even. But what sold you? You had to make the hook Google's attractive. Was it a clean sheet of paper? Was it the personnel? Yeah, it was all that. It was just the, you know, look, I looked at it and said, I looked at it and I said, you know, what's the next chat? I always look at the breakdown, you know, things in five and ten years. And he said, the next five or ten years, when I look back on that, not what I want to do tomorrow, when I look back on that, when I'm glad that I made that decision. And so I decided, I really believe this is the next wave to bring about massive change both for consumers and enterprises. And instead of watching it from afar, why not participate? So when I had the opportunity to actually participate in a big way, yeah, I jumped on it. It was a pretty short conversation. Kubernetes means helmsmen of the ship. You know, as they say, if you don't want to hit icebergs, don't sail in the North Atlantic. So good luck with your new opportunity. Certainly, thanks for having us here on theCUBE. Brian here inside theCUBE live at the Google Developer Conference, the cloud platform live here in San Francisco where all the actions happen. Developers, this is theCUBE. I'm John Furrier with Jeff Prick. We'll be right back after this short break.