 Live from San Francisco, California, it's theCUBE at VMworld 2014. Brought to you by VMware, Cisco, EMC, HP, and Nutanix. Now here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. Live in San Francisco, this is theCUBE, our flagship program. We go out to the events, we have an instructor, Susan Linois, some say the ESPN of tech, as Dave Vellante and I always say. We'd love to go out, talk to the tech athletes, I'm John Furrier with Dave Vellante, my co-host. We'd love to talk to the experts, we'd love to talk to the CEOs, and of course, get all the relevant content and data, share that with you, and our next guest is Craig McLucky, who's the product manager at Google, with really this announcement around Kubernetes with the Docker integration. Actually it's a very sexy, relevant announcement for VMware to be in the mainstream cloud, embracing Open, embracing Docker, embracing containers, really is a shot across the bow for their entire company and ecosystem that hey, they want to go cloud, they want to do it right, welcome to theCUBE. Thank you. So I've been to pretty much all the Google IOs there have been, I didn't go this year, but just because conflicts, but Google's culture, very engineering centric, very innovative, they do things their way from their IPO to whatever, but Google Cloud has been out for multiple years, Google Compute Engine, App Engine, certainly before that, real core competency in the cloud, and Google is the cloud. And we're the only cloud company. And certainly the ethos of open source, absolutely ingrained in Google's culture. They just plow through the fields, minding their own business innovating. So being involved in this announcement to me really is a game-changing statement for VMware, because Google's an innovator, and Google's kind of picky to who they work with. So comment on the Google relationship here at VMworld, why is it so important? Kubernetes, obviously open source cluster management, VMware essentially embracing, some say maybe extending, but you can't do that in open source. What's your take on all this? All right, so at Google, obviously, we have a growing cloud business that we're very serious about. We've put a lot of effort into building out a series of very high quality cloud products. And one of the things that's become increasingly clear to us as a company as we've progressed from supporting the sort of modern internet companies towards supporting more of a enterprise-type environment is that our customers demand choice. They really are invested in multi-cloud environments, multi-cloud worlds. And we have to find a way to meet them where they are and provide them a framework to actually move to a more progressive and more cloud-native way of thinking about building their applications. And so this announcement is really important to us because it gives us an opportunity to extend the reach of our technology and to take technology that's been born in the Google environment and pioneered over a decade in running at sort of internet-scale, hyper-scale, and bring it to enterprises in a way that is comfortable for those enterprises and let's just meet the enterprises where they are. So I got to ask more about specifically the propensity for old versus new. Google always does things have new way, as I mentioned on the intro. What do you like about the VMware announcement? I mean, why are you guys here? What's the big deal? I like the idea of, we are a cloud-native company and one of the things that I've learned over my time here at Google is that we can go and speak to a sort of a bank or a corporate customer. And if we enter that sort of situation, we provide very progressive, forward-looking statements. They want to go there but they don't know how. They don't know how to go from where they are today running in a mainstream VM-based environment with a Generation 2 architecture to get to that Gen 3 space. And for many of them, it's a bridge too far. But what we're doing here with VMware and with a number of other partners, Microsoft, IBM, Red Hat, and others, is bringing Google's way of thinking about technology delivery to mainstream enterprise. And so it's really creating that bridge so that people can move from the old to the new in a comfortable fashion. All right, so let's break down the news at hit. Yesterday, just take us through the quick highlights, read or digest what was announced and your involvement. All right, so Google a few months back introduced a new technology called Kubernetes. Kubernetes is an open source cluster manager licensed on the Apache license that is intended to bring Google style application management to a multi-cloud environment. And what we announced yesterday was VMware's participation in the Kubernetes program which meant that the Kubernetes technology, this new container-based cloud native way of thinking about developing applications is now available to VMware customers everywhere. That's a great part of the announcement. The thing that's even more exciting in a way is that VMware is not just committed to providing support to their existing customers. VMware is also contributing technology directly to Kubernetes to make it better. So they're doing deep integration with the Open V-Switch initiative to make sure that it works really well in the software defined data center. And what have you guys found in the open source community? Because one of the things that I found with Docker in breaking it down to folks that aren't super technical is that the community aspect of the code sharing is really a fundamental feature of this, right? It's not like about containers, right? I mean, we have the engagement container with CrowdChat. Containers do make sense in this always-on real-time world, certainly with the cloud and DevOps, but it's the DevOps kind of ethos of open source. Can you comment on why that's important? How the sharing is now part of the code base? I think it's an incredibly important idea because closed systems, if you're invested in building a closed system, if you're a technology company and you create this closed ecosystem, you're taking responsibility for everything. So you're taking responsibility for your tools, taking responsibility for the platform, for the infrastructure, for everything. With these open source initiatives and these open ecosystems, you're empowered by the ecosystem. You actually get contributions from everyone. Your technology gets better, but you also have an open surface area in which other people can innovate and find new ways to use your technology to accelerate it. So for instance, with Kubernetes, we use Docker extensively as our packaging framework. We love what the Docker guys have done to make containers more accessible. Explain for the layman out, the layman developer, how this works, just sharing a piece for Docker and specifically, can you describe that? Because it's a little bit unique, but it's still open source. So Docker is a packaging format and a runtime environment for Linux application containers. So what you can do is you can take all of your code, you can package it up into a container and then you can run it in the Docker environment. Now Docker is open source and the Docker crew have worked very hard to make sure that Docker works very well everywhere. So one of the lovely things about Docker is you can take it, you can build a package on your laptop, run your code, make sure it works and then forklift it into a cloud of your choice and have a very high degree of confidence that's going to run effectively there. And it's entirely accelerated by the open source community. It's been embraced by a very broad array of partners, everyone like Red Hat, ChloroS, you name it, folks are embracing Docker and actually bringing that open container format to their environment. I was talking to some developers, some really well-respected developers in Silicon Valley and they said, the thing about Docker is it's hard not to share in Docker because you're actually penalized for not sharing. What do you mean by that? I mean I don't, is it just like open source, sometimes you gave up a storage, you put your stuff around it and you say, okay, that's my product, I'm not going to share that, but in Docker it's a little bit different, isn't it? Can you explain that? So the interesting thing with Docker is that it's a couple of things. The first thing it is, is it's this runtime that lets you build something and then run it anywhere you want. But the second piece of it is, is that it's a platform which lets you build on other people's work. So you can take a piece of code and publish it to the Docker registry and then someone else can discover it, take it, extend it and use it. Now by putting it out there, you're obviously going to get a lot of use, but you're also going to get other people building on top of it and doing things that you might find useful yourself. This is one of the things that Google's consistently found with the open source communities. If you put something out there, people take it, they'll make it better and a year later you bring it back in and it actually makes your life a lot better too. So Docker's created this very strong framework for code reuse that's going gangbusters. The flywheel definitely is going on, absolutely. So this idea of right wants, run anywhere is obviously very powerful. What you said was that you're bringing this Google style of application management to a multi-cloud environment. So that to me is also a very powerful statement because you're defining cloud very broadly. You're defining cloud like the half a million VMware customers define cloud. You're including what they're doing on premise. Can you talk about that philosophy which is different from some others? Yeah, so for us, the reality is that we believe at Google we have the best infrastructure in the world. We believe that it's quantifiably better than anything else you can buy and we feel really good about the quality of infrastructure. And so as a business we're invested in succeeding by creating a platform that attracts people to us and lets them benefit from our infrastructure. So we are incented to openness and we find that our customers demand openness and it actually creates an environment where we have much more access to a much broader customer base because they want to run in multiple cloud environments. They want to run with two public cloud providers. They want to run in a hybrid cloud environment. And with what we're doing with Kubernetes is we're inviting the community to work with us to produce a Google style way of running applications that works well for them everywhere. We could do it ourselves, we have amazing technology but most of our customers aren't interested unless they can take it and actually get their hands on it and build it and integrate it and run it and use it. And so we're trying to do something quite different with our next generation of cloud offerings at Google. We're inviting the community to participate. We're inviting the community to get involved right at the early days and to help define the APIs and help define the tool chains and be real owners of the code that we're producing so that they can shape the shape of that cloud product that we offer. So Craig, you said your cloud you believe is quantifiably better than anything else that you can buy. Can you help unpack that a little bit? How do you quantify that for customers? You know, one example of this was some time ago when I announced Google Compute Engine we brought up a company on stage that broke the World Terrasol record on our basic data map. That was MapR. That was MapR, yeah. MapR's first big benchmark. We covered that, we had that. And so that was quantifiably better than anything that existed in the space. The cross-sectional bandwidth and network capabilities, the boot time, there's a lot of properties of the Google Cloud that are very difficult to replicate elsewhere. We have some really great technology. And so our invitation is to our customers, come and try it, we think you're going to like it. We're not going to try to lock you in but if we can provide a framework where you're running on open technologies, you're using open APIs, makes it a lot easier for you to come and use our platform. Can you talk about the security model and Kubernetes and how should we be thinking about security differently? I actually think that's a great question and I think this is a very topical point. There's really two schools of thoughts around container-based security. There's the one school of thought which says that the Linux container environment is sufficiently robust to support multi-tenant security and there's another school of thought which says things that's not. At Google we tend to be quite conservative in terms of the sort of security dimension and when I look at the Linux Cisco interface, it's a very broad and it's a relatively hard to defend surface area and we see one to two elevation of privilege exploits a year that would let someone that was running in a container actually break out of the container boundary. So for us on Google Cloud Platform, we always run containers in a single tenant VM that we'll pack those containers in really efficiently to get every ounce out of that physical piece of resource that you've got. Okay, so you guys, like I said, you're conservative, maybe sort of skeptics of traditional security models and maybe that's too strong a statement but how about network, how about the networking aspect? How should we be thinking about networking differently in this world? That's actually a great point and I think that networking is going to be absolutely crucial. One of the things I've really been excited to see with VMware is that they're bringing the SDN based model to Kubernetes. One of the hardest things for us is, you know, obviously everything inside Google is software fine networking and as we think about the way that we build and run applications, it's very easy for us to, for instance, deploy a set of containers and assign an IP and then do very fine game control of what can talk to what but our customers were struggling with this model when they were running in other environments. You know, some of the older architecture clouds didn't offer some of that flexibility around networking and so what VMware is doing with the Open V switch initiative and their NSX integration is actually providing deep SDN integration so that you can align the container with the actual network SDN definition and then your scheduler can actually wire everything together very nicely and so you'll be able to very easily and comfortably integrate full SDN sort of network partitioning, you know, instead of having to do weird port mapping between, you know, inside of the container, you'll be able to get a very effective way of actually controlling who can speak to what and creating these partition networks for your applications. Now, it sounds like Google's philosophy on cloud is very open, inclusive. I mean, you've got two ends of the spectrum. Amazon on one end says, you know, everything will go to the public cloud. You talk to Joe Tucci and it's, you know, hybrid, hybrid, hybrid. What's Google's position on sort of that spectrum? It sounds like you're less dogmatic, you know? On the left side of that spectrum. Well, we try to be pragmatic and, you know, the reality is we think that in the future most people will want to run in a public cloud. We think that the cost advantages, economies of scale, and outright performance is going to make it a no brainer for most people. But we want to be pragmatic about it. We recognize that it's going to take years to get there and there's a lot of constraints and frankly, a lot of customers want to have multi vendor relationships and so we have to be open. We have to make sure that we're not attempting to lock our customers in and that anything we do, there's a reasonable alternative, you know, out there in the open ecosystem to use. So I wonder if I could follow up on that because there's a prevailing sentiment with amongst enterprise customers that renting is always going to be more expensive than owning. Personally, I don't agree. I think the marginal economics of the cloud at volume are going to be so compelling that renting is actually going to be less expensive long-term and only because of the volume aspects of it. Having said that, again, there's a lot of debates there. You've said earlier, Craig, you believe that ultimately most of the world, much of the world is going to move to that cloud because it's going to be more cost-effective. Why will it be more cost-effective? Is it that sort of volume play? Is it that just you guys do things better? You have more resources? The fundamental economics. I think there's a basic economics. They run the biggest cloud in the world. I mean, Amazon and Google are the biggest monsters out there. What are the fundamental economics of that that are driving your thinking is really what you're trying to get to? I think there's several things. I think there's operational efficiency. We're able to, we have an incredibly streamlined supply chain management provisioning deployment process. So the operational overheads, amortized out over the volumes that we run are great. The economics of power, construction, location are going to fundamentally change things. If you're building mega data centers next to dams and you have preferred agreements, power is a major driver of cost. Most corporate data centers are going to be running on commercial power. That's two, three, four times more expensive than what you're getting with those facilities. And then you have, I mean, you have this purchasing power which is a non-trivial consideration. The ability to negotiate great rates with your suppliers. But a lot of it's also going to be just being able to bring differentiated technology to bear. It's hard for people to do fully provisioned, high-end SDN rollouts. It's hard for people to provision an heterogeneous fleet that's highly tuned to different workloads. There's things you can do at scale that you just couldn't do anywhere else. And we're going to continue, we reduce prices aggressively quite recently and that's benefited all of the cloud customers. And we're not done. We'll continue to push down the economics and I think it's going to become very favorable. Last question, John, if I may. So in history of the industry, you've seen a winner takes all. Microsoft won and the software Intel wins and the hardware will cloud be a winner take all? Why are we not? I don't think so. I don't think it will be a winner take all. I think that customers ultimately demand choice. And most customers aren't going to want to be in a single provider relationship. And I think as a result of that, there will always be choice in the ecosystem. And I actually think it's good for customers and it's good for the companies to just stay competitive and stay moving. Craig Mclucky here, product manager at Google. Google and the enterprise is not new but you guys are putting the full force on. We see the progress. Congratulations. Thank you. I'll give you the final word to end the segment. Share the folks out there in a quick summary, short and sweet why you guys are here partnering with VMworld. So we're committed to bringing Google style cluster management and technology to the enterprise. We're partnering with VMware and a number of other high-end technology providers to make sure that the technology that's been pioneered in the Google data center is available to enterprises everywhere. This is the Cube, Dave. I love the conversations in the media. Everyone talks about wars, cloud wars. This is really more cloud paradise. You got Google, Amazon and VMware, all creating value. This is really the action. You call it a kind of- Bringing that knowledge to the enterprise. I guess it's a cold war but the benefits are significant. So congratulations. It's great to see you here in enterprise show like VMworld with VMworld, congratulations. Google in the enterprise, Kubernetes, open source. The consumerization, the open source of everything is all goodness here at VMworld. This is the Cube bringing you live right back after the short break. Thank you.