 Live from the Mission Bay Conference Center in San Francisco, California. It's theCUBE at Google Cloud Platform Live. Here are your hosts, John Furrier and Jeff Frick. Okay, welcome back everyone. We are live in San Francisco for Google Cloud Platform Live. Google's developer conference was first, in our role event for the cloud team, kind of breaking out on their own from IO, which we've been to all the events there since it's founding, been great. This is theCUBE, our flagship program. We go out to the events and expect a civil noise. I'm John Furrier, the founder of Silicon Angle. Join my co-host, Jeff Frick. Our next guest is Greg D. Machelle, Director of Product Management at Google. He runs the cloud except for the networking piece. You have the keys to the kingdom, as we say in the product world for Google. Welcome to theCUBE. Thank you very much. I like to say I work on the fun bits, but I'm sure my colleagues who work on other parts, those parts are fun too. Great keynote up there. I really appreciate you guys. Really humble, really laid back, very cerebral in a Google way, but really delivering a methodology well thought out plan. You talked about the pass layer, and you guys claimed to be the original pass, first pass. Can you clarify this? We had this on Twitter earlier for the folks watching. Yes, I, it was- Please clarify. It was not- It wasn't intended to be a chronological statement. I would say we are certainly we're the first very widely adopted platform as a service when you look at App Engine and the companies that have built incredibly large scale applications on it, whether that's Snapchat or others. So yes, Twitter corrected me that chronologically speaking, there were some pass products before App Engine, and I humbly stand corrected on that. So we've been watching you guys for a long time, since you guys kicked it off, kind of on a private, private, directed availability, whatever the buzzword is. You know, getting out with the cloud, really taking a real soft approach to not making big moves, getting ahead of yourselves, really taking a well thought out, kind of a clean sheet of paper. Was that by design? Did you guys look at the cloud and say, you know what, we don't have the baggage that Amazon has, we don't have that lead. So let's take a step back. Let's look at what tools we have. Was that by design? Did you have a clean sheet of paper? Did you have something in mind? Is it the normal cadence of Google? Give us the insight into the early days. Well, I think when you look at us in cloud platform, you, we look at it a couple of ways. One is, we know from our own experience that there are certain just better ways of building cloud scalable applications, just based on our own experience in building search and Gmail and ads and maps. But at the same time, you have to meet developers where they are. You can't come out to developers with a product that doesn't look like anything they know. They want things that are, that are also comfortable and familiar to them. So App Engine is in many ways an embodiment of that. It builds on architectural ideas that we developed in the course of building out Google itself, how to scale front end services, how to do effective development. But it embraces languages like PHP and Python and that are very popular. The web scale, the web scale stuff. Exactly. And then you look at infrastructure as a service, which is all about meeting developers where they are, right? Taking a system you already know and how do you move it from running either on a server under a desk or in an on-premise environment and into a cloud environment where you can start to get better economics, better scale, better agility. So we try to be, we try to have a mixture in terms of how we approach the cloud platform. One part is building on Google's innovation history, building on things like BigQuery, containers, while also being very involved with the community as well. You can't be in an ivory tower until developers come to my way of building systems. You actually, you have to- They'll run the other way. No, you got to be in the pubs in the streets, not, you know, up in the ivory tower. And so it really is a combination of leveraging Google's history of innovation with just talking to a lot of developers. So you guys, you guys have the scale, that's well documented and totally buying to everything you're saying, but I want to ask specifically, was there by design saying, look, we could make, we got muscle, Google has muscle. What you basically said was, Google's got some strong muscle relative to scale and they know the developer community well because you guys are developers. You make your own cloud, you run your own cloud, you program your own code and you're very involved in open source. Great, got that. But on this strategy, four years in now, you're seeing the emergence of the cloud from Google. Tell the folks, how do we get here? Are you where you want to be? Is it by design? Cause by the way, you guys have a clean sheet of paper. Yeah, we're never where we want to be. I mean, we're never satisfied. I think, you know, my, one of my VP's said at the last cloud platform live, the one thing we know for certain is five years from now, when we look back at the state of cloud today, we're going to laugh, right? I don't know exactly how it'll be different, but I know that we will look at what we have now as an industry, not just Google, as an industry as incredibly primitive. That's a given in the pace of evolution. So what we're trying to do is innovate as fast as possible because I really think, I really think when you're picking cloud platform right now, what matters far more than what's happened in the last five years is what you think that vendor's going to deliver in the next five, right? You have to take a forward-looking view. So, you know, so we are trying to innovate on really all levels of the stack and that's what we talked about today. So you're taking your time. Mail the developer community first, job one. Yep. And bring the web scale with interconnecting other little hard things to do at first. Yeah, I mean, that's the first thing. Ultimately, if you look at most technologies, developers are the first ones who get excited about it. Who are the first ones who build mobile apps those developers not really enterprise and IT? Where did cloud first get its first adoption? It was the startup community. And that makes sense because if you think about it, those are the people without any legacy, without any existing systems. They start from a clean slate. But what you're seeing now is even inside enterprises, there are within the largest enterprise, there's new development going up. There's if not green field, green shoots, right? And we're seeing tremendous traction in both large companies and small companies who are looking to do more than just cut 10% off their budget year over year. Well, that's a weird comment before we came on. Jeff and I have a hot tub time machine stuck in the 80s IT. And there are a lot of companies focusing on that kind of focus, which really isn't innovative, just more of a blob of, you know, Op-Ex and some spend they got to manage, really managing costs. And then there's a whole another class of enterprise that want hyper scale, web scale, that are global. There are new ones. So can you take us through that? Because that's a phenomenon on the consumption side that really is only like five years old. Yeah, no, I think that's true. I mean, what's happened is there's a whole bunch of really interesting technologies that had their origins in the web scale companies. And they are technologies like containers, like Hadoop, like BigQuery, like platform as a service. These all started in the web scale companies and the enterprises have been now adopting them picking them up and incorporating them in their application development. So I think in many ways, the web scale companies are leading the way in terms of defining what the next generation of architecture looks like. And today we were sort of talking about that again as we talk about containers and container management and orchestration. These are technologies that companies like Google and other web scale companies have figured out how to do and there are, but there are benefits even within the enterprise. So yeah, I mean, I do think that this whole notion of enterprise IT, really good enterprise IT is not trying to just cut cost 10%. They're trying to add value to the business. Would you consider Salesforce as part of those web scale companies or they're just more of a, you know, data center driven company? Well, I think you have to give Salesforce a lot of credit for popularizing cloud and a lot of enterprises before, you know, before a lot of other companies were. I mean, even before us, before Amazon, there was Salesforce out there telling people that you should stop installing software, you should start doing cloud software delivery and just look at the revolution that's happened since then of all the other companies. Every ISV that's interesting now is a software as a service ISV, right? Whether it's, you know, Workiva, fast forward five years. How many companies do you think are still going to be doing their Fortune 500, their financial filings with on-prem software versus switching to a software as a service company? If you're not on SaaS today, softwares of service or have some sort of one on the cloud, you're nothing. I mean, to me in my opinion, as a startup at least, because you've got, it's the easiest place to start because why would you want to produce in any hardware? But even enterprises are starting to shift from running installed on-prem software to doing that same function through a SaaS ISV, right? I mean, those are, these companies are the same ISV ecosystem in the 90s that drove the PC. These are the equivalent of that. These are the ISVs to drive the cloud. Well, Erika from Bitnami was commenting on that lag piece where literally their business managers are running to them saying, I want you to do it for me, not IT, because they're talking eight to 12 weeks to do stuff. I mean, just who the hell operates on that anymore? That's a huge issue. Agility is a major factor, right? And it takes many forms. It takes, the simplest form is, I provision a new computer by a server by API instead of a purchase order, right? That's the simplest form of agility. But the other form comes when you really start to get more experimental and you try new things as a company and your developers start to just do A-B testing. And your system is able to actually scale with that. If you do an experiment that's wildly successful, the system doesn't fall over, right? So I think that whole notion of agility is the real driver of why cloud platforms are so interesting, much more than capex and op-ex. So we got a question from the crowd here. I asked the question on our crowd chat. Go to crowdchat.net slash GCP live and join the company. Cloud Market Share by vendor Google can really take some share. If they get the product right for developers and enterprises, it lists you guys at 1% Amazon at 22% market share from TBI. Floyd, an analyst said, can't get baseline developer traction when critical enterprise apps run on sophisticated yet proprietary tech they are not interested in. And he said to you guys, it won't happen unless you Google wins the enterprise that don't know what Kubernetes or what Docker does. That brings up the ease of use simplification piece to that comment there. Well, I mean, I think that comment addresses a couple of things, right? Which is that, where is containers in its development life cycle in Kubernetes? Are they something that enterprise IT is already deploying? I would actually argue though that if you look at the adoption curve that containers have seen in Kubernetes in particular, it is steeper than any other cloud technology I've seen since the beginning of Cloud. In terms of the ramp. Oh, in terms of the ramp. This is a vertical adoption curve nearly. So, certainly Docker for sure. Certainly Docker, and exactly. So I think although the comment that how many enterprises are using Docker today is a true statement. Ask me again in 12 months and I'll bet a beer that we're going to be saying something quite different. And the reason is. I'll just eat the beer because I want to have a beer with you. But the reason is because there are some technologies that developers love and they push IT to adopt them. There are some technologies that IT love and they pull developers to start using it. Docker and Kubernetes is this rare technology that actually has both. Developers push it because they get more productivity. They get to access a giant repository of tested software images. They get to run and deploy on their laptop and know that it works in production. CIS admins get. Well they also develop compatibility. Compatibility. But CIS admins get the ability to deploy to Google Cloud or private Clouds with the same system. They get out of the hell which is well why did the deployment fail? Let's roll back half of it. Let's try to sort all that out. So it really is I think that rare technology. Developers are going to push on their IT organizations and IT organizations are going to pull from their developers. That means you're going to see a much faster adoption I believe than some other technologies. So as Gartner would say we're the slope of enlightenment after the trough of disillusionment. So now you've got the timing. So what I hear you saying is Docker and Kubernetes gives from a timing standpoint couldn't have come at a better time. Absolutely. And I think you guys really hit the scene just as a compliment. I want to say congratulations. So your announcement at VMware. The Kubernetes really adopted at VMware which is a huge testament to the fact that they want to speed the adoption and are willing to essentially acknowledge Docker and Kubernetes in classic IT. And they have a lot to lose with that too. So that's a huge give at VMware. I think if you look at the whole Kubernetes consortium VMware, Red Hat, Microsoft, you know if you'd have like two or three years ago would you have imagined Google, Microsoft, VMware and Red Hat actively co-working on an open source project and it's awesome. And they've all been terrific partners because we all actually believe that containers and Docker are going to be the way applications are built. And it's in everyone's interest that this be an open source non-lock-in system. So I mean, I think every member of the Kubernetes consortium deserves huge amount of credit for this. I got to ask a tough question to you or a difficult question. And it's been come up on our chat here. App Engine has been kind of has a steep learning curve but once you get there it has some smooth sailing. So there's some nuance to getting up to speed which implies that they'd be pretty smart to code on it. I think you have a good developer base there. So that's cool, but as it goes mainstream how are you guys making it easier? What tooling are you providing? I'm a developer, I read the manual, I try to read, no, I read the manual. It's like they wanted to read the manual. But how do I get up to speed on the App Engine? How do I make my life easier from a developer? So I mean, I think the power of App Engine is that you just get to program your app code. You just write your application code in Java, PHP, or Python, and you don't even have to have a back end developer. I mean, we have numerous companies that simply start writing a Java application, deploy it out there, it gets no traction initially or no traffic, costs them nothing, then if they go viral, they're not having to carry a pager to keep the system up and running. The challenge has always been like with most... That's the dream of developers, by the way. But with most platform as a services, there's the trade-off and that is to accomplish that, you have to, I called it today, I said you have to color within the lines. You have to use the specific set of tools and APIs and operating systems and libraries that your platform supports. That's called getting down in dirty with the nuts and bolts of the infrastructure. Right, and so one of the things that we really believe is important is that you should be able to use App Engine, but you should be able to color outside the lines when you need to. You should be able to step out and say, for this part of my application, I'm gonna use C++ or I'm gonna use native code or I'm gonna use an open source library. So that's what this project we have called Managed VMs is really all about. It's about giving you the flexibility of a VM, so choose a two core, four core, eight core machine, choose an open source library, choose an operating system and still have the auto management capability of App Engine there for you. So really, when we look at what keeps most people from stepping into platform as a service, it's not the ease of getting use, ease of getting started because they love that. It's people looking go, oh, but what if I need operating system X down the line or what if I need an open source library and Managed VMs allows people to go ahead and start using App Engine and you know that if you run into some custom need, you can color outside the lines when you need to. Okay, so you got some onboarding with that, sounds like an onboarding, that makes it easier. But what about like horsepower? I need elasticity, I need to have really jam this thing up, I need to spin up some cores, I need the Compute Engine. How does it connect into the Compute Engine? I know you talked about it on stage. No, but that's my point, is these Managed Virtual Machines, they run on Compute Engine. Okay, got it. So basically you can start with App Engine, get the incredible scalability that companies like Snapchat and Secret get and if the need arises, you can easily step out to a Managed VM, runs on Compute Engine, you have access to all that sort of flexibility. And again, what keeps most people that we talked to have said I looked at App Engine but didn't use it, it wasn't that App Engine didn't have the capabilities, it was this notion of well, what if something happens down the line? I got to port it somewhere or I got to need to put it. And Managed VMs really gives you that sort of flexibility. So basically if you get the green light on a good app and get back end support, okay cool. Just in general, I want to ask you just, what's next? I mean, honestly, you guys are breaking out, you got a developer conference here, what's on the product roadmap? You're inventing, you're shipping the future and planning for tomorrow, what's next? I think the four major themes that we talked about today, compute, network, mobile, and big data. I mean those are sort of evergreen areas that we are really investing in. Obviously with the acquisition of Firebase and them joining the team, you can look forward us to investing significantly, both in continuing to help the Firebase team improve that product, as well as integrating it in with the rest of the platform. So you can look forward to see us do significant work around mobile. Compute, with Kubernetes and Container Engine, we just opened this up, we're going to continue to rapidly evolve in that space. Ride the ramp, if you will. Yep, exactly. Big data, back at IO, we announced Cloud Dataflow, which is really a revolutionary, completely managed big data service. That will be going, being opened up for more developers leading into the beginning of next year. So, I mean, across all, across the team. You guys are busy. Always. So I got to ask you the final question or final comment, big meaty question. I want you to comment on these three things. Corrimitalist security, multi-cloud, and mobile infrastructure. Three hot areas that people are talking about. And for the folks out there that aren't in the weeds in the industry, you know, inside the ropes as you are, we are, those are three hot buzzword areas. Primital security speaks to no perimeter, speaks to cloud. Obviously, you have multi-cloud. That's like, okay, just in case, I might want to move things around and workloads. And also, mobile infrastructure. It's more mobile-led than what's going on. So we obviously firmly believe in workload mobility. The essence of doing Kubernetes and open-source project was to ensure that when you're using Kubernetes to deploy containers into a cluster, that you can deploy those same containers onto Google Cloud Platform, or Azure, or your on-prem. We think you should pick a cloud provider on the merits of the cloud provider, not because you've got a lock-in on a particular capability. So we are huge believers in that. That is an essence of what we, of why Kubernetes is an open-source project and why we're partnering with so many people. On the mobile front, I mean, I think you're right, that all the interesting apps that have been built today are mobile apps. And they're not only mobile. They're mobile, they're social, they're real-time, they're connected, and they're incredibly difficult to create and manage. That's why the acquisition of Firebase and having them join Google is so important. Because at Google, we get how to build systems that scale to hundreds of millions of users. Firebase has a tremendous developer experience and incredible ease of use of getting started. So we think the combination of that is really going to make us be just the obvious choice for somebody trying to build a mobile app, whether you're a startup or an enterprise. So Peruminos IT is one of those things where security around the data, it's a little out of religion going around here, different sectors of views of what Peruminos is, how you operate Peruminos security, because the old way of having a perimeter is gone. If you, notification economy, API economy, APIs are persistent in enterprises now. So you're on-premise, you're talking to the cloud, you got stuff going in and out, mobile apps. What is the perimeter less, you know, state? Obviously Docker's great for stateless applications. No, I think the point is, is that customers need, there's always this perpetual challenge of, I have my on-prem resources, which I traditionally guard through VPN and lack of connectivity, but then I've got my cloud resources that are somewhere else. So the simple thing is to just start doing a VPN connection, treat them as one logical network. But we are seeing customers more and more realize that as they build mobile applications that have to access that data from outside the firewall, you're seeing more and more move away from protecting your network just by locking down IP addresses or those sorts of things and switching to trust the person and the machine, not the network connection they came in from, right? If I bring a laptop into your office and plug in a wired cable, why is that somehow a more privileged network access than me, an employee of the company, accessing across my phone? So I do think this whole notion of authenticating the user and the device, not the connection. It's certainly Google, that's the way we work with our networking, right? We, I have access to networking resources from any device I'm on, as long as I have my proper two-factor authentication, as opposed to having to run a VPN client to connect and then be trusted for everything, right? So absolutely, in terms of the way we see network evolve, the way we do that at Google, I think, is going to become. I think you guys should lend a lot of leadership there, a lot of conversation certainly happening in the area. Greg, great to have you on the queue. Thank you. I'd like to summarize the show for the folks watching. What's going on here? It's your big breaking out shows, not part of IO, although IO is a great event. So IO has always been a great event for us. IO is a place where we can come together and talk about everything Google's going for developers, across mobile and apps and browsers and Chrome and Android. But really, Cloud Platform Live is a place for us to come and share specifically what we're doing with cloud. And I think if people take anything away from this, it's again, we're really trying to help people build the next generation of great cloud applications, whether those are container-based infrastructure applications, or whether those are mobile applications, or whether that's big data analysis. We've really put together the pieces so that we're ready for your most demanding applications across any of those spaces. And also, we're just getting started. It's been about six or seven months since the last Google Platform Live we did. With each six months, the volume of the releases we do is greater than the last. And I don't see that slowing down anytime soon. Final, final question. Every company has a cadence, Moore's Laws, Intel. What's the one thing that makes a Google Cloud team different? I mean, as Google's got its own DNA, but within the Cloud team, what's that one feature that makes you guys a unique culture? Well, I mean, I think we are passionate about how developers build applications. And we at Google know we've been fortunate enough to build a system that Googlers use. And we experience, as Googlers, how incredibly productive it makes us, and how it enables Google to innovate rapidly. And we want every developer to have that capability that a Googler has. The Google goodness for the developer. Exactly. We think that should be able to every company, whether you're an enterprise or whether you're a two-person startup. And we just want everyone to experience that same capabilities that we get internally. Greg D. Michelli here, inside the Cube, director of product management. He runs the store. He has the keys to the kingdom. Whatever metaphor you want to use, must be a fun job. Thanks for sharing your time with us. Exciting stuff here live in San Francisco, the Google Cloud Platform Live is the Cube. We'll be right back after this short break.