 Live from San Francisco, it's theCUBE. Covering Google Cloud Next 2018. Brought to you by Google Cloud and it's ecosystem partners. Okay, welcome back everyone. We're live here in San Francisco. This is theCUBE's exclusive coverage of Google Cloud's event Next 18. Google Next 18 is the hashtag. We've got two great guests talking about services, Kubernetes, Istio and the future of cloud. A partner seen as the group product manager of Kubernetes and we have Henn Goldberg, director of engineering of Google Cloud, two amazing CUBE alumni, really awesome guests here to break down why Kubernetes, why is Google Cloud really doubling down on that? Istio, of right, who are the great multi-cloud and on-premise activities. Guys, welcome to theCUBE. Great to see you guys again. Thank you for having us. Always a pleasure and again, you know we love Kubernetes, a CNCF and we've talked many times about, you know, we were riffing and, you know, Luke Tucker was on from Cisco who loves Istio. We thought service meshes are amazing. You guys had a great open source presence with CUBE Flow and a variety of other great things. The open source contributions recognized by Diane Green and the whole industry as number one. Congratulations. Why is Istio so important? We're seeing the big news, at least for me, this kind of nuance is one dot is available, you got general availability. We're supposed to be kind of after Kubernetes made it. But now Istio is now happening faster, why? So what we've seen in the industry is that it only become too easy to create microservices or services overall, but we still want to move fast. So with the industry today, how can you make sure that you have the right security policies? How do you manage those services at scale? And what Istio does really in one sense is to expand it. It decouples the service development from the service operations. So developers are free, they don't need to take care of monitoring, audit, logging, network traffic, for example. But instead, the operation team has really sophisticated tool to manage all of that on behalf of the developers in a consistent way. You know, Hen and I did a session yesterday, a spotlight session, and it covered cloud services platform, including Istio. We had a guest from eBay, and eBay has been with Google Kubernetes Engine for a long time, and they're also a contributor to the Kubernetes open source project. They talked about how they have hundreds of microservices, and they're written in different languages. So they're using Go, Python, Ruby, everything under the sun. And as an operator, how do you figure out how the services are communicating with each other? How do you know which ones are healthy? So they, I asked him, you know, so how did you solve that complexity problem? And he said, boom, I used Istio, and I deployed Istio. It deploys as just kind of like a sidecar proxy, and it's auto-injected. So none of your developers have to do anything. And then it's available in every service, and it gives you so much out of the box. It gives you traffic management, it gives you security, it gives you observability, it gives you the ability to set quotas, and to have SLOs. And that's really, you know, something that operators haven't had before. Describe SLOs for a second. What is, why is that important? So you have network-level objectives. So you can see, so you can have an availability objective that this service should always be available, you know, 99.9% of the time. That's an SLO. Or, you know, the response rate needs to be, have a certain type of latency, so you can have a latency SLO. But the key here with Istio is that, as an operator, previously, Jeff was working, Jeff from eBay. He was working at the VM or container or network port level. Now he's working at the service level. So he understands intelligence about the parts of the application that weren't there before. And that has two things. It makes him powerful, right? And more intelligent. And secondly, the developer doesn't need to worry about those things. That's the power of Istio. And I think one of the things for network guys out there is that, it's like policy, it brings policy to the equation. Now I want to ask you a question on the auto injections. What's the role of the, how much coding is involved in doing this? Zero coding? How much developer time is involved in injecting the sidecar proxies? Zero. Zero. Zero. From a developer perspective, that's not something that you need to worry about. You can focus on the chatbot you're writing or the webpage you're writing or whatever logic you're developing that's critical for your business. That's going to make you more competitive. That's why you were hired as a developer, right? So you don't have to worry about the auto injection of Istio. And what we announced was really managed Istio on GKE. So that's something that Google will manage for you in the future. I want to add one last thing about Istio. I think it also represented changing the transformation because before we were all about Kubernetes and containers but definitely when we see the adoption, the complexity is much broader. So in GCP we're actually introducing new solutions that are appropriate for that. So Istio for example, works on both containerized applications and VM-based applications. Cloud build that we announced, right? It also works across applications of all types. It doesn't have to be only containers. We introduced some tools for multi-cluster management because we know all customers have multi-cluster, the large ones, so really thinking about it, how in a holistic way we are solving those problems. We've seen Google evolve its position in the enterprise clearly. When John and I first started talking to Google about cloud, it was like everything's going to cloud. Now we're seeing a lot of recognition of some of the challenges that enterprises face. We heard a lot of announcements today that are resonating or going to resonate with the enterprise. Can you talk about the cloud services platform? Is that essentially your hybrid strategy? Is it encompassed that? Maybe you could talk about that a little bit. Cloud services platform is a big part of our hybrid cloud strategy. I mean, as a Google platform, we also have networking and compute and we bridge private and public and that's a foundation. But cloud services platform, it comes from our heritage with open source. It comes from our engagement with many large enterprises, banks, healthcare institutions, retailers, you saw many of them here. We had HSBC speaking, we had Target speaking. We know that there are large portions of enterprise IT that are going to remain on premise that have to remain on premise because they're in a branch office or they have some sort of regulatory compliance requirements, or that's just where their developers are and they want to have a local environment. So we're very, very sensitive and knowledgeable about that. And that's why we introduced cloud services platform as Google's technology in your environment on-prem so you can modernize where you are at your own pace. So some of the things we heard today in the keynote, we heard support for Oracle Rack and Exadata and SAP, that's obviously traditional enterprises, partnership with NetApp, Cloud Armor, Shielded VMs, these are all traditional enterprise things. What enterprise grade features should we be looking for from cloud services platform? So the first one which I actually love the most is the GKE policy management. One of the things we've heard from our customers, they say, okay, portability is great, consistency great, but we want security portability. Right, they now have all of those environment. How can they ensure that they are compliant with the GDPR in all of their environments? How they manage tenants in all of their environments in the same way? And GKE policy management is exactly that. Okay, we're allowing customers to apply the same policy while not locking them in. Okay, we are fully compatible with the Kubernetes approach and the primitives of RBAC and roles, but it is also aligned with GCPIM. So you can actually manage it once and apply it to all your environments, including clusters, Kubernetes cluster everywhere you have. So I expect we'll have more and more effort in this area and making sure that everything is secured and consistent. What about auto scaling? Is that enterprise grade? Auto scaling. Yes. Yes, I mean, auto scaling is a inherent part of Kubernetes. So Kubernetes scales your pods automatically. That's a very mature, I mean, it's been stable for more than a year or probably two years and it's used everywhere. So auto, pod auto scaling is something that's used everywhere. The thing about GKE is that we also do cluster auto scaling. Cluster auto scaling is actually harder and we not only do it for CPUs, we do it for GPUs, which is innovative, you know? So we can scale and auto scale and auto implement, auto provision your GPUs, if you're running machine learning. We're going to bring that on-prem too. It's not in the first version, but that's something that with the approach that we've taken to GKE on-prem, we're going to be adding those kinds of capabilities. How hard is that going to be to go on-prem? Is it just an extension? You just got to get the job done? Or what timeframe are we looking at that? So the API that we've built, it's a downward API that works with some sort of hardware clustering technology. Right now it's working with vSphere, right? And so basically if your underlying technology has that capability, we will auto scale the cluster in the future. You know, I got to say, you guys are like the dynamic duo of Kubernetes, senior in the shows, you have the Internet Linux Foundation events. Talk about the relationship between you guys, you have in engineering, you have product management. How will you guys organize? You're moving fast, I mean just the progress since we've been interviewing you to CNCFs ago has been significant. Since we've started talking on theCUBE, you're seeing Kubernetes, obviously you guys have some inside knowledge of that, but it's really moving fast. How is the team organized? What's the magic internal formula that you guys are engineering here? You see in Google you mean? And you guys are working as a team, I've seen you guys open, is it just open source, is it the internal? Talk about some of the dynamics. So first of all we're working as one team. One thing I love mostly about the Google culture is about doing the right thing for the user. Like the announcement you've seen yesterday on the keynote, there are many, many teams that have been working together to get that done. But you cannot see that, right? You don't see that there are so many different teams and different product managers and different engineering managers all working together. But what I think where we are right now, I know is that really Google is backing up Kubernetes and you can see it everywhere, right? You can see with our announcement about Knative. For example, so the idea of portability, the idea of no locking is really important for us. The idea of open cloud, freedom of choice. So because we are all aligned to that direction and we all agree about the principles, it's actually super easy to move forward. So she is very modest, you know, this type of thing doesn't just happen by itself, right? I mean, of course Google has a wonderful culture and we have a great team. But I really enjoy working with Hen and she's an amazing leader. She's the leader of the engineering team. She also brings together these other teams. You know, every large company has many teams and the announcement at the scale that we made it and the vision that you see, the cohesiveness of it, right? It comes from collaboration. It comes from thinking as a team and you know, the management and leadership that Hen has brought to the Kubernetes project and to Kubernetes and GKE and cloud services platform is phenomenal. It's an inspiration. I really enjoy working with her. And the progress has been great progress. So I hear a lot of customers talk about things like, hey, you know, they evaluate vendors. You know, those guys have done the work and it's kind of a categorical way of saying, it's complete, they're working hard, they're doing the right things. As you guys continue this mission, what's some of the work that you're continuing to do? What's the work that you guys are doing the work? We see some of that evidence. If you have to describe this, someone says, hey, have you done the work to earn the credit in the cloud? What would it be? How would you describe the work that you've done and the work that you're doing continuing to do? What is that work? What would you say to that? I mean, I hope that we have done the work to, you know, to earn the credit. I think we're very, very conscientious. You know, in the Kubernetes open source project, I can say we have 300 plus contributors. We are working not just on the feature functionality, but we work on the testing and we work on the QA. We work on all the documentation stuff. We work on all the nitty gritty details. So I think that's where we earn the credit on the open source side. I think in cloud and in enterprise, well, you're seeing a lot of it here today. You know, the announcements that you mentioned, we're very, very cognizant. And I think the thing I like about, one of the things that Diane said I like very much is I think the industry underestimates us. Well, when you talk about, well, we look at the Kubernetes, if I can call it a playbook, it took the world by storm. You obviously solve some of your own problems, you open source it, develop the community. Should we think about, it's the same way? Are you going to use that sort of similar approach? It seems to be working. Yes. Doing open source is not easy, okay? Managing and investing and building something like Kubernetes requires all of effort by way of no just from Google. We have a lot of people that work in full time just on Kubernetes. The way we look at it, we look about the thing that we value the most, like portability, for example. If there's anything that we would like to make a standard, like with Knative, those are the kind of things that we really want to bring to the industry as open source technologies, because we want to make sure that they will work for customers everywhere. We need to be genuine and really stand behind what we are saying to our customers. So this is the way we look at things. Again, another example you can see about Kubeflow. So we actually have a lot of examples of where we want to make sure that we give those options. So that's one, it's one is for the customer. The second thing I want to actually emphasize is the ecosystem and partners. We know that innovation, not all of innovation will come from Google. And we want to make sure that we empower our partners and the ecosystem to build new solutions. And this is, again, another way to do it. Yes. I mean, we were talking before we came on camera about the importance of ecosystems. Dave and I have covered many industries within enterprise and now cloud and big data. Now you see blockchain on the horizon, another part of our coverage area. Ecosystems are super important when you have openness. And you have an inclusionary culture around building together and co-creation. This is the ethos of open source. But people need to make money, right? So at the end of the day, you guys are not a not profit. You guys make profit. Instead of the partners. So as the world turns to cloud, there's going to be new value opportunities. How do you guys view that ecosystem? Because is it more educational? Is it more just keep up? A lot of people want to be on the right side of history with cloud. And a lot of things are changing. How do you guys view that ecosystem in terms of nurturing it, identifying it, working with it, building it, sharing? What's your thoughts? Sure. You know, I believe that new technology comes with lots of opportunity. We've seen this with Kubernetes and I think going forward, we see it. It's not a zero sum game. You know, there's a huge ecosystem that's grown up around Kubernetes. And now we see actually around Istio a huge ecosystem as well. The types of opportunities in the value chain, I think it changes. It's not what it used to be, right? It's not so much, I think, taking care of hardware, racking and stacking hardware. It's higher level. When we talked about Istio and how that raises the level of management, I think there's a huge role for operators. It's a transformative role. You know, and we've seen it at Google. We have this thing called site reliability engineering, SRE, it's a big thing. Like those people are God, you know, when it comes to your services. I think that's going to happen in the enterprise. That's going to be a real role. That's an operations role. And then of course developers, their life changes. And I think even like for regular people, you know, for kids, for you and I, and normal people, they can become developers and start writing applications. So I think there's a huge shift. That's a huge thing. You're touching on a lot of areas of IT transformation. You know, talking about the operations piece, we've touched upon some of the application development. How do you guys look at IT transformation and what are some of your customers doing? IT transformation is enabled by, you know, this raising of the level of abstraction by having a multi cluster, multi cloud environment. What I see in the customer base is that they don't want to be limited to one type of cloud. They don't want to be limited to just what's on-prem or just what's in one, you know, in any one cloud. They want to be able to consume best of breed. They want to be able to take what they have and modernize it even if it's, even if they can't completely rewrite or even if they can't completely transform it, they want to be able, they want it to be able to participate. So they even, they want their mainframes to be able to participate. Keep up. Yeah. I had one customer say, you know, I don't want to have two platforms, a slow platform and a fast platform. I want just the fast platform. Exactly. So how about the future now? As we end the segment here, I want to get your thoughts. We're going to see CNCFs coming up to Seattle in a couple of months. And also, Istio's got great traction with, obviously with the support and general availability. But what's the impact to the customers? Because GKE, Google Kubernetes Engine, is evolving to be the single interface that's almost as ease of use. Because that's a real part of what you guys are trying to do is make it easy. So the traction layer is going to create new business models. Obviously, we see that with the transformation piece you were just mentioning. But at the end of the day, I got to operate something. I'm a network guy. I'm now going to be operating the entire environment. I'm going to enable my developers to be modern, fast, or whatever they want to be. At the end of the day, you've got to run things. You've got to manage it. So what does GKE turn into? What's the vision? Can you share your thoughts on how this transforms? And what's the trajectory look like? So our goal is actually to help automate that for our customers. So they can focus elsewhere. As we said from the operation perspective, making things more reliable, defining the SLO, understanding what kind of service they want to provide their customers. And our hope, again, you can see it in other things that we are building, like AutoML. Okay, actually giving more tools to provide those capabilities to the application. I think that's what we'll see more and more. So the operators will manage services and they will do it across clusters and across environments. This is a new skill set. You know, it's the SRE skill set, but even bigger because it's not just in one cloud. It's across clouds. It's not easy. They're going to do it with centralized policies, centralized control, security, compliance, all of that. So you see SRE, which is site reliability engineer, it's a Google term, but you see that being a role in enterprises. Yeah, that's what STO enables. And it's also knowing what services to use when, what's going to be the most cost effective, the right service for the right job, if you will. That's really an important point. I agree. I think security, I think cost perspective was something definitely that we'll see enterprises investing more in and understanding and how they can leverage debt, right, for their own benefit. The admin, the operator is going to say, okay, I've got this on-prem. I've got these three different regions. I have to be that traffic coordinator to figure out who can talk to who, where should this traffic go? Who should have how much quota? All of that, right? That's the operator role. That's the new role. So it's an opportunity for operations people who might have spent their lives managing loans to really transform their careers. That's right. There's no better time to be an operator. I mean, you can say, I want to be an operator. And you know, like. I can't tell you how much the SRE impacts our team, like the engineering team, how much they bring the focus on customer, the service we are giving to our customers. Thinking about our services in different ways, I think that actually is super important for any engineering team to have that balance. Okay, final question, just put you on the spot. Real quick answer. Great stuff. Congratulations on the work you guys are doing. Great to follow the progress. But I'm a customer. My customer hat on. Power in the hand. I can get that on Amazon. Microsoft's got Kubernetes. Why Google Cloud? What makes Google Cloud different? If Kubernetes is open, why should I use Google Cloud? So you're right. And the wonderful thing is that Google is actually all in Kubernetes. And we are the first public cloud that actually providing a managed Kubernetes on-prem. We are the first cloud provider to have a GCP marketplace with a Kubernetes application production ready with our partners. So if you're all in Kubernetes, I would say that it's obvious. Yeah, I see most of the customers wanting to be multi-cloud and to have choice. And that is something that is very aligned with what we're... Let the best cloud win. Open source is winning. Great to have you on a part of hand. Thanks for coming on. Dynamic Duo and Kubernetes, Istio. A lot of new services are happening. We're bringing you all those services here on theCUBE. It's our content here from Google Cloud. Google Next, I'm John from Davi Lonely. We'll be right back. For more day two coverage after this short break. Thank you.