 Computer Museum in the heart of Silicon Valley, extracting the signal from the noise. It's theCUBE covering OpenStack Silicon Valley 2015, brought to you by Morantis. Now your hosts, John Furrier and Jeff Frick. Okay, welcome back everyone. We are live in Silicon Valley for theCUBE Silicon Angles flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, my co-host Jeff Frick here. In Mountain View, one exit from our offline Palo Alto and of course, one exit from the headquarters of Yulia Packard and our next guest is Mark Interante, Senior Vice President of HP Cloud. Welcome and Senior Vice President of Engineering at HP Cloud. Welcome back to theCUBE, great to see you again. Good seeing you too. We love chatting with you because the insight's great. Last interview at OpenStack Vancouver was awesome. A little bit different vibe there, a little bit chiller with the patches. You were more, you were more jeans here in Silicon Valley. You got the nice shirt on, looking good. This is the 408 area code walk. You know, the 415 is a bit different. Okay, so I want to jump right in. So with Cloud's hot, we all know that. OpenStack is at an inflection point now. We're seeing great adoption. We're seeing some stability in the winners with virtualization, compute, a lot of stuff moving quickly and up into the winner category. Kubernetes is hot, containers are hot. What's the big deal? I mean, people want to engineer DevOps and they want to do it with standard components. What's the status in your mind from an engineering standpoint? Were you guys seeing traction on what's rising, what's trending? So I think what, I think one, containers, everybody's talking about containers, it allows people to get the kind of, the kind of stuff they had in virtualization, they get it faster, easier. And it allows people to develop on their, on their PC and their cafe and be able to decompose their services and then deploy those onto a larger system. So certainly there's a lot of momentum there. The other thing we're seeing is with the maturation of OpenStack is we're seeing people deploying OpenStack private clouds as an internal service provider and getting really big benefits from their application development teams. Because they have to have teams who want basically infrastructure services on demand in a virtual manner so they can go spin up, dev test and even production environments without having to have meetings. And the thing they're trying to get away from is they do not want to meet with IT. They want to be able to do the work they want to do. They want to get the resources they want to have and they want to be able to do their daily life without having to negotiate for, in the old world we would say fixed limited resource. OpenStack gives you the opportunity to see a near infinite resource for developers point of view. They can go spin up lots of different things. They can spin up databases and storage systems. But they don't have to go negotiate and get services deployed for them. They can do it themselves. They're much more empowered. It's all about the OpenStack. Are you happy with where things are engineering wise and what areas you see that are really accelerated from an engineering standpoint? Where's the coding getting done? Where's the action? There's a lot of good action in Neutron right now. I was at the, we hosted the operator summit a couple of weeks ago. Had about 250 people there. These people operating OpenStack and people were saying, well look, are you using Nova Networks or using Neutron? Now raise your hand if you're still using the old Nova Networks. One hand went up. Everybody else is on Neutron. So I think that was a good example of the fact that we passed the barrier on Neutron being production grade, having the features and capabilities at once. The refactoring that we did that allows things like VPN as a service, firewall as a service, load balancer service to be sub-projects and independently evolve has been a very good architectural change. And now those projects are picking up steam. Particularly, I'm excited about the Octavia project which allows very large scalable load balancing in the Neutron context, but it's decoupled. And upgradeability's made some big improvements this last couple of releases. So I'm seeing people move forward in that area. Okay, so the number one question that I get when I talk to customers in the industry, what's going on with HP? I'll say the splits coming up on November versus the official date. You guys have been operating as a split company from the enterprise team and the printer, whatever they would call it, HP, whatever that is. As separate companies now for a while, we kind of, everyone kind of knows that's out there. What does that mean for the cloud group? Does that make you guys a little bit more free to do things as a business as usual? What new changes is happening? What can we expect from HP Cloud in the new enterprise? Oh, sure. If you could share. Whatever, I can tell you what I can tell you. So one, I think this really is going to allow the enterprise team to be much more focused around its customer base, which is large scale enterprises and enterprise developers. So that focus for us is helpful. It allows us to be able to get more people's time around figuring out how to solve the core problems and making sure we've got good routes to market for the different solutions that the cloud BU is developing. So I think our partnership with the different other groups that are going to make composed HP enterprise has really accelerated here in the last six months. Both working with the hardware group, the EG group, the services group and the software group. What's going on with the trend that we're seeing with Docker containers? Because that was certainly a euphoric moment. Go back two years when you saw them at Red Hat Summit. It wasn't even the big deal at that point. Now it's kind of blown up to be a huge deal. Got Kubernetes around the corner. It speaks to this new abstraction layer that's being developed, which speaks to accelerated deployment. So I got to ask you from a perspective of DevOps. We're moving to a lightweight environment, it seems. We want lightweight. They want standard plug and play. That means the old model of heavy fat code is going away. So where's OpenStack related? Is it lightweight enough, or is it doing well and being lightweight, horizontally scalable yet with the integrated stack model? So I think that's part of the evolution that OpenStack's going through is making itself more componentized internally. If you look at the journey of the last three years, we've continued to refactor and build out smaller and smaller services. If you go back a few years ago, I mean storage is completely part of Nova. Network's a part of Nova. Almost all this started off as one set of code. And as we've continued to recompartmentalize and make it smaller, independent services, I think that's helped our agility. It's also helped our focus in driving up higher quality. This is clearly where people want to go. And I think OpenStack is well seen as a DevOps project. It is a project, not of only a piece of complex code, but it's one that people run in a DevOps manner. Many of the people who are developing it are running it. And certainly we are. So you guys at HP, and I was there back in the 90s, I worked there nine years from the 80s and 90s, and HP always was a leader at new stuff the biggest intranet at one point during the web days. First website on the Mosaic Browse, one of the first handful of websites. You guys always kind of eat your own dogs with whatever the metaphor is. Can you share some insight of what you guys are doing with OpenStack internally? And what are your kick, what tires are you kicking? What are you learning? And what can you share with folks? Well, sure, a couple of things. So one, we use OpenStack increasingly more internally in our development environments and in our, so we run something we internally call GOZR, externally it's called OpenStack Infra, which is a very large CICD system. We use that internally for all of our builds. And so we're continuously running against thousands of machines trying to basically to build and deploy our both upstream OpenStack and also the patches that we're holding and the things that we're doing for our customers. We also have a really good partnership with HP IT. That's one of the largest IT organizations in the world. It's probably in the top 20 or 50. I don't know exactly how big, it's quite large. It's pretty massive. It is. And they are early adopters of our OpenStack technology of our healing development platform, which is based on Cloud Foundry. So they're a great partner for us to give us real good, clear, honest feedback and clear- Great customer internally for you guys, right? They really are. And we're making strides over our next couple of releases to get more and more adoption with them. So I got to ask, I asked every question, every guess the same question on the cube this week at the OpenStack. So I'll ask you, does hybrid cloud exist? Is it like a concept? Is it an actual product? Is it an outcome? We know public cloud, what that is. We know what private cloud is. Hybrid cloud certainly is being the buzzes at an all-time high. But actually, is it a product? Is it a category? Or is it an outcome? Is it how you deploy? What's your take on that? I think it's a category. I don't think it's a product. And I remember when I was talking to one of my CEOs a long time ago, and he wanted me to build this thing. And I said, got it. Okay, let me give you a taxonomy. There are bugs, there are features, there are products, there are product lines. And there are industries above that. And kind of categories, which is an industry. I think hybrid cloud is a descriptor of a large way in which clouds are being deployed today. People use a variety of different ways to describe a private cloud. I mean, a hybrid cloud. Sometimes it's public and private. Sometimes it's either public or private and non-virtualized. Just, I have a bunch of regular gear that is all optimized to work the way it is. I want to easily connect that into the cloud resources. And you see companies doing that every day. People aren't doing wholesale rewrites of their massive systems. They're looking at their massive system and they're going, the massive system is actually composed of 20 subsystems. These two are the ones under stress. So how could they need to, they're evolving for some reason. Let's go evolve those in a cloud manner and still keep the other 18 just in maintenance mode where they are or in low evolution mode. And the other two, we'll put in a DevOps environment, we'll have it heavily evolve, but it will interconnect in. I think you see that all the time in the industry right now. So what makes it hybrid then? If you've got, say, one call between a public and a private or you've got a public cloud and it needs one piece of data that's not coming there. Is that just a connector with direct connect? Does it take some critical mass of connections or dependencies between the two to call it hybrid cloud? I mean, when you're talking to customers and they're trying to conceptualize, where do I put what? I think what's interesting is you don't just put things anywhere by themselves anymore. They're all connected to other things. So is it the definition of the number, intensity, variety of connections that defines whether that particular stack of gear is a hybrid cloud or a private cloud? Or a public cloud. And if it's semantics on some lines that are quite fuzzy, but we end up with is most large applications kind of fit into a broad definition of hybrid. And then the real conversation is at the next level down. How do I refactor or rewrite a subsystem easily and quickly and evolve it? And you get companies well within a mile of here that are deep on that evolution and have been for, I was doing some of this work eight, nine years ago at Yahoo. We had oldstools, we had Yahoo Finance which is one of my properties. It used to run 20 or so big subsystems. And we'd evolve three of them rapidly. We would use instant hardware which is what we called cloud back then. Because I mean, you get a thousand physical machines in a couple of minutes. It seems pretty instant to me. Well the web guys early on, they were all running the cloud native. That was native cloud essentially at the time. Exactly, so as we evolved that, we evolved it in those manners. And that's what you see kind of more mainstream enterprises doing kind of every day. And so I end up having conversations with DPs of Infrastructure or CIOs about let's walk through your portfolio. How do you want it to evolve? Where are the rates of evolution highest? Which tells me, how do you need to get a more agile infrastructure underneath that to reduce the friction? Right, right. And then build lots of connectors as a toolkit so that it's easier to move data back and forth. Right, so I was going to just follow up on that. When you're talking to the customers, there's clearly some stuff that there's just no reason to touch it. Works, it's sitting, it's been running forever. And clearly if it's a brand new application, right, you should probably want to go cloud first or at least tell me why we should it. When they're evaluating what potentially should move, what are the kind of the really key criteria that you review with them that says yes, this is worth the effort to move it from the existing infrastructure to more of a cloud based infrastructure? That's a great question. I kind of think of it as kind of portfolio analysis. You're looking at your portfolio of subsystems and you start to ask some questions. Like one is are there any, what does the requirements change per month or quarter? And sort it, is it really low? Okay, that's good. Is it high? Okay, this is under more pressure to evolve. What's my maintenance cost on it? Like maintenance per subsystem? Is it kind of low and stable and it's a small team and it's kind of under control? Or is there a pretty high maintenance burden? And those are kind of two of the, probably the first two factors you look at. And you start to say, okay, if it's, if it's in that quadrant with high evolution and relatively high maintenance costs, you either want to do automation, let's go make sure we can automate the deployments and the QA and the CI infrastructure. And how do you get an agile way to test it with an infrastructure that allows CICD? So Mark, we have one minute left. I want to just drill on and give you guys a chance to talk about what's going on with HP. Share what's happening with the engineering, the product, the vision. People want to know the HP story. Obviously the big news is the HP Enterprise is going to be a focused organization and the consumer stuff's going to be separate. That's pretty much out there. With the cloud group, what are you guys doing? Obviously any updates, any share any story updates, personnel updates, product updates? Well, we're obviously working on new releases of products. We've got some stuff in beta that I'll be happy to tell you about next time we talk. We'll have customer successes by then. When's that going to be released? Not soon. It's already in some customer hands. Probably by London, if I had to guess. We'll drive over by this right? Come over to London, hopefully. You can overtake a look at it. You know, we just bought a sticcata, which is going to help enhance our cloud foundry development application team. They're excited to have that team join us and I'm looking forward to kind of the next evolutions of our cloud foundry, more developer oriented products and families. What's the number one thing you're hearing from customers as you talk to customers? Features, product line updates, what are they looking for? What's the number one thing they're looking for? They're looking for experienced help in making a transition to be an internal service provider. They said, I want to do it, but like how people, processes, obviously technology, training, how do I go engage my app devs and say, I've now got a fully featured private cloud for you. How do we make this happen? After it's all installed, happy and running. It's day two. And how do we make that transition happen? Having a lot of conversations about that. You're lucky to have worked at Yahoo. You've seen that generation of people being first generation cloud native, DevOps that cultures ingrain and all the Yahoo, Facebooks, Google, you've seen it. Everyone wants to be cloud native right now. So I got to ask you, what advice would you give customers and people watching? What road map? How do I get started? I want some of that cloud native infrastructure. I want agility. I want OpEx. I want workload mobility. I want all those benefits. I have Yahoo, Google, Facebook, NV. How do I get there? You got, so here's how you do it. One, it's a journey. Step one is commit to getting a CI process in place and commit to automating your infrastructure. Do that and start getting releases done every month. If you're not, if you're getting them done every month, tell everybody in two months, we're going to do them every two weeks. And it's fractal. And just tell them, it keeps going in half every couple of months. And at some point, you'll have teams that have reduced all the infrastructure costs and complexity down. So you'll be able to say, I'm releasing every day. And you'll tell them, great. You're now within a factor of 100 of the most advanced tech companies who release 100 times a day. But 100 once a day is awesome. You never thought you could be there a year ago. So baby steps kind of, and if you will, go down and get iterated, get the CI going, get the continuous innovation and development agile, if you will. And just keep going. Double it every couple of months. There's no silver bullets. It's an iterative process. It's an iterative process. And just find the roadblocks. Find what costs time. Find ways to do it smaller and easier. Mark, thanks so much for sharing that insight here in theCUBE. Again, Senior Vice President of Engineering has been there, done that. Doing some great stuff at HP. Great to have you back on theCUBE, sharing your insights. We'll be back more here in Silicon Valley after this short break. This is theCUBE at the Opus Tech Silicon Valley event. Thank you.