 Live from the FIIA Barcelona Grand Villa Compensator in Barcelona, Spain, it's The Cube at HP Discover Barcelona 2014. Brought to you by headline sponsor HP. Here are your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. This is The Cube. We're live here in Barcelona, Spain at HP Discover 2014. This is The Cube, our flagship program. We go out to the events and extract the scene from the noise. I'm John Furrier, my co-host Dave Vellante, our next guest, Tom Norton, VP of HP OpenStack Services. Welcome to The Cube. Thank you, John. So we just chatted with the head honcho engineer. I knew he was tired so I went for the hard questions early. More him down a bit. But a great guest, Martin Tarante. He's awesome. And the open source ethos is so awesome at HP. See you guys really doing a great job. So we're certainly impressed. Big fans and rooting for HP. But OpenStack is going through some interesting times. Dave and I always commentate that we don't know if it's a consolidation or a land grab or misfiring. Is it working? What's shipping code? All this is a lot of fun going around. Certainly a lot of fun. But we always say the rubber hits the road with customers. Who's shipping code? So I got to ask you what's going on with state of deployments? What are you seeing for that migration for the services and offerings out there today? Well, it is really interesting because customers are going to be a little more mature. But in one way or another, they're still looking at a singular workload or looking at an environment that supports a test case. I think it's really important with any new technology like this when you, especially in an enterprise case, you decide you want to understand the performance elements, understand the productivity elements, understand predictability, high availability, security elements of it. So doing a massive change is very difficult. I mean, the knowledge base is difficult. The integration aspects are difficult. Understanding performance can be difficult. So we see our customer base today where people are actually very interested because there's a lot of advantages in the enterprise. But they're being careful. So regardless of how long they've been working with customers now, they've been doing this for 18 months. And they've been working from trunk. They've been working with their own distributions where we have people who are just trying to understand it because they've got some requirement, maybe scalable storage or something. So for each one of those, though, it's a case study almost to begin with. I think that's where most people are. They're adopting for a particular purpose and they're starting with a singular idea of this workload is going to provide value and I want to use this new emerging platform to do that. So thinking about workloads and how customers decide to put what where at the highest level, you can think of, well, some stuff, this stuff can go in the public cloud, this stuff, I'm going to keep on premises and I'm going to build a brick wall around it. This stuff, I'm going to do hybrid. Now that's a very rudimentary discussion of workloads. Right. And so point number one point number two is and it's it's apples to oranges to compare Amazon to OpenStack. One's a, you know, for service, one's a platform of which you build products out of. Right. Having said that, Jerry Chen, who's a partner at Greylock came back from OpenStack Paris and we asked him to do that apples and orange, orange comparison. And what he said was very interesting. He said, Amazon's trying to be one thing for all people. OpenStack is trying to be all things to all people. Which again, I thought was kind of clever. Whether or not you agree with the apples to oranges. And I know this arguments around that. From a services perspective, it's clear that OpenStack has more use cases, as I described earlier, public, private, hybrid, et cetera. From a services perspective, how do you approach that? It's got to be somewhat more challenging. Can you talk about that a little bit in terms of how you evolve services? Sure. So I think from, especially from an OpenStack perspective, it's really going to be driven from a requirement. We talked about the workloads. You talked about performance or what you can put publicly. But there are a lot of levers in it. So if I look at a customer that may be in Germany, for example, it'd be very difficult for them to put their data, whether it be health care, or it might be public sector or something else, very difficult to put data in a public market, in a public cloud environment, and have storage extend that way. Because there are privacy issues, there are restrictions in terms of what you can take out of the country and what you can. So in that particular case, being able to keep the data private, surrounded and protect the data, but having the flexibility to say that's only the data platform. The application or the driver for the data, maybe it's the app, maybe it's information and analytics that need to be done. You can combine that with someone that can best host that because they've got better performance. You might have better storage performance on premise. OpenStack can allow you to network that or combine those approaches in a seamless way so that the consumer doesn't really get a look at where those two are combined. They just see the end result. And that's how we can use OpenStack. We can help people develop a storage base, people develop an application base, and be able to integrate those two, connect those two in a seamless way so the end result is still a unified application data platform. So I like that Germany example, because I've used it many times with the public cloud guys. And the answer used to be, well, we have a facility in Ireland that's part of the EU and that's good enough. And you talk to certain customers and segments in Germany. Now, Amazon just recently announced a data center in Germany. So my question hearing you describe that, is that maybe not enough? There may be other corporate edicts that I have to meet that just having a physical location there itself isn't even enough. Yes, I know I can't move it out of the country, but even in the country, I have to have certain edicts. Is that what you're finding and are you able to deliver services to assist customers in those more complicated situations? Well, I think it still exists. It's probably not the norm. You know, from what we've seen from a service provider perspective, because that's where things seem to be kicking in. And we've seen that kind of partnership between service providers, where they've got some that say, well, just talking today, within a country like Nepal, or a country like Pakistan, so they've got some restrictions, but they also have certain elements within that country that are problematic in terms of blowing out a service provider concept. But they can combine services to be able to have that combination of things with it. They may want to have the network where they observe from Singapore because it can be refreshed. There's a lot of compute. There's a lot of network bandwidth that may come in, but they might want to restrict the data. Now, to answer the question, though, is, yeah, there are certain business requirements or certain governmental compliance requirements that say, I want to protect the application as well. For example, very recently, there's been some compromise of mobile apps where people are taking, without naming anything, people are taking videos or they're snapping pictures. And the application that resides in the mobile device was actually a front to be able to capture the data before it actually got deposited someplace else, right? So they're saying, now I need to protect the app and wrap that app, control the app, regardless of what device may be using the app, and I have to protect the data. So it complicates the cloud approach. Very much so, yeah. But you can do that still in some ways where you can protect the app, protect the data, but provide it outside of the country. So in country, there are the regulations, but you'd be able to provide it outside of countries to countries that aren't going to require that the app would be wrapped and protected within country. So sometimes I get confused around the services at HP, because there's a lot of them, right? So you've got technology services within EG, you've got sort of former EDS sets of services, and you're specific to OpenStack services within HP's cloud group. Is that right? Well, so I have OpenStack and the development platform as well. So really, for Helian professional services is what we're running, and our focus is going to be on OpenStack and Dev platform. So how do you relate, connect to those other services organizations? So we're really a service provider too. So if it was a more traditional cloud delivery. You're the expert zone. Yeah, so if it's going to be an ES, and they're going to be serving up a virtual private cloud for example. And that, one of the requirements may be connectivity or scalability or something, and they want to use what HP has the expertise in is around OpenStack, I'll provide that service to them. The same way with technology consulting or HP software, we're going to provide those very specific deep dive experts who have worked in the enterprise and have done Swift in the enterprise and done it in a scalable way. Where we're talking petabytes of access to data. Where we're going to talk about compute. Where you may go from 100 nodes where you might have or 20 nodes per cloud and you have 20 clouds where you're going to have 400 nodes or go to 600 nodes and so on. So the enterprise adoption of OpenStack seems to be where the major challenge is today. Because a lot of the development work is being done at a usable scale, a testable scale, but the adoption in the enterprise where you would have a big service provider or a global telecommunications provider. They're all really interested in OpenStack because of the services that they can provide. But providing high availability and full tolerance and scalability, that migration to an enterprise grade type of version is where we can adopt helium and the HP approach and some of the true content that HP is embedding within the Healing OpenStack environment or dev platform. And we can help the customer scale that to a level that they want to scale it to. You're talking enterprise customers. Enterprise customers, man. And so you're starting to see demand for that. What's driving that demand? Well, I think you mentioned at John to begin with. There's a lot of consumption of businesses coming in and the reason is because there's a demand that says I want to be able to address a larger volume of customers. So mobile devices is really drawing a demand for access to applications. But it's not just apps, it's data. So if you want to, say you're a multimedia company, it's talking to somebody in Europe just last night and they're a multimedia communications and media company and they're doing all this video analysis work where they've got 20 petabytes. I think in this case it was five but there are others that can have 10 or 20 petabytes. And they want to find an easy way to provide search access to that as a cloud service where they can say I've got this repository of video and I want to analyze the video for a certain criteria and I want to be able to offer search of that video environment to that end user. Not necessarily just to a business where you're going business to business, you're going to go business to consumer now where you can do analytics or even machine to machine where I could say I'm going to offer this data pool and I've got somebody that's generating analytics within their service so they can run the service engine, I can host the data and so as a service provider you can expand very greatly when you have something like OpenSec. You can expand storage very quickly to consume extra data in retail right this time of the year. The retail data coming in from endpoint machines and from the web is huge but it's not valuable all year long. So maybe I want to collect information and I want to scale information as a retailer because I want to collect trends that run from maybe late September, October to February 14th, maybe January 1st. And I want to keep that data for a certain period of time and then I don't need it anymore. I can put it to tape. I can do something different with it but I want to run an analytic equation against it for a segmented period of time to scale up from one petabyte to three petabytes or from a certain number of terabytes and quadruple it or more and then scale that back can be very expensive in certain circumstances. So if you think of that kind of scalable storage opportunity and you know you've got a segment transportation is the same way. You know if you're a delivery company and you know you're delivering packages it doesn't matter whether you're a Deutsche Post or you're someone else. Federal Express, UPS, whatever the delivery company is from Thanksgiving to Christmas and January 1st, huge volumes of data. You want to be able to run analytics against it. You want to index it and you want to push it to tape. Well to do object storage over that period of time can be very expensive. So what we're doing with services is trying to understand, okay, how do I create that environment where I can expand that very quickly with low cost type of physical infrastructure below it and then contract it and maybe redirect that portion of it somewhere else but I'm doing something with Opus Act that allows me to expand and contract and that's growing in terms of its ability. You know we've got- So agile and flexible, what you're saying is that the developers can program on stuff that's important to business but having that flexibility with cost and technology. Yeah, absolutely because if you say you are say you're a data scientist for example you want to have access because you're going to run certain studies or say you were running an application this season you got a tax application and you have something else and you want to- There's always something seasonal, right? The world of clouds, the world's flat, right? And I think that's a very common use case and you can't do that if you're talking notes or you're talking numbers of environments. The licensing requirements can be very expensive if you maintain that all year round. Yeah, I mean Opus Act is definitely the Lego blocks model. We've talked about that in theCUBE all the time at the Opus Stack SV event we had. We had it, but Dave the quote that you mentioned but Jerry Chen, what he actually said on theCUBE was a year ago Amazon was trying to be one thing for all people Opus Stack was trying to be all things for all people. Which is what he's basically saying is Amazon is the one trip pony. You go there and it's it. Opus Stack has a little bit more legs in terms of use cases and a lot more legs. I mean people want to tailor Opus Stack for making and building infrastructure. Right. With cloud, right? With legacy stuff in the back end. So would you agree with that statement that that's one of the main objectives? Absolutely I do. So I think part of the quote earlier was agility is kind of tough to define in some cases, right? Because it could mean a lot of different things. But in this particular case being able to customize or address a particular business value of business issue is very important to the value of Opus Stack. So you could say something about, I mentioned retail earlier, you could want to do something in retail that you want to speed up the transaction but make more information available at retail. That's a very common use case. But with Opus Stack you can actually do more. Well I guess we always like to smile because we always roll our eyes. We've been saying this for four years now. Shadow IT really drove Opus Stack because I think the people we talk to, CIOs that we talk to in the Wikibon community were like all privately admitting to us that Shadow IT is running rampant in their organizations and they actually think it's a good R and D but they got to bring it in. And so Opus Stack became the hope, the bridge to cross to this new world. And I think that was an initial motivator that we heard. So I got to ask you, that is what people think. They want Opus Stack to be a really viable bridge so I don't have to get locked into an Amazon or another club. Now Amazon is trying to change that by going to enterprise. So that's where we're at right now. So talk about the use cases. Is it Shadow IT? Is it like push heavy computing system cloud, R and D, test dev? What use cases are going mainstream with enterprises that you're seeing that are taking the Shadow IT back in with Opus Stack or those little Shadow IT I define as non-mission critical stuff? Right, well I think- Ungoverned stuff. I mean I wouldn't call file sharing using Dropbox. It could be mission critical. Non-mission critical. Security holes left and right. Yeah, I think that it's still there though. But I think that it's less prominent now. I think people are saying that IT can be less on the defensive with the way Opus Stack is today. So I think taking a defensive posture that says I have to maintain relevance in the business, I want to stay important. Well they can take offense if it's clear that you're driving the car. Yeah, and you can actually take something like this and say, what are you trying to drive from a business? What is the end result that you try to get that you want to go outside the business for? And you can bring, and you can say, actually we can react faster. I can give you more than just infrastructure. I can give you more than just a database. You know, I can give you some extensibility. I can give you some other security elements. I can be flexible enough that I can segment some of the requirements of it. And I can actually control the network internally as well. Because Opus Stack, what we're seeing now is this really idea of using the platform which is Opus Stack and layering a development platform on top of it. So I've got the whole dev base. I've got things like Trove where I can start experimenting with big data. I can start looking at virtualizing different characteristics even virtual desktops, for example. I can look at how I initially did that and I can get ahead of that with this. I can react very quickly to requests that come into this. But I think even what's the most exciting aspect of it is you can actually say I can segment that and I can react even faster than I could have in by going external. Because you can have partnerships that you can establish ahead of time. You can network between different service providers. You can network between different IT providers and say that I can take that negotiation aspect away. You know, one credit card swipe doesn't necessarily give you everything you need for your business value. You may have to go two and three different places to get exactly what you're looking for. And what IT can say today is I can broker that. I can actually, using Opus Stack, have those predefined networks already connected where I can have a storage base provider and I can have a compute provider. And I can have someone coming in and doing network virtualization for me and I can monitor that. And I think the challenge for IT becomes that how do you establish that governance model around that? And how do you tune Opus Stack so that it can be connected? How can you tune it so that I've got a trusted compute model between those different partners? And that's why they're calling us today. So that brings me, there's one more challenge I want to throw into the mix there. I'm going to get your comments on the developers themselves. That's a big draw. People need more developers. I want to ask you, I asked Mark the same question. I don't know if you heard me asking the question. I just got it. Okay, good. So what are the top three targeted developers that you are looking at, not to hire, but that you want to engage to see the Opus Stack ecosystem be successful? Is there a categorical like mainstream developer? What's the top three targets? Give me the personas. Languages that they code in, skill sets. I still think that it's going to be mobile apps. I think is one of the key targets that we're still seeing because people are still, the mobile world is still driving. So if someone is coding or developing for mobility or a mobile application base, and it could address any business issue, but I think mobile apps is driving a lot of that. The second example I'd use is going back to the original premise that says, I think it doesn't matter whether it's healthcare or not, I think applications that are going to need access to data and have data that they can, that will create a variant of the information or the response from the data is going to vary. Say it's a medical application or it's an insurance application or it's an engineering application. It's going to go back to a repository of data and return some result. I think developers that are looking at high-scale data platforms and how you can take something which has been, big data has been talked about for a while, but you can virtualize that. Using things like Sahara and other places, they can actually virtualize that access today that used to be a physical constriction in terms of a data center size and being able to expand and contract. So Python, data wranglers, no JS. All that is becoming a more and more. Mobile apps might be something different. What are they coding mobile apps in these days? But for OpenStack itself, I mean. Well, I think our work is mostly geared towards extending what is part of OpenStack. So we don't do a lot of application development. Most of my work is understanding the requirements of the app and then going from there. So most of our work is still going to be work with Ruby and work with Python and other things like that. Okay, so there's really two talks. There's people working on OpenStack and then there's the consumers of OpenStack, the developers. Right. Okay, we're getting the break here. Tom Norton, thanks for coming on. Vice President at HP OpenStack Cloud for services, appreciate it. Thanks for coming on. Day two, coming down a little wire. Here in Europe, I'm John Furrier with Dave Melanthek. We'll be right back with our next guest after this short break.