 Live from the FIA 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 Welcome back everyone. We are here live in Barcelona for HP discover 2014. This is the Cube our flagship Welcome back to the Cube. Thanks. It's good to be here. You know, we're all from California, so we're all Like what a wall But helium is pumping right now. You guys have a great momentum We chatted our last conversation and we saw each other in San Francisco the big party Great success across the board you've seen the marching of the of the team more investment you guys are contributing a lot to OpenStack And now you had a new leader in Martin Mika's Nervous up there in the keynote, but it is playing cool. I'm proud to be here humble What's it like? What's going on? Tell us that here's the update. Oh boy Yeah, to me. We're really excited because we delivered helium. This is we built this out for For many many months and it's very exciting to actually get the full release delivered Along with the development environment all set up. We're getting we're onboarding customers onto it We're getting really good feedback and the conversations we've had of the customers here have been been quite exciting and people are really Pushing us to get more features added in and to build out So the first question I want to ask you is what's going on internally because we saw We've been chronic loss of the evolution of the cloud group from the rubbing nickels together and then SAR taking it over You know get the things going and now that really is bulked up What's going on inside HP you guys service in other organizations the conversion stuff is going great storage is moving great No, there's a lot of technology sharing so you're the engineering guy So yeah, what are you guys building in the cloud group and what are you sharing back or help us sure We're sharing everything we're totally about open source, but we're partnering very closely when looking at converged systems We're partnering with the networking team. So what we do is we realize or talk about Where do we want networking to go? We want all these new features and neutron and and and Nova and things like that So like okay, well, how do we go achieve this over the next three six nine months? And so I work with my peers and we're figuring out okay. You do this I do that and the networking guys are profoundly deep and Software-defined networking skills and my team doesn't have a lot of those we've got more distributed system skills So they team up and write certain areas. I write other areas And so that's a big part of it right now is we're really trying to broaden it from just the cloud Do you writing the code to many of us all collaborate? So you guys have a great growth on the number of contributors on the stack you guys do a lot I mean, I think that 30 40 people Are growing We asked about committers. I'm sorry to the numbers the committers are all about right engineers engineers Well, I'm trying to understand the how many people are engineering pure open source. Yes versus engineers like internal HP working on HP stuff. Oh, are they separated? Categorize that way. No, they're not like they're not really categorized We've got developers that are working on the compute infrastructure some of them for a period of months You're going to be working on a feature an element that we care about so may not be a sub-strain Other folks are going to be working on a pure upstream work So we don't believe in kind of sorting people out into different categories like that. We're all engineers We're all working on these distributed systems You know, I had to decide which goes open source Is there more like a Google model or 20% of time towards open source or people full model on our principles Or that that basically anything that is practical goes open source There's certain things that kind of aren't because it's a it's relative to like a piece of our firmware or something like that Is it would make sense or it doesn't make sense to have Something that's outside of open stack. It's like we're not trying to push stuff. It opens like it doesn't belong Okay, so when you say you're totally open source Yeah, I mean any contributions you make to open stack up is yeah, yes Cinder drivers for fiber chain Absolutely Okay, so Control panels are horizon, you know our control panels are horizon We're contributing SDKs into the open source SDKs that exist. We're trying to help make those better We do good members of those communities So as a customer of healing on a basically buying it to an open source platform Yes, that's right. And like I said some stuff maybe is not There's not much something doesn't make sense. I need to cover it so nobody would care Yeah, exactly. So compare that Other clouds Is there a spectrum or is it just like Closed open why well certainly there's a set of really awesome closed clouds, right? Where you get the API and you've got their world That's Amazon. Yeah, Amazon is an API Stuff yeah, exactly take from open source. We don't get much back Okay, yeah, that's right. That's their call great. And that's great business model. Okay, and then you're at the other end We are which is totally open except for a couple things that nobody cares about right and then I think there's some things in the middle They're coming there and I'm not completely familiar with all the company's internal Open-source strategies, but you could easily say look we're based on open source Like I use Nova use Swift. Well, we had a ton of bits around that So you're over using cloud foundry as it has So there's there's a spectrum and look we come down to the fact that we believe being a member of the community Putting as much of stuff upstream as possible is a high value us This is a this is a cultural shift for HP. I mean it's always been open. Yeah, but in terms of open source software Is this relatively new? I mean, I know Lennox I think I'm sensing a greater emphasis on open source for years There is a greater emphasis And I again I'm relatively new to HP so I don't know all of the last decade of history But if you go back even further go back into the 80s massive members of the Unix community the open software foundation In different communities like that. So then you're very involved in Unix and the evolution to Lennox for a very long time But now in the last couple years has been a really big focus around open source contribution and just open source as a way to build software that's Deep deep internet infrastructure. So from the standpoint of managing an engineering team What does that mean? How is it different and it kind of touches on John's question? And how do you leverage that open source community and versus your internal resources and how do you allocate that? How do you predict how much the open source contributions really going to deliver that you might not have to That's the key thing is how the predictability does go down And I'm welcome close source software is not very predictable either, right? But at least everybody is on your team and all of your do all of what you're doing You can order them around Do this by Friday But when it's open source and you're teaming up with a dozen other people or dozens of other people You know, you're not sure if this is going to land in March or May depending on how project goes If it's not if it's kind of halfway done in March, I'm not shipping it So what we have is less predictability of any individual feature But the overall aggregate velocity, I believe is as far as to carry out There's the number and the kind of people that we're having on these projects are people you can never bring into one company There's networking people from a variety of awesome networking companies. There's hypervisor experts that You couldn't assemble the opensack community under one roof. So is the aggregate velocity predictable? Yeah, you could say we're going to get about x interesting features per quarter I don't know exactly what you can kind of predict this at some level Yeah, but these features will probably land other ones probably won't so there's a little bit of art in there From there is managing the team separate and an open source project sometimes projects Water events, you know, the new project picks up It may explore a couple of different options before it kind of hones in on this is what it's really about Right and that may mean some predictability in the middle of the project before actually releases and you can inject Resources into the open source. All right, so guys get on this Accelerate it. I've turned those knobs. So for example, we in the last Six nine months, we really wanted to contribute more into neutron the networking project And we significantly up our our neutron contributions. We brought people in we found people who are interested in this work And we got them engaged on it and our and that really I think helped stabilize and increase the quality of the throughput of In this last year Because you know, we love open source work out We're we're a media version of open stack with the way we do our content everything's free We believe that you release stuff free good stuff happens and whether it's crowd check it's a momentum But that's the kind of thing we believe in so we're with you on that So I got to ask you the question of the big but And the plug the the big but is how do you explain to folks out there how you're successful with open source? No, Martin is obviously involved. You have a lot of great people on the board HP has good culture with open source good alignment there But how do you manage the successful transition to shipping product in an open source framework? A lot of companies are trying and they say they want to do it don't know how to there's no real You know how to play with open source for dummies book Maybe there is but but what is the secret to your success and what are your challenges and how do you manage that? That's a great question. I think I think some of the challenges are that the code evolves rather quickly and it's only Has stability areas so any since months releases for open stack. It's most stable in about a month or two During that room because after that you're adding more stuff in it becomes a bit less stable You're adding to features so you've got to understand the dynamics of open stack And I've been lucky enough to be having been an open stack for four years now And so I've been I think released everything since Essex It's kind of like an organism you kind of know when you inject new you know something's a reaction And you got to kind of have a feeling for it. Yeah, absolutely cadence if you will there's a case There's certain points where you kind of get a feeling that the code is gonna be more stable around a month or so after the big So I guess the question is kind of like the entrepreneurial question. Are you born with that? Inherent skill or you can you learn it you get totally There in fact, I'm happy to we're happy to show this kind of stuff We've learned I think it's just helps make everybody's products better. So what's what do you guys do? What's the secret besides the cadence so firing? Yeah, hire people to love open source I hire people that both have software development experience and an online services experience So myself for example, I built software. I've also run services like for example, Yahoo Which have a continuous integration continuous delivery model So we do continue and you can get on the back channel and rip someone apart and say hey Here's your code done. Look at the IRC channels. We love the IRC channels Online collaboration, too. Oh my we're very much in IRC So we've got all kinds of different channels for different projects if everyone to go Hey guys, what's going on or where are we here? I just hop on the channel and see where things are It's a very distributed team. We're In Have it as in countries and we're many locations in the US So it's highly fortunate. So you got your budgets for 2015 How did it look you guys investing more absolutely investing more. I am I'm actively hiring folks So if you're out there, do you want to do some great work on open source and work it open set come talk to me So what areas are you looking at specifically I see up and down the stack or there's already focus areas that you're hot on You're you're you're zoomed in like a laser on well No one's given up the names on those they're hard to they're hard to hire right now. They are hard to find They're hard to hire. Oh once a once the bubble burst. No Internet So I got to ask the technical questions now around and then strategy questions if you will just an engineering around NFV So we were talking with Saar earlier NFVs got his new business, you know, they're huge telco opportunity Really perfectly fit for open stack, right? So you actually get Amazon you get Google out there with their own clouds They're not gonna be really kind of warming up to the telcos anytime soon So that's pretty obvious for the people who know their business strategy Almost direct competitive Open stack seems to be a great fit So what's going on with NFV? Who writes what code are you writing that code? Is it all open source? I mean, I'm gonna say you need to see clearly on cloud team doing it Yes, I think the Saar just integrating everything together. No, Saar's got development teams I have development teams Saar is so our collaborating going like He needs certain things For NFV for his customers and his telco customers Okay, here's how we can help Nova evolve in a specific way and it's more natural for us to do that work There's lots of things about VNF's virtual functions and things like that are much more natural for his team to go focus on Great, so we're really trying to harden it and make open-sack more Fit for purpose for those kind of tilt. So what do you guys talk about in your management meetings? When you're sitting in there with the as execs you're sitting around the table, you know relaxed having a good time What's on the agenda? What do you guys talk about the most is it hiring is a market penetration custom first thing It's customers. We're talking about, you know, who's had what customer success What are what are we hearing from the from the people we're talking to? All of the all the executive team is very very into that including myself and my leaders also second thing is like We're sure projects. How are we are we getting things done? How are we looking at upcoming releases and how are we prioritizing our work? Those are quite the two main things and how about growth you guys talk about, you know, injectors and hitting your targets growth targets and you guys talk about growth numbers and stuff. We do yeah, and so You can share some of that stays in the vault Can't talk too much about that. No, we're growing very rapidly. We're excited about both the helium growth areas and the cloud system and it's You know, it's startup growth rates. This is nice to see so Meg wants velocity, right? Yes, you got you got a budget good budget, but it's not limited Yep, so I wonder if you could talk a little bit about sort of where you put your effort You got some references out there. You said customers really what's driving What are they asking for me this this is there's a good reference model obviously in the public cloud You know, you got to have a Object store you have a compute you have a block store you got a networking whatever it is But what are really the priorities to talk talk roadmap a little bit, okay? Okay, so I think a couple of things So one we want to get it to Our public cloud scale. So right now as we build helium built it First target seven cases is more for the private private cloud scale So as we're building in more and more features and capabilities to make it a carrier-grade or service provider level Offering there's things that you would need if you're managing Tens of thousands of machines that you don't need if you're managing At a more private clouds that are related to management and automation management automation just bring it for example When you're at a public cloud scale, you're bringing in racks of machines Every week or every day is on basis, right? So I mean it's so literally all the pre-certification. How do you? Connect that new capacity in certify that it's good. Make sure that it's effective before you turn it on There's a ton of automation you can do just right there It's stuff that you know, if you're running a private cloud of many hundreds of nodes You may have to pass any wants a quarter twice a year. Is it requiring you guys to engineer some more custom? Hardware that you can't necessarily get from HP off the shelf not yet. No, okay. Do you do you think it will? It's a mega scale. I've been pretty impressed by the variety of hardware. We offer actually I'm becoming more and more aware of that But you know, we have our cloud line. We have our SL and DL lines The cloud line is really optimized for minimum footprint Maximum compute per power and so it's very similar to what So let me ask that differently because years ago people would say oh public cloud guys They just use white boxes and see gate disk drives and that's it and that's changed so so And you've got a big portfolio to choose from so You're not building this just on where are you on white boxes and off-the-shelf disk drives We're building it on HP here. Yeah, so Full spectrum the full spectrum. Okay, and then so the the challenge is how do you get scale with that diversity? Yeah, so we get so a part of that Maybe that's what a public cloud Yes, yes, absolutely. So, you know, we get the discounts from our HP partners and the CL line It's quite cost-effective which is really based on the Foxconn kind of open designs And so they're designed around not having the extra chips. You don't need if you're only running a cloud offering So that just seems to work pretty well So kind of an open reference model that you guys can use that they can build build to it. That's right Yeah And then part of what we think we're trying to make helium work the best on HP here so that we're trying to make each year here safe And there's additional features and support within HP here that we think we can be very fast very fast and effective And building up this product. I always use that I live in an old house So I use the old house analogy guys obviously going you and many of your Competitors going after a hybrid cloud strategy at any time you do anything to an old house You got to connect it and the new to the old and that's the hardest part So I wonder if you can talk about that. Is that a challenge in the cloud land and how do you get around those challenges? I think I think it's a lot of features that we are building in have built into helium one And we're going to be building in the future versions of alien that make it more Easier to interact a couple of things we shipped in version one active directory say active directory integration for Easter Making it very easy to manage your users. We have ESX support for that hypervisor We're trying to be very open to the very common very popular Enterprise technology so we can integrate into a very effective way That's a challenge though, right in terms of in terms of getting to scale isn't that because you get so much diversity in your customer base But we're I mean we're building a lot of things for private clouds much more than a public cloud scale That's not the only target here. Yeah, sure, right, right, but that was sort of the next wave, right? Challenges how do we get the scale? Yes, that's public cloud scale, right? And obviously you want to do that so you can get you know the marginal economics go to zero or close to zero as possible, right? And do you do you have to get there though? In Kenya with that diversity or do you not have to because you have an install base that will consume that that diversity? I think two things one we want to get to a larger scale for two reasons one for ourselves for our offerings to many partners and customers that want to offer healing Clouds either private cloud more very large private clouds or public clouds For themselves, so there's a number a number of different cases for running it at a much larger scale We have a couple minutes left Going forward you got good momentum you guys have done great. I know you guys worked hard been following Your you guys moves great stuff. We're really excited to see that now. What's next? It's gonna get harder It's like the journey just begins with funding. You got funding you always had money But like when you get more just more gas for the for the rocket fuel I mean, I think I think the next set of challenges or how do we execute a multiple sign-up simultaneous prop? You know projects and initiatives really a lot of this year is it isn't healing and getting integrated into more and more parts of the HP business streams And so in that case we're getting more integrated with the minute of different hardware systems you see here and That means we need to sync up with their schedules and their Their goals and their overall integrations. So if that's I think that's a big a big challenge new customers to from that Absolutely customers All about the community is one of the things that the you guys have to also win on is building a community around That's a challenge, right? I mean, I mean there's a community out there. You got to win them Yes, win them over and gratiate in grow with them What's the plan there? What's just Didn't you visibility on that you can input on that? I mean, I don't have any I don't have much visibility into it. The main part that I do focus on is, you know after his open-stack application developers and getting more of those people a trained up to be able to use our infrastructure and our development environment And I think that'll be a big a big growth area for us in the next Which developers would you think if you could put together the top three list of wish list of genie in the bottle kind of thing where Your business success to be successful aligning with what are the top three developer categories? I say I think the I think getting getting Good strong enterprise developers who are very used to Java and dot net Having them start to use our platform and of course as we roll it out into their enterprise I think that's a huge deal for us second thing is for those sasper those ISV's they're transitioning their software into either ISV of the cloud or a sas model or make sure we're supporting them Very very well, and I would stop profile of those guys There's a there's a pretty high tons and across the board C++. That's one of the challenges It's a pretty wide profile. There's a lot of people that are doing stuff in And C++ actually and so you and it's not typically cloud oriented so helping them re-architect with queuing systems and Databases on demand integrated stacks basically dev ops. That's exactly okay. So dev ops. So you're probably going at node Absolutely, it would be a good one. Yes, Python may be for data management and mid wrangling Okay, cool Just wanted to kind of get the sense of that because you know some dumb we think about we talk about all the time Around dev ops. Yes, and then you got you know the older old school if you're over 40 like me close to 50 Then you know Java You're under 40. You don't do job. Well, that's what that general. That's my Young kids like a child And I'm very excited about the dev ops move I think it's evolving quite quite rapidly and it's an area where we're trying to continue to help enable the software developers To control their own infrastructure Mark really appreciate you coming I know you're tired busy all the time certainly jet lag from California We both flew out so and you guys I know how hard you guys do customer meetings Thanks for spending some time on the queue. Give us the update on the engineering front Final word. I'll give you the word quick summary of the difference between Paris show open-stack summit and North America. Was there a distinct difference between the two? So I gotta tell you I didn't actually go to Paris. I missed my first summit ever I was bummed by it. Sprite my Sprite my ankle couldn't go so I got to get back to Vancouver So, you know, okay, you can't comment. Okay, I wish I could I wish I was there We heard more telco smaller attendance great focus very focus a lot of new people There's a lot of new people in the European audience really had not People hadn't traveled over the US to attend a lot of NFD conversations and then a lot of good conversation about the Enterprise upgrade of the upgrade ability and modifiability I heard great traction and getting new people in the growing community. So awesome. Thanks for coming on the queue Mark the senior vice president of OpenStack within HP cloud Senior vice president engineering at HP healing in cloud of talking about open-stack here in the queue. I'm John Foyer with Dave Blasio We'll be right back after this short break