 The Cube at OpenStack Summit at Lata 2014 is brought to you by Brocade. Say goodbye to the status quo and hello to Brocade. And Red Hat. Here are your hosts, John Furrier and Stu Miniman. Okay, welcome back everyone. We are live in Atlanta for the OpenStack Summit. I'm John Furrier, the founder of SiliconANG. I'm John Myko, Stu Miniman. This is the Cube. This is our flagship program. We go out to the events and they strike the signal from the noise. And we want to get the action. We go wherever the action is and I want to thank Red Hat and Brocade. They're the only ones who stepped up and helped us out with sponsorship. We really appreciate that. They're the best MVPs for the week here. Red Hat and Brocade, really appreciate the support. Our next guest is a startup coming out of stealth called Stackstorm Evan Powell, Chief Executive Officer based in Palo Alto. Welcome to the Cube. Thank you very much. You're no strange at OpenStack and the systems we were just talking before we came on. Accentive, which a Cube alumni is running that. We've got Tarkin Maynard taking over your previous company. You're coming out of stealth. Congratulations. So tell the folks out there about the new startup and what you guys are doing and then we'll just jump in on the commentary on OpenStack. Okay, great. Well, as you alluded to, I've been talking about software defined storage for years. But it kind of dawned on me like, what's defining all this? What software is defining and orchestrating the whole thing? And when we looked at it and I looked at about 50 companies as EIR of Exceed last fall, what I saw was there's an opportunity with DevOps and OpenStack for a new leader in operations automation. Operations automation's been around for a long time. In fact, my co-founder helped create it as a head of engineering at Opolis. He's now Microsoft's system center orchestrator. Obviously Horowitz, Ben Horowitz did pretty well with Opsware. Then virtualization came along. You had another wave of operations automation. We think we're right at the cusp of a third wave and that's where we're focused. We also apply some learning algorithms so that the system can learn. We talk about automate the automators as well. So the system can learn over time to be a better piece of operations automation. So StackStorm has what kind of product? Software defined? So it is software, and it's operations automation software, but you can think of it as providing those definitions. What is on top of all this software defined data center? Well, it turns out in the top operators, Facebook, they talk about F-bar. They've actually written software that closes the loop between monitoring and actions. That's Facebook, right? But we're not all Facebook or another one is. Everyone wants to be Facebook, but in terms of Ops. Yes, absolutely. PayPal is another. We have an advisor who runs cloud at eBay, PayPal. They've written their own, but what about the rest of us? How can the OpenStack community start to close that loop? We helped sponsor a project with our friends at Mirantis and Intel called Mistral, which is workflows of service, but StackStorm is 100% open source. We use Mistral and other pieces to give you this holistic view into what the heck is happening and hopefully the ability to learn and collaborate over time. Yeah, Evan, I mean, when we look at OpenStack, one of the white spaces in the area is there's lots of different projects and there's lots of pieces, but I still need to help simplify that overall operations. And it sounds like you guys are looking to fill that need. Yes, that's exactly it. I mean, the good news is those who figure it out, I'm talking that PayPal's, the Facebooks, et cetera of the world are not a few times more productive. They're like 10, 100 times more productive. If you look at the ratio of orders of magnitude. Literally two orders of magnitude more productive. If you look at how fast somebody like a PayPal can get new features out today, it blows away what my friends, many Nexenta customers frankly could do in terms of change cycle and agility, completely blows it away. But for the rest of us who can't afford to hire 50 deep computer scientists who know the full stack, right? So you gotta be a full stack guy. No operations. This is DevOps, right? We call them like and know how to work with developers. We call them like the dancing uniforms, right? There's only so many of the dancing uniforms in the world. What we need to do is have software that enables the rest of us, takes our scripts, makes them more intelligent, right? Nables us to collaborate with each other and ties into your entire operating environment. Yeah, Evan, I mean, that resonates so much with, Wikibon's been talking probably for the last four years. We've talked about hyperscale and what those guys are doing and how can that bleed into the enterprise. So you're right, the hyperscale guys, they've got their group of PhDs. They tweak their own applications and that takes care of everything so that infrastructure, for most of them, can be pretty stupid, but some of them do tweak it. You look at Amazon, creates like hyper-specialization. If I go to the enterprise, I mean, they need pre-built solutions and most of them aren't gonna build all their applications and they've got all that legacy baggage. So how do you help deal with that? Well, the good news is OpenStack itself, as you know, is super open, right? And the notion of a fully API-driven cloud is revolutionary and we sort of take it for granted, some of the OpenStack stuff and I think your typical viewer might think, oh, yeah, it's neat, it's got APIs. No, no, this means that someone like us or some other orchestration product or solution does not have to write natively to all of these environments, that's a big deal. But it's also true that we have to work with whatever's there. So we're not coming in and saying, we can bring you Nirvana first to rip out your old stuff. You've gotta work with what's there and have a hybrid environment. And so we do that, we fit into whatever monitoring or orchestration solutions you already have and that is absolutely, absolutely fundamental. And you can start small. It can be what are the top five things that I have to do manually to my environment? Let us help you figure out how to automate three or four of those in a fully auditable, transparent way. And then we think over time that carrot approach will encourage your operators who probably don't want to carry the pager anyway. They want to get away from that. They want it to really automate not just the things but start to automate themselves or at least the late night troubleshooting piece of themselves. So can you explain to us a little bit more about your software? You said it to open source. That means there's a free version of it. People can try out. You've got kind of a freemium model or a kind of open source plus. We will. So as you know, we came out of stealth last week and we are in private beta today. But yes, it is 100% open source. Kind of the engine on workflow is called Mistral. Again, an open stack project. We have other components as well, including some of these integrations, including sometimes we call it the decider which we could talk about, but a piece of if-then logic, some of the auditing capabilities. Those will all be open source. Some of them already are. And yeah, the freemium model. For open stack users, they want us running on-prem. If you're more of an Amazon user, most of our users are both, but you might be comfortable with us running in the cloud on Amazon, let's say. So what's the open stack have to do? Ashley, you heard the Rackspace guy on the bottom of the foundation. The community board media. So some love tweets last, oh, went great, fantastic, we're all aligned. We're trying to get out if that's reality or not. It sounds like it's pretty true. I'm not saying that they would lie on Twitter, but certainly Optics is everything. Your previous company's doing a lot of deployments, in a sense, with Tark under Tark. Is it selling? I mean, certainly people in a room or developers are all building an open stack with the customers. So what's your take on that? Are they missing out on it? Or is it just not there yet? Are there customers buying open stack? I think, first off, it's a great question. A company like Nixenta does in a month about 100 sales, net new customers. And my understanding is that all the open stack 1.0 companies put together have done about 100 customers. So the scale that a lot of startups, at least in the storage space where I came from operating at is just massively greater than the typical distro reseller today. Having said that, open stacks everywhere, right? I mean, it really is. Certainly, if you talk to people at VMware, they view it as a real issue slash opportunity. The assumption is in the field that every significant vCloud deployment is competing with open stack. So what's the disconnect? In my opinion, the disconnect is business models. I think trying to out Red Hat, Red Hat is a recipe for capital destruction, frankly. And I'm going to be the Red Hat of open stack. Well, I think Red Hat has decided it's going to be the Red Hat of open stack and IBM's doing well, HP's doing well, but the big guys can stand behind these customers. Red Hat's doing well. I mean, we have the Red Hat Summit and they're coming out. I'm thinking some big guns. They're talking about systems guys. They have guys from deck. All their top leadership are, they're not chumps. Well, it comes to systems development. But if you talk to the top operators and we've talked to a bunch of them, about 100 of them, in fact, there was at the operator summit at PayPal, somebody asked, so which distro are you guys running? Is it X, Y, Z? Is it the Red Hat distro? And the answer, almost everyone said none of them. We're just open stacks good enough. We can get the components we want direct from the source. We don't need one of these companies to come along. And that is gotta be concerning. Our approach is, and we know a lot of investors who are still looking at the distro space, but I think the way to play it personally and the approach we're taking is let's build the world's best solution for what we do. And open stack is a huge enabler and we tie into the fact that big operators are agnostic, not just big operators, enterprises. They don't really care. I don't view, in short, open stack as a product. And trying to sell open stack as a product is a challenge. I view it as a platform and a movement and a community, V-grade community probably operating in my lifetime and in infrastructure, huge believer in it. But it is 30 something packages, as you were saying, and growing, and that's not a product. That's a bunch of different products. And if you can sign a support agreement with one of the big guys, but the smaller guys, I think, are having a challenge showing top line results. So Evan, you've got experience going up against some of the big guys and trying to help drive change in the software defined environment. Tell us what you think about the open stack community and what's the opportunity for startups? I mean, it sounds like you guys aren't only on open stack, you can also do Clouds like Amazon or the like, when is there going to be enough revenue and customer adoption that we're going to see startups just all trying to, I mean, there's already a bunch of startups trying to hit open stack. But what's the opportunity that you see there? I really do think it has got to be the best at something. And trying to be the best at we support open stack is tough, unless you already are at scale and you're supporting Linux or et cetera. That's a tough one. I think you've got to look at, you've got to do the lean startup thing. You've got to listen to the customers, really understand a particular pain point, and figure out how you're going to own that in a way that is disruptive. This is what we did with Nexenta, I think. We certainly disrupted the heck out of that industry. In this industry, this piece has already been disrupted and there's a screaming need now for somebody to come through and make it intelligent. We can't wait on our schooling or what to train millions of dancing unicorns, full stack engineers who can code and understand operations. There's never going to be enough of them. We need software to step up and do it. We're also applying some learning algorithms on top so this thing can become smart. We talk about it being the brain in between the muscles, the config management, and the nerves of monitoring. It may be a lizard brain. It's not going to start writing Shakespeare for you, but it's going to do some basic level. I think that's an example. I think storage. Obviously, there's been some exits there. That's another one. I was just chatting with a friend who's a CEO of one of the software defined networking companies. I think there's some opportunities there. If you understand particular use cases. I mean, one good one I'll throw out there is lab manager, right? If you look at, OK, Jenkins does some great stuff, but VMware made a killing early on really understanding that QA use case. Well, maybe somebody could come up with a package that really nails that for the developers out there. Something beyond Jenkins, which is great, I know CloudBees and some others are working there, but something that then encompasses some of the networking side, just brainstorming here. But there's some obvious pain points that people are writing themselves. I think companies need to stand behind those pain points as opposed to saying, hey, we've made the ocean 0.1 degrees warmer. We support, we're boiling the ocean. We're trying to make open stack easy. Well, that is a freaking huge mission. We're talking billions of dollars being invested, or at least that's what they say. I think you gotta focus and own a set of capabilities. Well, they got a marketplace that they're announcing. What do you think about the marketplace? I think it's a great idea. I think we at Nexenta and this company as well, we try to partner with integrators, right? Who is actually helping the enterprises make that transition? And they need recipes. And a marketplace can be a piece of that. Already pre-certified building blocks, Legos, if you will. I think some of the top system integrators as well are building these pre-stacked set of Legos. And again, it's not just open stack, it's you bring in Chef or something like this for configuration management. You bring in something else for security. That's what I think enterprises wanna buy. Evan, I gotta ask you the senior member of the industry. I love the quote about open stack as one of those great communities of your lifetime. Obviously, there's been other ones. Linux has been fantastic. Patchy certainly has done an amazing job with open source. There's a ton of open source stuff that's been great, great communities. But the young guns that are coming out of computer science programs or music programs or hell, I talked to a guy who's a big data on developer. He was an anthropology major. So it really doesn't matter. If you've got the chops, you can come into the industry. The development mindset's changed. You mentioned Agile earlier. DevOps is one of those things now where it's a movement. It's not necessarily a program language, it's a movement. So we were talking about systems programming. Our generation, I got a systems degree. You know things about compiler design and systems, linkers and loaders and whatnot. But now with the cloud, it's an API economy. Amazon has proven there's momentum there. Certainly economics, there's some trade-offs, but if that's the thesis, APIs are the interfaces. What does the programming models look like? What's changing for the young guns out there? Share your opinion on that. Yeah, so I think this gets back to one of the prior questions, which is the good news is that through continuous integration and development, you're having much smaller pieces of functionality spun out frequently. So how does GitHub, on day one, you join GitHub and they say welcome to the company. Today you will push code into production. Well, how do they do that? Well, part of it is they have continuous integration. They have some of the stuff, frankly, that we're doing already built around operations, automation, they have a mindset which is similar to Facebook's, right? But they also are not, when they think of an application, it's not, you know, 120,000 lines of code, right? So they are taking this incrementalist approach. Very device driver. Yes, you push some code. Today. Before lunch. Yeah, I hope you have a lot of coffee. But so they're taking this incremental approach. It's cultural too, right? It's like, you know, get your feet wet. That's how it works here. And it's wonderful. The flip side of it is now instead of having monolithic stacks from, let's say EMC, you know, a company we know well, a tremendous company, but or, you know, name brands like that, you have loosely coupled software, which is getting pushed out all the time. That's good news, we think it's tremendous, but actually then being able to orchestrate that and make it work together is actually harder. So I would argue that what is happening with Agile has happened with Agile, I should say, has really boosted productivity. But often it's like a tube of toothpaste, right? Then you're pushing some of the pain down to the operators. If you don't embrace DevOps as an approach, you will get killed. All right, so I'm a customer, right? I've had an IT department, I've laid off most of the stuff, the expensive staff, outsourced my help desk, brought in consultants, have a core staff, they're doing five different things at once. One's provisioning some storage, one's DBAing. You know, since these are jobs, they're comfortable, they got their jobs. Now they got to be a DevOps guy. Okay, see you at lunch, now push some code. I mean, this doesn't work that way. So how does a company, whether it's an enterprise or a service provider, get DevOps up and running as besides cultural? What's shifting programmatically? What are some of the tooling? And what are the key things that need to get done to make it truly DevOps, or is it just CloudOps? Is it a different term? First off, it's great, again, good question. I think it always comes down to people. Every transition I've seen, including software-defined storage, but before that, I was at the forefront of the shift to IP-based communications, mainly financials, it always comes down to people. And you could say politics too, but what skills do people have? Is the organization aligned, et cetera? So how do you get over that? You got to get some wins. And we talk a lot about that. And actually, one of the great leaders, I think, of this movement happens to be an advisor of ours, Ryan Grinard, who's now at eBay PayPal. And he talks about marketing. And it took us a second to realize he was talking about marketing within the organization. He's not giving me advice on our brand. It's, okay, you're gonna do something, really think about the end user, and really do something incremental so they start to spread the word. Another thing, again, not unique to DevOps, but worth remembering is find your skeptics, embrace the skeptics within your organization. Who is that? And it'll probably be the SVP of storage because they move at a glacial pace, but anyway. But I love my SVPs of storage, but you get the idea, conservative, what are you talking about? We don't change configs. Okay, let's get them involved. Get them in a user group internally and try to get a win for them. So there's some of those human approaches, including leadership and marketing, within the organization, start small, spread the word. Embrace your skeptics ahead of tweet that out, at the quote. I mean, that's true. I mean, you need to face, because everything's now transparent, you gotta address it right head on. Absolutely. So Evan, when you talk about the skill set, one of the big leaders out there is VMware. VMware did a good job of helping greater utilization of what they're doing, change the skill sets a little, and they've got to play in the orchestration layer with everything they're doing with their vCloud suite. Can you maybe give a little compare and contrast where you see the industry needing to go and where VMware is today? Yes, I mean, VMware is an amazing company and StackStorm wouldn't be here without them. My co-founder ran vSphere client engineering for years and helped on some of these acquisitions and so forth, Dimitri, Dimitri Zameen. So a great source of talent. They've done incredible stuff. I'm a little skeptical though that I think vCloud sales are declining. I might be wrong about that, but and the reason is when we actually, I did a blog about it last week, but when we look at the architecture, the fundamental architecture, like how our software is written, infrastructure as code is not just a marketing slogan. It has real world implications for where do your configurations live, right? They have to be totally externalizable. You have to be able to rebuild us as a management solution completely from the source. We can't go into a DevOps environment and say we help you manage your DevOps environment and not be DevOps friendly ourselves in our fundamental architecture and getting to the point that a monolithic piece of software now has all of its APIs are all fully scriptable, I haven't mentioned, but that's worth mentioning. And secondly, again, that your configurations are totally externalizable, meaning you can share them. I mean, Dimitri gave a spiel at the first time we ever really spoke publicly. We didn't talk about ourselves, but at scale Southern California Linux users earlier this year and he just has a screenshot of a question was asked by a VMware user. Hey, this is great stuff, this vCloud stuff. I need to export my configurations. How do I do that? And the answer almost verbatim was you don't. Not allowed to export. Why would you want to do that? Next question. And so VMware is a big force. They have some of the brightest people in the industry, lots of customers whom they've made happy, but it's a fundamentally different architecture if you're taking it true. That's why we think it's the third wave. They're winning at the second wave, more power to them. They have won or winning at the second wave of operations automation. The third wave, different rules, different approach. Probably not a big enough market maybe for them right now and we're not encouraging them to enter it right now. We'll take it, thank you. But there is some rewriting that has to happen or refactoring. What do they have to do to be successful in that wave three? I mean, vCloud has had some issues in adoption. What do they need to do to be successful? There's, you know, that's what I gotta kind of punt on. I have strong ideas, but, you know, they could partner with Stackstorm, that would be good. Great answer. Yeah, but no, I think there is architectural limitation. I do try to build a transparent company. So I mean, we've basically blogged about what we think they're running into and we're open about it and I wouldn't put it past them. And what's the URL for your sites and we give the people the blog. It's just stackstorm.com slash blog. Okay, yeah. Other industry ideas around this platform, what do you worry about for OpenStack? I mean, if someone's been in the industry, seen the cycles before, you know all the tactics from the strong arm vendors that might want to infiltrate and slow down momentum. What do you worry about on OpenStack? Is that one of them or are other things you worry about? What are your big things as an industry participant and a CEO of a startup where a thriving ecosystem matters? So what do you worry about on OpenStack? What are you hopeful for? I think mis-set expectations would be the key thing that I worry about. Back to embracing your skeptics, I think we need to embrace the trough of disappointment, right? Every one of these adoption cycles, I mean, Gartner is interesting. I think they're right about this curve and there's a reason why it's been broadly adopted. I think OpenStack's gotta go through that trough and we need to be careful about overselling based on promises that it is easy or simple today. I think on the flip side, we need to talk about two orders of magnitude. Why is software eating the world? I mean, Andreessen's right, right? It is eating the world. But eight years ago, it was every software development project ever is late and too expensive and it doesn't work. So what changed? Agile and then DevOps changed. Every industry in the world is getting, I'm sure there are some that aren't mining. I bet mining's being disrupted by software done right too. I don't know, thermostats evidently. I mean, basically, we religiously believe in that software eating the world philosophy with some of the research we've done. But if you look at what big data, and that's gone through the same evolution. You're seeing very vertical penetration of terms of use cases, clouds the same way, you're kind of hinting the same thing. And if you think about it, I was asked earlier in the last panel on the crowd chat was, the rack space, hey, your stock's down. The people even understand the value of what you're doing with OpenStack because a skeptic or they say, ah, what's a shit investment to do that? OpenStack's done, ah, no payout there. When downstream, the benefits are massive. Talk about that. I want to get your perspective. Well, what is the value if we can get through this new software revolution? Right. I mean, everything's disrupted. Every single piece of the value chain is disrupted. Every single piece. So a reconstruction means wealth transfer, right? It does. I mean, isn't that where we're going here? And I think in part because of the open approach, a lot of the benefits really do accrue to the users, right? And that is a big difference. And I believe in the open approach. So the chips are on the table and it's anyone's game at this point. Would you agree? I agree. I think one comment would be that there's a lot of OpenStack developers or developers of Open DevOps approaches, software, who then run everything on proprietary clouds. So they need to kind of step back. I think we as a community need to support, frankly, Rackspace. I mean, so it's a little more expensive but a lot more flexible. I would gladly, I do gladly pay that tax versus going with someone who is seen as more of a open source strip line. Yeah, and Stu kind of peppered the Rackstates guy and said, hey, did you fumble the ball? And that's the question everyone's asking. And I got to give Rackspace props. They step back at the right time for the betterment of the community and they deserve kudos for that. They don't necessarily get recognized for it because at that point in time it's very fragile. Absolutely. So Rackspace is a gender. What do they know about software? They kind of take a step back and then the party got full and they're like, hey, we're over here. So, okay. I think it'll pay off. But when you can take development of a piece of software that does all sorts of things, pick one, you know, we just saw a train go by, right? So traffic control. And you can take those cycles from years down to a day. You get smarter networks. I mean, that's one I know because of, anyway, we have an advisor who does that kind of control. I mean, everything's connected to an IP address. Everything's addressable through software. And as API economy with the cloud, then you can, in essence, build essentially a mainframe for everything. Yeah, every application could be a mainframe. Absolutely. I mean, that level of intelligence is now out there for those developers and shops that take this journey from Agile through to DevOps operations. Evan, thanks for coming on theCUBE. I'll get you the final word to the folks out there. Share with them in your opinion why, in this place in Atlanta, at this event, why this cloud stuff is so important. Why does OpenStack matter at this point in time? I think OpenStack is that infrastructure that allows us to finally fix IT. I mean, it's that fundamental. Let's get intelligence in the hands of end users massively faster, orders of magnitude faster, the right intelligence that they need to actually automate and fix parts of their lives. And so it would be wrong for me not to mention StackStorm again. So we're part of that puzzle, but OpenStack is a massive shift and it's tremendous to be here. Evan Powell, industry guru. Great to have the commentary. Congratulations on coming out of stealth. CEO of StackStorm. CloudStorm is here, the cloud collision, all the action here. This is theCUBE. We're right back with our next guest after this short break.