 Okay, we're back live here in Santa Clara, California. This is the Velocity Conference, O'Reilly Media's Velocity Conference. This is Silicon Angles theCUBE. theCUBE is our flagship program, where we go out to the events and extract the student from the noise. I'm John Furrier, the founder of Silicon Angle, and I'm here with James Turnbull from Puppet Labs, technology, VP? VP of Technology. VP of Technology, okay, good, get that right. You guys have grown out of the Puppet Chef early days of recipes and provisioning and whatnot, so that is great, first generation cloud, but now cloud's growing up, mainstream. I want to get your take on two things. First, what's going on with Puppet? Okay, you guys have got actual Puppet code base, and also the company. And then let's talk about what's going on, DevOps and Agile and mobile. Sure, so Puppet Labs is about eight years old, I guess, and we started off as our CEO, Luke Kenees, as sort of a sole operator, and he was largely the open-source software sales model. And we're now about 160 people, and we just closed the $30 million round, led by VMware in February. And we sort of fairly, we sort of started off in the open-source world, and we now sell an enterprise, version of the product, Puppet Enterprise. And we come to that sort of world with a sort of, I guess, a long heritage of operations people and CIS admins, and our primary customers, the people we sell the product to are largest CIS admins and heavily used in the web ops and sort of cloud world. Why VMware putting 30 million in, I see VMware's out seeing part of EMC, Federation that we've covered in the past. But they must be interested in something because they sell to a lot of enterprises. What's their interest in you guys, besides making some money on a good investment? So I think obviously they're keenly interested in making some money on a good investment. I also think that they're interested in seeming and standing near innovative companies. I think VMware is conscious of the fact that they launched the virtualization revolution. Cloud is now sort of an extension of that revolution. And VMware is evaluating where they fit into that world and the companies they sell to now, those enterprise companies that they sell retail software to, whether it be vSphere, right up to the vCloud cloud stack. I think they're trying to stake their place in that world and we're obviously a player in that space and I think they're interested in looking at how our technology can help them be more innovative and more agile. What mega trends are out there fueling the growth of puppet labs and what specific problem are you guys solving right now? So I think that the probably the two big trends that I see at the moment is virtualization and cloud. So a lot of our customers are looking at automation because they've basically reached the limit to which they can humanly scale problems and they now need to automate those problems away or they need to use automation tools as a force multiplier. And the second thing we're seeing is that the business is much more demanding of results. They want their service delivered in, instead of waiting six weeks for a server to run up their marketing application, they want it tomorrow, they want to be able to change it on demand. And as a result, we're seeing a lot of customer self-service, I have a self-service console, I'm a developer, I want to get a copy of the production stack and I want to be able to develop against it and I want to deploy it to the, no, all of that sort of stuff. And those sort of requiring those sort of more aggressive sort of time frames and service levels has meant that, you know, you can't do that with people. You can't throw people out of that problem, you need to throw tooling people and culture in that problem. Sean Douglas, who's XEMC Ventures now over at Service Mesh, shout out to Sean and Service Mesh, well, good plug. This wasn't planned, but a plug to those guys. He always uses the term with me, infrastructure as code, which speaks to the DevOps kind of mindset. You know, you got the Facebooks, you got the Googles, people have to build their own and really make it work at an ops level, a high availability at scale, but it's really programming mentality. What does that infrastructure as code mean? I mean, what is he pointing to? Obviously they do orchestration, but you're more on the end. Sure, actually, interestingly enough, the term infrastructure as code was one of my co-founders, Andrew Shaffer. So he has the one that actually coined that phrase and shout out to him. What a coincidence. Well, I think the infrastructure as code meme is really important because it's treating, instead of your infrastructure being a special sort of snowflake, you treat your infrastructure like you treat your code. You know, you are, you know, you build it, you deploy it, you know, it has a life cycle. It's no longer a sort of closed system. It's now closely tied to the applications you want on top of it. And it's also keenly sort of, it's become keenly apparent that if you want to be flexible and scalable, that you need to be able to take that infrastructure, run it up in a new data center, run it up in a new cloud, take that workload and replicate it. And you can only do that really if you have a way of abstracting the sort of complexities of your infrastructure way the same way you abstract some of the data. And that really is, that really points to the DevOps perspective because you think about it, I was talking to one IT developer and he's like, you know, look, I got code and local host and I'm ready to push it. I'm waiting. I'm waiting for some servers to be installed. That's the old days. Bear metal stacking up, rack and stack. And people don't want to install Linux again. I mean, although Linux has been very popular and successful, but that's going away. The new schools like, hey, I want Elastic, I want resources, I want to be able to move my service levels around. I want to have service levels that are highly reliable, high quality. But I don't want to have to be provisioning gear and reinstalling patches, waiting for a security update. That developer doesn't care, you're right. That developer doesn't care about the stack. And the business doesn't care about the stack. All I care about is the application that's delivered to the time frame they want it to. And it operates in a way that with the right sort of performance characteristics that the customers are happy. Otherwise, they don't care what version of Linux you deploy or what version the package is or whether the security thing has happened. All they want to know is the results. Well, since you brought up the stack thing because this has come up multiple times here, Velocity. Share with the folks out there, what is Velocity? Because Velocity is a very cutting edge conference. I believe it's a leading indicator of where the development community is going with DevOps in particular. I'm going to say DevOps. It's probably the most generic categorical thing I'm pointing at right now. But it's not just DevOps. It's not just guys with infrastructure. It's developers, guys on the developer edge, programming codes that were this JavaScript, Python, whatever it is, they're programming it. And they need to be more agile on the infrastructure side. What is this conference? Why is this conference so popular for the AlphaGeek community and not yet in kind of the mainstream eyes? Because, I mean, it's out there, it's mainstream. I don't say it's getting there, but it's not like a cloud show. You can't say it's a cloud show. You can't say it's just a front-end show either. So it's kind of a mix. I think when Velocity started and addressed the need where the fact that there was no real place for the web performance and web ops people to articulate their concerns, to talk about the things they wanted to talk about, about advancing the technology in the industry. I actually think it's changing. This year, in particular, I've noticed that there's a lot more sort of, I guess you call them mainstream IT people here. And that a lot of us who are in the sort of, that sort of web ops community are very much seeing the, very much seeing that this is almost a social interaction for us. It's a catch up with the tribe sort of experience. And that the people joining the conference now are actually, so previously it's a, one of the cool guys at Facebook and Google and Etsy and a lot of things doing. But now we see people from State Farm and Bank of America and organizations you wouldn't traditionally see in this world who are sort of starting to adopt some of the practices and principles. I guess one of my Riley friends calls it, calls it not just the cool kids anymore. The smart kids. And I think that's what's happening. And again, this comes back down to the trends that we were pointing out. I brought up the VMWare mainly to kind of tease that out. I mean, why would VMWare be investing in puppet labs? Well, because you guys are doing some really good work around you really start with provisioning and now you're in essentially cloud automation. There's a lot of that stuff going on, dev ops, cloud ops, whatever you want to call it. It's data center ops now. So where this is going is going to the data center, right? So clearly now a whole new ball game emerges. The State Farms are here. They want to invest in robust infrastructure. They're starting to reboot their IT enterprise infrastructure. They want to be hyperscale and be small scale at the same time. So that's a challenge. What do you see for those guys out there? What are they learning? Where are they starting? What's the sentiment? What's their environments look like? And what are their challenges? So I think there's two things happening though. The first one is that a lot of those organizations are discovering that things they thought they were risk-averse about are not actually true. They're like, we can continue to deploy things. We don't have to wait six weeks to deploy an application that they're reevaluating their risk appetite. And the second aspect of it is that they're starting to look at things like cloud and virtualization and they've discovered that the biggest challenge is not tooling and it's not technology, it's not infrastructure, it's how you change the culture of those organizations. So they're inherently siloed, conservative organizations where change is a threat as opposed to an opportunity. Unlike the web ops world where if you don't deploy every day multiple times a day sometimes, then you don't keep up with your competitors. You can't scale out. Pinterest is not going to go to a median users in whatever three months or whatever it was if you don't deploy on a fairly frequent basis. And they're starting to understand that they need to change the way their organizations do IT business. What about the big guys like IBM, HP, they're trying to get back in the cloud. They're going to be fully tooled up, where they need to outsource, where they need to be partnering with. What's your take just as a participant in the community? I mean, because they have a lot of inertia. They're like the big aircraft carrier. It doesn't really turn on and dine, but when it moves, it moves pretty, pretty big and has a big wake to it. So my immediate experience for Puppet Labs was working in a very large bank. And so I've been a Tivoli and a BMC customer in the past. Those tools are, I've always described Tivoli as doing lots of a huge suite of things, not very well. And I think BMC falls into a similar sort of category. I think they're starting to understand that they need to actually change the way their tooling works. You can no longer have, particularly when you look at open source tools, where the development cycles are quite fast, where the sprints are happening with new features appear. You can no longer survey your customers once a year and then go, you'll get that one of those features may be in Q2 of next year. I think they're having to build smaller tool sets, integrate more with other vendors and deploy solutions that are no longer monolithic, but in fact, combinations of best of breed tools. So if I'm an IBM or a BMC and I have conversations with those guys, they are reaching out to organizations like us to say, how do we integrate? How do we leverage the sort of, the agility you have in the market? Well, agile is a big theme. I mean, I've been, I love agile, but you know, agile doesn't make for bad coding. And we always debate about QA and the old days of coding with QA and test and dev. You still got to do the, you still got to do the work. Interready doesn't mean put out crappy code. You still got to have some strategies and architecture to it. So I got to ask you, with the multi-cloud vision that most big companies are having, they want to use multiple clouds, like they use multiple vendors. Like I have, you know, EMC drives, but I might have NetApp Filers. I might have, you know, a little bit of HP gear here and IBM over there, or IBM services managing it. So, you know, these are multi-vendor, large-scale data centers. And not everyone has the luxury of building their own data center full turnkey. So those guys have to deal with that. So how does a company deal? And you guys have a different view because you guys are providing those key software components, helping them manage these clouds. What is that reality in multi-cloud? Is it real? Do you see that being a real factor? And if so, what things should be people be thinking about as they step back and start architecting out their solutions? So I'm starting, I think that there's some companies that are leading indicators that are looking at things like, they have multi-cloud solutions, they're deploying workloads based on cost. So for example, you know, it's 4 p.m. on the east coast. That's a peak time for us. I'll buy some spot instances from the USCs, then I'll shut those down and when the time zone changes in the west, I'll bring those up again. Or I'll move them between local cloud and then burst out to cloud service like AWS. So I think there's a bunch of those companies doing that. I think they're leading indicators and I think the rest of the, when things like the PCI DSS standard changes and you can actually do things, some of that transactions in the cloud and things like that, that we'll see a lot more companies follow that. I think the key changes they need to make around their architecture is they can no longer, they probably have to take a bit more risk when it comes to where they manage the data and how they manage the data. And they also probably need to build things that are far more sort of service oriented architectures where there are small components that it's easy to deploy, where the state is maintained in a central location as opposed to the sort of very traditional monolithic enterprise sort of SAP-like applications which you literally are not agile. You can't move them around. You can't deploy them from a, you can't, you know, doing your disaster recovery with SAP is like a multi-month sort of testing exercise. You can't flip between data centers the same way a web application can. And the workload movement too. You have workload management too. It's another force that's driving it. All right, so I got to ask you, Mousa, you guys are, you know, like I said, very respected company. You know, it started as a startup again, 30 million, still growing. Still kind of in that, I wouldn't say startup phase, but you guys are still early. You're on the front edge of all the action. What are you looking at as VP of technology within Cuppit and beyond that you're watching in the marketplace? You look at that saying, hey, we're watching those forces and we pay attention to that. What are the important things that you guys are watching? I think we're seeing a move up the stack in terms of tooling. So I think we're seeing a lot of people perceive in the web ops community that the configuration management problem is kind of solved. I don't think the future is evenly distributed in that sense, like the state farms of the world are coming to this later than other people. But in terms of the innovation, I think the innovation is moving into a couple of new areas. One, I'm keenly interested in monitoring. What's happening, like traditional monitoring has been things like Nagios and it's been around a long time. And the fact that it stayed around for so long suggests to me that it's ripe for disruption. And there's a few companies Boundry and Datadog and a bunch of other people sort of doing some stuff. Datadog will be in the queue later tomorrow. And there's a few interesting tools like SENSU and stuff like that that are coming out of them. So that's an area that I think is ripe for disruption. The other one is that people want to move up the stack and become sort of application deployment, application lifecycle management. That stuff has become, most of the tools that are out there are not very agile and they're not very good at doing that, the same fast moving sprint sort of stuff and certainly the deploying the applications. So I think we're seeing quite a lot of movement, CA's purchase of NolioSoft, IBM's purchase of Urban Code. There's a bunch of interesting acquisitions happening there where it seems to be that that's the next problem that needs to get solved. We've been provisioning the systems fast, we can configure the systems right, but we now need to get our speed to market for application right. Great, so I got to ask you about preferences. I don't want to say one's a Democrat or the Republicans or religious, and check it's all about the religious wars. This stack versus that stack. We've all been on the old listservs and the old forums when people go bloody with these arguments. It's kind of really philosophy. So Shaft and Puppet has always been kind of a personal choice. How is that going between you guys and, say, Shaft? Is it still the same kind of thing? Is it more of a feel issue? Is there any advancements on what you guys are doing over Shaft and other things? I think we have a slightly different approach. I could go into a lot of technical, philosophical details, but I don't think it's necessary. It becomes a bit of a Vim and Emacs argument, and I'm kind of agnostic about this stuff, but I think we have slightly different customers, and I think we have slightly different market. I rarely run up to them, run against them in accounts. We're heavily used in on-premise enterprise. They probably have a little bit of an edge on us in the sort of the cloud space in terms of sort of marketing and lots of the stuff. But as a tool, we're very good at handling the on-premise and very good at allowing customers to sort of burst out to the cloud or to move their workloads around. And that's a traditional enterprise. And that's a traditional enterprise. And also we have a very low barrier to entry. Like I can teach a system in how to use Puppet in like half an hour. Like I build a basic stack in like half an hour, it's worth the work. Versus years ago, yeah. It's a Ruby-based product. You need to have a little bit more programming skills. I think the barrier to entry is a little bit higher. And our tool is less threatening to a lot of traditional enterprise system that don't want to become programs. I mean adoption is really a key thing about the human. We just talked about human labor factor resources, you know, around deployment. Look at MongoDB, right? I mean, you wouldn't say how fast that thing grew because developers love Mongo, it's easy to work with. So getting things going, that's a human factor. Would you agree? That's a kind of a big part of it. Okay, James Terrible, thanks for coming on theCUBE. We really appreciate Puppet Labs, company to watch, funded by True Ventures, just recently scored a $30 million investment from VMware. Again, on-premise, we're seeing with the Flash technology, on-premise clouds are real, private clouds also known as, and hybrid clouds is kind of the delivery to the public. So I think I'm going to see a lot of different traversals of cloud deployment, certainly from the enterprises and people are looking at continuously deploying. So I think you got a good vision there. This is theCUBE, we are live at the Velocity Conference. This is SiliconANGLES theCUBE. We'll be right back with our next guest after this short break.