 Okay, we're back here live on day three of the CUBE, our flagship program. We go out to the events, extract the signal from the noise. I'm John Furrier, the founder of SiliconANG. I'm joined by co-host Jeff Frick here. Day three with Ross Turk, who's the vice president of Community at Ink Tank, which is you guys manage around CEF and all the other components around. Obviously involved in OpenStack pretty heavily. Welcome to the CUBE. Thank you. So obviously a community you're involved in dealing with all the open source activity and all the people developing. So let's share with the folks first about the company, your company, and CEF and what's going on around that and then we can dig into the OpenStack conversation. Sure, Ink Tank is the primary sponsor of the CEF project. It was spun out of a hosting company called Dreamhost last year. The CEF project has been in development since 2004 and it's an open source distributed storage platform. It started off as a Department of Energy sort of experiment or funded project to figure out how to make a distributed file system that really, really scaled. But out of all that thought came CEF, which is more than a distributed file system. It's a complete storage solution and what makes it different is how it scales. So what storage is open, I mean is it pure like commodity drives? Is it specific vendors? Because we've had NetApp on earlier, they're doing some work in this area. What's storage in particular? So CEF is, it's different from a lot of that because it's software, it's not hardware, right? So CEF is a software layer that sits on top of everyday hardware and exposes those storage resources of that hardware in a way that's useful. So it exposes it as object block and file storage depending on what you need. So it's commodity storage. It's commodity hardware. Not like a brand named like a NetApp or EMC. That's right. So they're competing with that from an open source standpoint basically. So the open compute initiative must be interested in this. Yeah, absolutely, absolutely. Yeah, so NetApp obviously was on Monday and they have their filers out there. So obviously the storage is a big part in fact one of the comments was storage is afraid of OpenStack because there's a lot of disruption going on. So what is the big thing about CEF that's getting traction in the enterprise and the OpenStack community in particular? Is it the positive compliance? Is it some of the things you guys done with it? What's the big aha around CEF? What do people need to know? I would think the big three ahas. The first one is lowest cost per gig because it uses commodity hardware and because of its self-managing and self-healing capabilities it's a little less difficult to administer than a lot of other storage systems. It was built exactly for that. So keeping the costs low on storage is super important especially as the scale is starting to get crazy, right? We're seeing this hockey stick in scale. Storage is not scaling linearly. So the technology shouldn't scale linearly either. So that's the first. The second I think is the really good level of integration that CEF has with OpenStack and CloudStack in Proxmox and Gennady and a lot of the other CloudStacks and integration with the Linux kernel and that sort of stuff. So the focus on integration is definitely number two. And then I just completely forgot what I was going to say for number three. So, well maybe we'll serve us in the next question. The slides are not rollers. No, no slides. No slides, no slides. Vendor hype-free zone here. So let's talk about OpenStack in particular. So what are you guys doing with OpenStack in the storage area? Obviously software-defined, software-led as we call software-led infrastructure is our vision. We don't really call it software-defined because it's not yet defined. It's evolving very fast. And you also have the advent of Flash which David Floyd or Wikibon covers extensively. Those things are changing the game. So scale out, open source is really the key. How do you guys work with OpenStack? Can you share some of the things there? Sure, so our integration with OpenStack is sort of in two different areas. The object and the block, right? So on the object side, we speak the Swift API, Northbound, so applications written for Swift will work with stuff in an OpenStack context. And same thing with S3. If you have an application for S3, it works too. We also integrate with Keystone over there for authentication management and that sort of stuff. On the block side, we have integrations with Cinder, Glant and Nova so that you can boot virtual machines off of images that are distributed amongst the entire storage cluster. And do interesting things also like create volumes from images instantly. So it makes it a lot easier to manage large numbers of virtual machines. What is the big interest from the enterprise? Because I talked to CIOs and they're saying, hey, we need to have these boxes checked off, POSIX compliance, NFS support, all these nuances. That open source developer is kind of not really thinking about that operational side. So what has been the traction in the enterprise? What's the interest level like? And what use cases do you guys fit? Well, the primary use case in the enterprise is private cloud for us today. And I think it's just because it lines up really nicely with CEF's architectural theories, right? The scaling out in every sense. A lot of the more traditional storage stuff is in the future for us. We're looking mostly at private cloud deployments today just because it just fits the scalability profile stuff really well. So we're targeting, if you look at the financial industry, you have data needs there that are growing just like crazy. So we try to focus on being the most cost effective and most easily managed. What use cases and who are you disrupting? Which vendors? Obviously EMC has a large presence in the enterprise. So as the guys buying EMC drives, what do they look at this? It's a lot of extra work. I got to bring in commodity storage. I got to set that up, stand that up and I got software, how do I manage it? Those are really legit questions they have. And they might not even have the people to actually do it. Sure, sure. Well, it's really important just as anything that we're seeing here at the OpenStack Summit to have a lot of the build automation and systems like Puppet and Chef and Juju, we're working on getting integration with those. So it's easier to manage these distributed systems because as you suggested, instead of having a single box with a sort of single purpose, now we're talking about hundreds or thousands of boxes. Management is a big issue, but we're relying heavily on the efforts of the folks who are working on Puppet and Chef and Juju and Robart. And Chef wants to ask a question, but I want to get this out, but you're seeing clearly that the scale out is the preferred way, right? For the environments you're in? Absolutely, absolutely. Yeah, we do best in environments that are focused on scale out. And is it really to be able to implement, excuse me, new applications? Are these new business applications that are enabled because of this massive storage? Or is it just simply a less cost alternative to do what they were doing before? Or are you enabling? And if it is new things, what are some of the business applications that people are using this new cheaper storage that they couldn't do before? Interesting. Well, again, I mean, I go back to the cloud, right? A lot of the greenfield opportunities that we have in the enterprise, at least, are our private cloud. And I think it's because people are starting to look at how they can make their resources more elastic, right? I mean, you can say a lot of things about NetApp and EMC and all of that, but the SEPH strives to be ultimately elastic. So any type of workload that needs that, HPC, big data workloads are also really well suited for SEPH. Okay, so it's really the elasticity adding and subtracting and really morphing the requirements dynamically, as opposed to kind of a static, as somebody else earlier said, and you're kind of mapping your storage to your application needs on a one-to-one basis. So a little bit of a shift gears, you know, we're here at OpenStack and it's interesting that there's all these kind of open source projects that are now coming from different parts of the infrastructure and merging. You've been involved in SEPH, how do you see kind of this confluence of open source projects now coming together to really provide an open source option for the enterprise with these, you know, SQL and Linux and SEPH, and it just seems to be going on on MongoDB. I mean, it's just curious that they've all been kind of independent projects, but now clearly to execute well in the enterprise, they all got to kind of work together. So how is that being involved in the foundations and how do these groups work together? Well, open sourcing the enterprise is nothing new, of course. I think it's just moving up the stack a little bit. In the early phase, in the early 2000s, the solution to that was a Linux distribution. Today I think the solution to that is really good definitions around interoperability, good adherence to API standards, strong systems like Chef and Puppet that can understand the interdependencies between systems and deploy them in an intelligent way. And the ephemeral state or the ephemeral nature of a lot of cloud computing, I think makes some of that easier as well. You know, things can, things like continuous integration will, they help make sure that things are integrated properly as they're being developed. So there's been a lot that's happened since Linux hit the enterprise that make it a lot easier for people to take that risk. Ross, I want to ask you about, I'll see the buzz here. So what are you hearing here at OpenStack Summit? I'll see it's a record crowd and they expect the growth to be bigger next time. It's a lot of, I'll see a lot of interest. Again, it's not a hype area right now. There's a lot of developers doing some stuff. What are the core issues you're hearing around storage here? And what are the hallway conversations and specifically things you heard in sessions? Well, I haven't had a chance to make it to many sessions because, you know, it's okay. Okay, what are you hearing at the parties and what are you hearing from your peers? I mean, storage is a hot area. It's the epicenter of all the action storage. You've got big data. You've got as a key element of the operating environment in cloud and it's not going away and there's, you know, flash memories accelerating different architectures and it's quickly disrupting the marketplace. What are you hearing people talk about? Well, I'm hearing people talk about Seth. So that's good. That's wonderful. OpenStack in general, I think we're hearing more about users, right? Then developers. And that's been the most interesting thing at this OpenStack Summit for me was hearing big companies talking about how they're going to be using OpenStack, which is something that is becoming more increasingly interesting, right? Particularly for us because we're starting to see some of our biggest users come out and talk about Seth publicly, which is a lot of fun. So what's on your roadmap right now? So what's the next year look like for Seth? What's the key things that you guys will be working on and where's the attention going to be focused in on? I think our biggest challenge, our biggest upcoming challenge is figuring out the multi-availability zone thing. Right now, Seth, the way it works, is a land scale technology because it does replication synchronously and is designed for consistency. So figuring out how to allow people to have one Seth cluster in this availability zone and another in a different availability zone and have them work together is a big challenge for us. And we've got a lot of good ideas, but that's a big challenge when we're working with the community to figure out. What about upgrades and automation and configuration management side? You mentioned Poppet and Chef, obviously those are popular tools. Are there other legacy environments that you're bumping into that you need to support? Like obviously the big enterprises and service providers that some, not all of them have a clean sheet of paper. They have to deal with their existing legacy. How do you guys integrate with that? Well, I mean, there are a couple of prototype ISCSI target implementations for Seth and the communities built things in Samba and Ganesha to allow Seth to speak SIFs and NFS for the typical scale-out NAS workloads and that sort of thing. But a lot of the people who are using Seth today are actually Greenfield. And they're like cloud providers? Cloud providers, service providers, exactly. Some people are not even cloud providers, some people are just storage as a service providers. Well, we know you're getting a lot interested in the enterprise folks we talked to and because of the compliance issues, right? That's a big thing. How's the community? I mean, talk about the community, obviously that's you're in charge of that. What are some of the dynamics going on? I mean, what are you seeing just in general, looking at over the open source horizon? Obviously they talk about Oskon is where REC space kicked off OpenStack, you have the Hadoop community, you have other different communities, Linux communities. So there's almost a blending now of the crowds in the communities. Can you share your opinion on what you see happening and help people understand how to navigate the different communities and is there any common threads together that you're seeing? Any signals? That's a lot of questions. Take whatever you want, just a jump ball. What's going on in the community? I'll start with the first. I'll start with what's going on in the Seth community. I think I've snatched this. What's going on in the Seth community? Massive growth and adoption, massive growth and participation, overwhelming growth and participation actually. We're at the point where the community management team at Seth is trying to figure out, okay, we have a lot of people who want to contribute to the project, how do we do that in an orderly fashion that maintains the quality of the project and that maintains its architectural integrity. So that's the big challenge for us. And it's a huge challenge. Governance, exactly, governance. And engineering processes and practices are a huge challenge for us. We look beyond that to the community. It's fascinating. I've been in the open source community for a really long time and I found that everybody who in the early years was saying, oh, I'm an open source person is now saying I'm a cloud person. I think that's a really interesting movement. The open source and cloud communities are really intermingling and it's funny. Back in my early career, I used to think that the job of a community manager was temporary because eventually everybody will be a community manager. Just like no company has an internet department these days. Native. Yeah, and so all kinds of companies who did not have community intelligence before have community intelligence now and that's fantastic to see. And you're seeing people really understand the maturation, maturing of these protocols, these implicit protocols, contribution. Not just a marketing statement. And that's been something that Rackspace, I think, hats off to Rackspace what they did with OpenStack. What they could have kept it kind of internal and put up red herring out there community and make it look like a community, but they went all in. And the benefit- You could call that faux open source. Yeah. What do you call it? Faux open source, like faux, like yeah. So I mean, but again, but they've proven the business model side of is actually interesting. Their lift as a company, they can move the ball down the field because they now have got huge pool of people developing on their platform or and it's helping their guys. So that's a new concept in the mainstream and not so much an open source, but the leads to the question is, what's your advice? I mean, just, you know, just on a personal level to folks in the enterprise that might be watching who say, hey, we got to invest heavily. We got to do transform our architecture to modern. We want to go cloud, but we got on premise. I got to get new tools and I want open source. We're obviously scaling out. I'm buying a lot of gear and I don't want to be restricted on data centers. I want to scale it out and use an interesting hardware. How do they deal with open source? What's your advice? Because that's a question we get all the time. So when you get that question all the time, what is the concern? Is it a legal concern? No, it's just like, what's the mechanics? I mean, do I hire different people? IT has always been kind of like this closed operational focus. Now the investment is coming in saying, hey, we want BYOD, bring your device to work. We want to have mobile apps. We want SaaS and we got to have our business model be Sassified. And so they got to do with open source. The best advice I can give to enterprises wishing to adopt open source is to actually become involved in the community. And that's something that a lot of companies are slowly figuring out is you can't just take the bits off the internet and deploy them and get them working. It's really much better if you understand how they're built and understand how they're developed and who's developing them and what their motivations are. And you insert yourself into the conversation to make sure that when you deploy, you have the intelligence, you have the lifeline if you need it as well in the community for help. And you have a way to influence the direction of the technology, all comes from participation and contribution. Well, it's interesting, just your title, you know, Vice President of Community. It's kind of funny to watch titles evolve over time, like, you know, CMOs and different things that kind of follow the trend of the technology. But I wonder if there'll be a lot more because as you said, it's, you can't pull a head fake on the community. You can't feed them a line that's not supported with other stuff. You'll quickly be found out and you'll cause more harm than good. And community is, I mean, it is becoming an executive discussion topic, you know, which is something that's new, you know, the discussion used to be about, with community used to be about, do I allow my engineers to contribute to an open source community? And now it's about, how can I align my company with the motivations of this community? Or if you're in the case of Ink Tank, you know, how can we build a broad ecosystem, you know, around this technology and still have a business, right? So these kinds of questions, I think we will see more VP of community titles because these types of questions are business questions and they're high level strategic questions. They're not as tactical as they used to be. What are some of the kind of legacy frameworks in terms of thinking that you guys have overcome as a company and maybe Ink Tank's not the right company because it's relatively new and it's kind of built out of the open source project? We got lucky because we were an open source project. Which means that a lot of the spirit and company culture were there. But we still have conversations every day about how do we differentiate our services from the other services in the ecosystem? How do we align ourselves as a business and still facilitate a community? Like what is that balance? It's a conversation we have every day and we'll have every day as long as the company is around. And you should, as soon as you stop having that conversation, then you're going to start blowing it. But again, you feel fundamentally because of where you are and what you've decided to do with your career that that's a solvable problem. You can effectively build a business around an open source foundation of technology and really do it effectively. Yeah, unquestionably. I mean, there are a lot of bad examples but there are a lot of good examples too. You know? A lot of companies have done it really, really well. And I think it's definitely possible. I actually, I think it's unavoidable. Yeah. Okay, we're here with Ross Turk inside the Cube talking about SAF and open source, all the great magic that's happening, awesomeness that's happening here at OpenStack. Final question before I want you to kind of get the last word in. I want you to, and take your time. There's no need to be a short answer. Obviously DevOps, we talked yesterday heavily. That's now mainstream. A lot of people recognize DevOps as a mindset and a movement that's legitimate. I want you to comment on your opinion of infrastructure as code. I mean, a lot of us folks talk about that in the communities because that's a great way to look at programmable infrastructure but that's now coming into the mainstream. And it's a concept that a lot of people are talking about like service mesh, et cetera. What is infrastructure as code? What does that mean to the world? Infrastructure as code. That's an interesting question for me because I've always thought about it that way. I've always been a program, I started as a programming system administrator. Yes, that is actually a really interesting question. I never saw the separation for DevOps that a lot of people saw. I think that that was somewhat manufactured by the way companies were put together. Yeah, I don't think infrastructure and development are different and I don't think they really ever have been. There it is, he's been living infrastructure as code since he's been breathing in IT. And obviously scale out open source. Ross, thanks for coming inside theCUBE and sharing. What's going on with Seth, appreciate it. This is Silicon Angles exclusive coverage of OpenStack Summit. Here we're day three live wall-to-wall blanket coverage. We write back, extracting more signal from the noise, sharing that with you guys. Thanks for watching. We write back with our next guest after the short break.