OpenStack Podcast #25: Ryan Floyd
I hate it when I have to miss a podcast. And I don’t miss many. But this week I was in New York for customer meetings so I had to miss OpenStack Podcast #25, featuring Ryan Floyd of Storm Ventures!
Such a shame. It was a good interview too. For those of you who don’t know, Ryan was one of the early investors in OpenStack vendors Metacloud (acquired by Cisco in late 2014) and SwiftStack, as well as other well-known technology companies like Marketo, MobileIron, and Airespace.
He’s in the heart of the action in Silicon Valley, helping his funded startups succeed, and listening to pitches from entrepreneurs about the next round of exciting new technology-based businesses. So he’s got a unique perspective and he shared it with our audience today. Subjects he covered include:
- How Amazon has leveled the playing field for startups
- The future of public and private cloud
- The importance of infrastructure companies supporting OpenStack
- Where the opportunities are in the OpenStack space
- The OpenStack business he’d start today (if he were in the business of starting companies)
- What makes a promising startup
- The importance of complexity when it comes to startup success
For a full transcript of the interview, click read more below.
Jeff Dickey: All right. Good morning, everyone. I’m Jeff Dickey from Redapt. Really excited today to have Ryan Floyd on the show. He’s going to have, I think, a great take on what we’re doing in the community, and I’m just really excited to have him on the show. By the way, Niki’s not here today, so we’re going to miss out on all the good questions. I’ll try to cover for her. Ryan, can you do an intro?
Ryan Floyd: Yeah. Jeff, you’ll do great. Yeah. My name’s Ryan Floyd. I’m a managing director at Storm Ventures. Been a venture investor for 15 years. All Storm does is early-stage enterprise tech investing. Today, what that basically consists of are things like SaaS, mobile investments, and the reason I’m here talking is because it’s a lot of cloud infrastructure. We’ve been investing in infrastructure for some time. It’s been interesting to see how things have evolved. That’s what’s brought me to OpenStack.
Jeff Dickey: That’s great. Get me some background. Tell me about how you got into technology and ultimately how you got in to looking at these OpenStack technologies.
Ryan Floyd: Yeah, it’s probably a different story than you hear from most getting into technology. My first job coming out of school … I was undergrad at Stanford … was working for a private equity firm out here, where I was basically cold calling. I was what a lot of companies call today sales development reps, SDRs. I was basically an SDR for this private equity firm. My job was to call companies, generally companies that didn’t really need money, and convince them to take some private equity, for a lot of very good reasons. Basically, it was a sales job. It was just an incredible first job out of school, but very humbling. Got hung up on lots and lots of times.
When I came out of school, I was under the false impression that there were a lot of opportunities in environmental technology. That’s actually what I had studied in school. I was actually a conservation biology major, called Earth Systems, was the group at Stanford. When I got to Summit, I quickly realized that that was not the case, and either I figured out technology very quickly or I was going to get fired. I quickly figured out technology and started focusing there.
That led me to get involved with an investment that Summit ultimately made in a company called E-TEK Dynamics, which we made that investment in ’97. It was a fiber optics business. This was during the bandwidth boom, when everybody was building out optical networks. E-TEK made some components that fit into the network that helped to split and condition light. We took that company public, and then ultimately it was bought by JDS Uniphase, which is still in existence. We sold it in 2000. In 2000, that’s when I left to start Storm. I really cut my teeth at Summit, and then really got an operational background at E-TEK. Then I’ve been a tech investor ever since 2000, when we started Storm.
Jeff Dickey: That’s great. What are some of the investments that we would know?
Ryan Floyd: Some public companies that you would know currently today, there’s a company called Marketo, which is a marketing automation SaaS company. There’s a public company called MobileIron that does mobile device management. It helps secure all the mobile devices, the iPads, the iPhones, and Android devices inside of an enterprise. We had a company called Airespace that was a wireless LAN company. It was one of the first that delivered enterprise wireless LAN. That was bought by Cisco. Then specifically in the OpenStack community, people are probably familiar with Metacloud, which was a company that I seed funded and backed Sean Lynch and Steve Curry, back in, I guess it was 2011, summer of 2011, when we got that business off the ground. Yeah. Many more, but a number of companies.
Jeff Dickey: Yeah, that’s great. SwiftStack, too, right?
Ryan Floyd: Yeah. Then currently, yeah. In terms of the ones that haven’t exited. Yeah, we’re in SwiftStack. We’re also small investor in a company called Nubeliu, which is the folks … It’s a really cool company. It’s the folks that built, that stood up the OpenStack cloud from MercadoLibre. They’re basically building out an OpenStack presence in Latin America.
Jeff Dickey: That’s great. That’s a big market.
Ryan Floyd: Yeah, it should be. Yeah.
Jeff Dickey: Great to have you on the show because of, again, the perspective that you have coming from the VC space. My inbox is flooded with folks that are in the community that want me to look at their technology and vet it out. You must see a tremendous amount of folks. What are you looking for right now in the space?
Ryan Floyd: Yeah. I do. It’s one of the things I love most about the job, honestly, is that I get to meet so many smart, smart people. I remember actually, going back to Summit, I remember one of their interview questions where they said, “How would you characterize yourself in front of the class? Where would you fit in?” I would characterize as I’m the guy that always works really hard but is never the smartest guy in the room. It’s great being in a position where I’m constantly inundated with really smart entrepreneurs that come into Storm and talk to me about what they’re trying to do and what they’re trying to build. It usually arises out of some problem or passion that they’ve had, or something they’ve been thinking about for years. It’s amazing.
I think especially today, it’s very busy, because the markets are pretty active right now. I think companies are buying a lot of technology or willing to buy from startups. Just to drill in on that for a second, it’s amazing today to think about the fact that, as a startup, you can go in, just take SwiftStack, and you can sell storage into a Fortune 500 company. I don’t know if it’s true of a lot of listeners, how much history they have, but that would have been impossible … Five years ago, maybe. Certainly 10 years ago, that would have been impossible to go do.
It used to be the case that the sales rep from the big storage vendors, whether it be Hitachi or IBM or EMC, would go play golf with the CIO. They’d figure out what their storage needs are, and they’d close it. Startups didn’t have a shot at selling anything into these enterprises. What we’ve seen over the last couple of years, or certainly the last five years, but it’s accelerated is just the ability to go in and sell if you’ve got a solution. We can talk a little bit about open source as a piece of that. We see a ton of just great, great businesses.
I think today, too, the other thing I’d say that characterizes what we see a lot coming through the door today, it’s all software-based. It’s just magical from a startup standpoint, because one of the things that always puts startups at a disadvantage, certainly on the infrastructure side, was this requirement that they had to build out hardware. That’s not true in all cases. There are some companies like Nutanixes that are doing incredibly well, Pures that are building out with hardware. But the vast majority of infrastructure-focused startups today are basically software companies.
What’s magical about that is it just doesn’t require as much capital to go build these businesses out. It really gives these startups an ability to go compete. Again, pointing back to SwiftStack or Metacloud specifically in the OpenStack environment, both these companies are software businesses. It’s just great, because it allows you to scale very fast and allows you to do a lot with a smaller amount of capital. It’s just fantastic. It just wasn’t true 10 years ago.
Jeff Dickey: Right. The SwiftStack thing in the enterprise is amazing right now. I think we’ve displaced some big iron with SwiftStack in these Fortune companies. There was no way … What are you seeing, the change in the enterprise, what do you think’s changing the buying process or the fact that they are considering these startups with new technology?
Ryan Floyd: I think part of it’s software. I think once you make that leap, that it’s running on commodity hardware, you can just get a little bit more comfortable buying from a startup, because you can replace that software. I’m sympathetic with the situation where they were making these big system purchases. They were not making a decision for today. They were making a decision they had to live with for a decade. That’s maybe not so true today, and so I think that gives some flexibility to startups and what they can do.
I think by far, the biggest push today is, and the biggest change agent, is what all these enterprises have seen happen with the public cloud. I think by far, if we did not have Amazon push the envelope on what is possible in the public cloud, we would not be where we are today. I think there’s a lot of trends that were pointing in this direction, but they have just pushed so hard. That’s put a tremendous amount of pressure on, I think, enterprises that do think differently. You no longer can go out and just buy a standard stack from, pick your Fortune 500 infrastructure vendors, whether it’s EMC, it’s Cisco, it’s VMware, whatever, and price that competitively to compete with Amazon.
In enterprises, for CIOs, for IT folks that are looking to do things differently, not only from … Price is part of it. I think there’s a little bit of confusion about that, too, that’s worth mentioning. Price is a component. People want to do things, obviously, for less cost, but I actually don’t even think that’s the big driver. The reason people love Amazon is it’s agile. The service delivery is so easy. Even if you have a robust stack from VMware and EMC, when you talk to most enterprises, it still takes them a couple of weeks to provision something. There’s still a somebody else in the middle.
I think all of these forces have pushed enterprises to be looking at startups more and more, because they just don’t have a choice. If they don’t do it, they’re going to lose their internal IT groups to public clouds. I just think that it’s both these forces, them being able to do things different because they’re not having to take as much risk because it’s software defined, but also this external push that’s coming from the public cloud that’s really forcing people to do things differently.
Jeff Dickey: I like that. I fully agree. What’s your vision or what’s your thoughts around public cloud? People are saying that everything’s going to be public cloud in a few years. There’s folks saying no, it’s always going to be the data, and the data’s going to be … compliance and whatnot around that. Where do you think this is heading?
Ryan Floyd: It’s going to be big. It’s just going to continue to grow. I think Stu mentioned it last week, too. It’s not a winner-take-all market. I think the public cloud’s just going to continue to grow, along with the private cloud. I actually don’t buy the whole data governance security questions. They are true, certainly, for some companies. I won’t deny that. Financial services, for sure, the government.
That’s not to minimize those issues, but I don’t think those are the big drivers. I think often the security concerns are raised as political air cover to not go do something in the cloud. Look, Salesforce had these … I think that’s the example I always point to. Salesforce had all these objections with SaaS a decade ago. There were very large companies that just didn’t believe that Salesforce could firewall things and could make it secure.
Today, very few people question any of that. In fact, the argument really could be made, it’s probably more secure than what you’d be running in your own data center. People hear about these public breaches, but it’s because they’re public. You don’t hear about a lot of private data breaches. You don’t hear about them, because unless they’ve compromised credit card numbers, companies are not publicizing any of that stuff. There’s not a lot of transparency there, so it’s hard to compare.
Back to your question, specifically on public cloud, I think it’s going to be large. I think the probably one thing that, maybe two things, that I think limit the growth there is, number one, data. I am not a believer that people are going to put massive amounts of data in a public cloud, because for cost reasons, I just struggle to really understand why anybody would do that. You really can do that much more cost effectively if you’re running that in a hosted center and you’ve got big data needs.
I’m not sure who [inaudible 00:13:41] quote. It’s not original for me, but it’s like data has, it’s got mass or it’s got gravity, and it’s like water. Everybody likes to compare compute to electricity and things move around. Data is more like water. It’s got mass to it. You can’t really move it around very easily. The public cloud it’s tough, because if I move my data there, it’s hard to know, if I have some other application I want run it against, some big Hadoop cluster, then I’m just locked in.
I also think the public cloud’s going to be limited on the higher-level services that people are going to buy into. I think people are very focused, at least the ones I talk to, very focused on being agile in terms of how they’re building out their applications and thinking about the future. Very few want to make a trade-off for simplicity today that may translate into lock-in tomorrow. It’s almost like people have just been burned by the big enterprise software vendors over the years, and no one wants to relive that story. It’s just that bridge has been burned to the ground. People are just very sensitive to that.
Look, it just makes so much sense for our startups to be starting in Amazon. Just compare that to OpenStack. Why would anybody go stand up an OpenStack cloud if you’ve got 15 engineers and a million dollars in revenue? It just wouldn’t make any sense. It’d just be ludicrous. At some point, if you’re spending a million dollars a month in a public cloud, in Azure or Google, or whatever, then maybe things start to change. The economics get to be a little bit different.
It’s going to be enormous. It’s so many applications. The scaling application, that’s not a use case. If you’ve got to spin up thousands of nodes to run jobs, to have that capacity that you can rent on demand, boy, it’s just amazing. It’s hard to imagine you’d ever want to build that sort of application yourself. Lots of just phenomenal use cases.
I guess I’ll end with, no big surprise, I am a believer in private clouds. I have spent a tremendous amount of time … and we can get into that here in a second. I’ve spent a tremendous amount of time thinking about it and talking to customers. I believe, as much as I believe anything in my 15 years of venture, that it’s going to be huge. I think both are going to coexist.
Jeff Dickey: What’s exciting you right now about the whole OpenStack ecosystem?
Ryan Floyd: Contrary to popular belief, I think it’s maturity. I think while OpenStack is still relatively difficult to use, it’s not simple, I would argue today it’s mature. I would argue today you can put it into production, for people that know what they’re doing, and they can get it to work. We couldn’t say that really easily two years ago. I remember when … I think it was in San Diego in 2012, when MercadoLibre guys got up. It was going to be this great panel session on a production use case of OpenStack. I don’t know if you were there or not.
Jeff Dickey: No.
Ryan Floyd: They proceeded to spend the next 45 minutes explaining all the hacks that they had made to whatever release it was at the time to get it to work. I was just like, “Ah.” It’s like, “That’s not production.” Most companies can’t do that. You don’t hear that today. Even a year ago, I know intimately from my time with the folks at Metacloud, they’re rock-star operators. They were definitely making some changes, but they were working really hard to just use the existing code base, because they wanted to make it compatible. That’s got me excited, because I think that maturity is really critical to adoption.
The other thing that’s got me excited from a startup standpoint is that you’ve got a lot of big companies that have adopted it and are really pushing for it, because I think it takes a village. It was never going to be Rackspace and NASA that were going to make this successful. You really need lots of companies all working together with NASA, with Rackspace, to push it forward.
The fact now that you essentially have every single infrastructure company behind OpenStack … Really, the only exception I can think of is really Citrix. Everybody else, and even Citrix “kind of,” is behind it completely. Obviously, they’ve got a lot of reasons to do that. It’s the biggest existential threat they face to their business. There’s a lot of reasons they’re doing it.
I think that’s overall, it’s really good, because I think it’s going to really push the whole community forward. There’s going to be bumps along the road, and it’s not going to be easy. There’ll be some issues at the board level. Overall, if you look over time, I think it’s going to be good, because it’s really critical for their businesses. I think a lot of people think, “The board, the dynamics, it’s going to screw it all up.” I’m not privy to any of that. What I do know is their success of their business long term depends on making OpenStack successful. I just don’t see an alternative for them. There’s a few other long-term priorities that are more important. That gets me excited, because it means that for startups drafting behind that, being in that mix, it’s great overall.
Jeff Dickey: How are the big players, the Oracles and VMwares, what are they going to do as far as making money? Are they going to have to fork the code? Are they going to start doing more proprietary things? Are they going to be all in and figure out new business models?
Ryan Floyd: No. I hope they don’t fork the code. I guess they could. I don’t think that’d be the smartest way to go make money. No, I think they’ll make money on delivering a solution and simplicity into the enterprise. We were talking about it before the show, that it’s not easy to deploy OpenStack. I don’t think that should be confused with whether or not it’s production ready, because I think they’re two different things.
I think that’s there’s a lot of opportunity for these big companies to go in and help large enterprises who don’t want to be in the business of delivering against the enterprise feature set, which we can talk about in a second, for them to build value on top of the open source core, because everybody wants to be able to have some kind of compatibility going forward. No one wants to be locked in. I think the fork, anything that creates any sort of challenges in terms of being able to move going forward is going to be hard. I think actually one of the drivers around that will just be hybrid cloud.
I really hope in … maybe it’s not a year, maybe it’s a couple years out, but there is actually API compatibility with Amazon, full up. If you talk to the folks at SwiftStack, they will tell you that they ship with APIs that are compatible. They do technically, but it’s not a … You can’t just point your application at Swift and then point your application at S3. It’s not that simple.
I was meeting with a senior engineer for NetApp last week, and he described it as “seams.” Instead of being seamless, it’s got some seams that you have to address and deal with. I think probably in a couple years, there won’t be any seams. It’ll be seamless, and you can be able to push. That will, I think, force all the vendors to really think that the value proposition’s going to have to sit on top. Maybe just to hit on that in terms of thinking about that.
I think, contrary to popular belief, enterprises are willing to pay a lot of money to solve their problems. They have money. If you’re providing a value proposition to them, companies are very much willing to part with cash to solve their problems. A lot of the enterprise feature set, like management or security or compliance, I don’t think the open source community is going to spend a whole lot of time focused on it, because it really doesn’t deliver value with respect to the open source core. It’s really about enterprise features that are very specific to an enterprise use case.
Just take compliance. The value to probably 80 percent of the developers in OpenStack is zero, because compliance comes with a tax. Nobody really wants that to be part of the core, because it just makes everything more complicated. Those are great things for the bigger companies to deliver against and provide value, and get paid for it.
Anyway, I think there’s a ton of opportunity for all the big vendors going forward, because of the complexity. Compare it to something like Linux. OpenStack is, I don’t know, 100X, 1,000X more complicated. It’s a business. You can even compare it to Hadoop. I don’t want to, I don’t know, make too many comparisons about how much more complicated OpenStack is. Even if you say it’s comparable to Hadoop, it’s just it’s complicated. It requires, I think, companies to stand behind it and really help enterprises get set up. I don’t think that’s going to change. It’s never going to be out of the box, push a button, and here you go. I’d be surprised.
Jeff Dickey: Yeah. What do you think is missing from OpenStack or the community? What are some of the gaps that people can address?
Ryan Floyd: I’ve told a lot of entrepreneurs that if I was starting an OpenStack business today, I would start a services business. I think that that is the biggest gap today, bar none. You talk to any enterprise, and they will tell you they have X hundred VMware-certified engineers. Ask them how many are OpenStack certified. Oh, whoops. We don’t even have OpenStack certification. How many even know what OpenStack is? Okay. How many have ever stood up a cloud? There’s just not that many. Part of it’s because it’s new. Look, OpenStack is early, and it’s just getting going. You need those people to really help gets things stood up. There’s just a huge lack of talent out there right now focused on that.
I think that that’s a huge opportunity. It may not be a venture opportunity, because it’s a people business. It’s maybe a little harder to scale. I think it’s just a huge opportunity. You saw eNovance got taken out by Red Hat. Mirantis still does some services, but they’ve worked very hard to shift towards product. You think about it. Who would you call today to do service work? You could call, Redapt does … You guys do a little bit of work. Maybe Blue Box does, but Blue Box [inaudible 00:25:09]. There’s not as many. I think it’s a great opportunity. I think that’s one.
I think the second opportunity is really … I alluded to it a second ago, talking about what the enterprise feature sets are that these enterprises are looking for that they don’t have today, things like security and compliance, which are really an afterthought in the open source community, because they’re a tax, they’re overhead. No one’s really thinking about those issues. They’re really not that cool, anyway. They don’t really push the platform forward. They’re this, I don’t know, baggage that it’s got to carry with it. I think those are good areas to focus on, because every enterprise needs to have it.
I think looking out a little bit further … I’ve seen a number of these companies come in and present to me. I think it’s a little bit early, but I think the hybrid cloud story about how do I provide a view for my developers to basically simply get access to an internal OpenStack private cloud instance or get access to an AWS or Azure or Google compute instance, and just make that seamless and easy, and be able to apply some policy across it, have some transparency against it. It seems like that would be simple to do, but there’s nothing out there today that really makes it easy. I think that’s a big opportunity. If, like me, you believe in the public cloud, people are going to be using both. You got to make something that works seamlessly across both.
Jeff Dickey: Yeah. I don’t think people realize, too, when they’re building these OpenStack clouds that a lot of them are greenfield. It takes almost 20 vendors to make it work. There needs to be some sort of consolidation, because everyone’s good at this one piece, and you need all of it. How do you manage your logs? How do you do reporting and monitoring? There’s all these different pieces that go into it. It would be. It’d be nice to see it under a single pane.
Ryan Floyd: Yeah. Yeah, and a single vendor … That’s I think where the opportunity is for the big guys. If I was Marten Mickos at HP, that’s exactly what I would be building out. Those components that just make it simple to deploy and provide all that functionality under one solution, one throat to choke, when things don’t go right, because customers like that. They don’t want to have a lot of finger pointing in terms of, “His log analysis didn’t work. Her API wasn’t configured right. His whatever … ” Customers don’t want to hear that. They want you to solve their problem. They don’t want you to create more.
It’s interesting you bring up the greenfield opportunity. That’s something else that I think a lot of entrepreneurs don’t always grasp when they come in and we talk through it. Maybe it’s because I’ve lived it. I’ve had the unfortunate experience of having gone through this. Things happen really slowly in infrastructure, and things are at greenfield, nine times out of 10. It is the rare situation where someone is ripping out existing infrastructure to put in an OpenStack cloud. It just doesn’t happen that often, because if you think about it, we’ve all got 10 things to do in a given day and only time to do three. That application that’s running, that is humming along, there’s nothing wrong with it, I’m not going to touch it, because there’s no reason to. It’s the reason why mainframes are still out there today. It is typically greenfield stuff.
I think one thing people have concluded, “Oh, OpenStack adoption. Where’s all the production use cases?” The reality, it just takes some time. What I’ve actually been encouraging a lot of people to do with OpenStack is, yes, production is great. That’s awesome. Let’s talk about production apps. Also, build out your test and dev cloud. Do that, because test and dev to roll into production, that’s something that’s going to be 2X, 3X the size of production infrastructure. That’s the vast majority of what Amazon’s infrastructure does today, is test and dev. Anyway, I think it’s important. I think as people think about OpenStack, it just takes time. It’s all heading in that direction. With every fiber of my being, I feel that way. It just is going to take a little bit of time to get there. Doesn’t happen overnight.
Jeff Dickey: I agree with the test dev. I’m out there every day. My biggest song and dance is, “Give your developers all the toys that they need to develop modern applications. They should have OpenStack. They should have Kubernetes and Mesosphere and everything, everything at their disposal, and start building what they want. Give them all those tools.” Yeah.
Ryan Floyd: Yeah. Yeah.
Jeff Dickey: That will help adoption. You’re seeing a lot of folks pitch you on their infrastructure stuff. What do you look for? We have a lot of folks that are doing startups around OpenStack that are listeners. I know it’s a balance, but is it about the actual team? Is it about the idea, the products, or is it about how much adoption they have? Is it about their pitch? What is that magic balance?
Ryan Floyd: Yes. Gosh, it is a hard balance. It unfortunately really is an “it depends” question. I would say, specifically around infrastructure and OpenStack, I think, stating somewhat the obvious, I think the team really does have to have technical chops. I think you have to be credible going up in front of a technical VP at a Fortune 500 company and really being able to explain what you’re doing and why, and doing it not only in very technical terms but having a grasp for the business reasons, and really get into the mind of why those customers are doing what they’re doing. That takes, I think, a special person to be able to do both. It’s not sales. It’s not just engineering. It’s really being able to really understand the whole equation. I think that’s one.
I think being thoughtful about how to build a business. I’ll give you just an extreme example. This isn’t an actual plan that I’ve seen. If someone showed up and said, “Look, Ryan, I want to raise 20 million to go build my own distribution of OpenStack because it can solve X, Y, and Z problem,” I’d be very concerned, because it just doesn’t seem credible. It just doesn’t seem credible that someone would be able to go do that today and go build a business around it.
As compared to, “Hey, I’m going to go draft behind these companies, because this is what they’re doing really well. This is what HP’s doing really, really well, or Red Hat, and this is where we’re going to go build on top of it, because we have all this domain expertise here. This is where we’re going to go execute it. This is how the market presents. It’s a good time for these customers to be buying, because they’ve already deployed an OpenStack cloud. That’s how we’re going to go sell in.”
Having people that have thought through that go-to-market piece I think is the other thing I look for, at least in OpenStack, because technology doesn’t sell itself. I think people have to understand how they’re actually going to get out there, and get in front of the customer and make a compelling value proposition.
Yeah. Those are the big things. Maybe it’s worth mentioning. I worry less about the, “Hey, what if Amazon deploys this and puts you out of business?”, or “What if HP builds this?”, because I think what’s happening right now is everything is moving at such a fast pace, relatively speaking. For those that have been in infrastructure for 15 years, they know exactly what I’m talking about. It is moving at just bullet-train speed right now, which means there’s just so much opportunity for startups that the big companies, I think, are not going to be able to be able to focus on everything, because the landscape’s just shifting so fast.
I worry less about competing with HP or competing with Red Hat, because I just think the market opportunity is pretty big. Like I said earlier, I’d avoid trying to … I wouldn’t want to go head to head with them on a distribution, say, “Hey, my distribution’s better than Red Hat’s because of this.” That’s tough, because if you’re … Think about it. If you’re a CIO or an IT, you’re going to want to buy from Red Hat, all things being equal, probably.
Jeff Dickey: It is moving fast. Is it moving too fast? It seems like innovation is outpacing adoption everywhere, each way I turn, in SDN, or SDS, or everything.
Ryan Floyd: Yeah. You could make an argument. There is certainly a fascination with new, shiny pennies. I don’t know what it’ll be this year, but I know there’ll be something that everyone will get enamored by this year. I think, in general, it’s all good. I wouldn’t say it’s too fast. I think it’s all good. What I love about engineering, especially in today’s market, is technology adoption, product adoption, it’s much more about a meritocracy, where people are really looking for value in those products that they’re adopting, versus the “who do you play golf with” or some kind of old sales relationship or sales model. I think that’s great. If there’s a new paradigm that comes along that makes more sense, then that’s good.
I guess the flip side of that, if I was inside of an enterprise today, I’d be pretty cautious. I would be looking to move forward, but I would be trying to be as agile as I can be, because I wouldn’t want to bet the farm on the shiny penny from 2014 only to have that change in 2015. I think OpenStack, fortunately, is past all of that. Now it’s massive adoption, so I don’t think OpenStack really fits that bill. Infrastructure in general, I think it’s a good point.
Jeff Dickey: Yeah. It seems like we need another type of Gartner for this new, fast technology, because what are the CIOs of these big companies looking at? GigaOm’s gone. Gartner is pretty conservative on this stuff. How do these people make these decisions, if they’re betting their careers?
Ryan Floyd: You raised a really good point. What I think’s needed is even different than what a Gartner does. It almost needs to be someone who’s running labs. It needs to be groups that really understand the technology, because I think, even with Gartner, the problem there is Gartner really skews towards established businesses, understandably. It’s an old way of thinking about things than where we are today.
I think if there was a better sandbox that people could really understand what works and what doesn’t and why, because the problem, even, in the media is it’s just hard to choke down sometimes, because it’s these press releases and these bullet points on features and functions that are “kind of” true, but they’re not really. A lot of these writers, it’s just hard for them, because they’re not technical enough to be asking all the questions. Often, even if they do, it’s one of those answers that “it depends.” It’s hard to really understand.
It would be interesting to see if you could almost see a lab environment get stood up, where you could qualify products and services, and really rate and evaluate stuff, because as we were talking about earlier getting actual production customers to talk about their stack and specifically what they’re done. It’s been challenging. Yeah, it’d be interesting. I don’t know what the business model for that would be, but yeah, we could certainly use it.
Jeff Dickey: Yeah. I like that. I like that idea. Working with folks, too, you can run OpenStack on your laptop and do pretty much everything, but you don’t know if it scales. That’s why I think about the whole OpenStack foundation is, what are the PTLs doing to … You see that it works, but you don’t see that it works at scale. How do you differentiate those pieces, and how do you validate … There should be a ranking system of where this feature goes.
Ryan Floyd: No, you’re absolutely right. The biggest objection, one of the biggest objections, we got at Metacloud when we talked to customers there about them moving, in Metacloud’s case it was usually existing production workloads onto OpenStack, was, “How do I know my application’s really going to run?” Because the whole, “Hey, trust me,” it works, but not really. You can’t really expect customers to trust you that it’ll work. How do you really test those workloads?
We tried … and I think it’s still up and running. I haven’t looked … to work with Internap where you can basically spin up OpenStack instances in a public cloud on demand. You could really see how things would work and how things would scale. I think more of that would help, because it would help people get more comfortable with it before having to take this plunge of, “Okay, I got to buy hardware. I got compute storage. An OpenStack vendor, I’m $100,000 in before I even figure out whether or not my workload runs.” It adds friction to the adoption process.
Jeff Dickey: We need Rent-A-Stack.
Ryan Floyd: Yes. Exactly. Jeff, it’s funny. You may have a perspective on this. Just being able to get hardware provisioned in time, it just takes too long, getting hardware from vendors. If someone makes a decision on a Tuesday, God, it would be amazing to be able to have hardware there on Thursday. Today, as you know, it doesn’t work that way. It can take 30, 60, 90 days to get hardware. That’s just another friction point in the process. You think about the advantages that an Amazon has or a Google has, or Microsoft. They don’t have to deal with that. If they need to light up more servers, they’ve got them right there. They light them up. It’s a big advantage.
Jeff Dickey: It is. On that thought, most people are using a best-of-breed vendor approach on their hardware. It just takes one piece of one vendor to get delayed for some weird reason, and everything, your whole project’s thrown off.
Ryan Floyd: Right.
Jeff Dickey: Yeah. We’re trying to solve that. It’s a tough challenge.
Ryan Floyd: Yeah. If you can, everybody will come to you. Every vendor will come to you, because you’re going to basically compress their delivery cycle. It’ll be great. If you look at what the webscale folks have done … If you think about it, you step back for a minute and think, “What are enterprises really trying to do?”, they’re really trying to build this web scale infrastructure that the Facebooks, the Googles, the Amazons have. And that’s really what they want. They want that same agility. They want all of those things that those companies have.
Most people think about it in terms of the software stack, which is a big piece of it, but there’s also the hardware piece of it that I think also is going to really evolve over the next 10 years, too, in terms of how that stuff gets delivered, where it’s much more standardized. I don’t know whether Open Compute becomes the direction everybody goes. I hope it is. I think it’s pretty cool. Something like that, where you don’t get delayed by a single vendor anymore and you can make these things happen very, very quickly, I think that will be a big piece to the overall solution.
Jeff Dickey: Yeah, Open Compute’s exciting. It’s going to have that fast pace where it’s becoming more consumable. It’s still at, we say, about 20 racks a month, a sweet spot where you want to be, which is pretty high.
Ryan Floyd: Yes. That’s pretty high. Yeah.
Jeff Dickey: Every year, it’s coming down every year. In the next couple years, it’s going to be a big deal.
Ryan Floyd: Yeah. As you see from your business, why 20 racks a month? With your help, why couldn’t it be five?
Jeff Dickey: It’s getting better. Most people that are consuming it … It’s hard to talk about folks, but they’re really doing it specialized. They are customizing this for a certain workload. They are testing that, and they’ve got folks on site that are writing drivers. It’s very specific to their workloads. They’re consuming it at a pretty large amount. It’s still at that area, in that phase, where we’re not quite down to the five-rack deployment yet. There’s still a lot of customizations that have to be made.
I was at the Open Compute show this year. Things have really … They’re accelerating much faster every year. It’s exciting to see. I had a customer at a large Fortune company talk to one of my gaming companies and was talking about that he gets 6kVA rack fully loaded on Open Compute. You just can’t do that with other folks. We are testing OpenStack on that. We are trying to create a smaller BOM on that that’s fully supported.
That’s the other thing, is where there’s a big gap on the support side. Who is going to support Open Compute? HP made some big announcements this year, and it was around switches. I was hoping that they were going to have a supported version of Open Compute that the enterprise could adopt. I think it’ll come. I think it’ll come. I think we’re going to see OpenStack on Open Compute very soon.
Ryan Floyd: Yeah. I think it will come, too. For an infrastructure investor, at least the way I see it, there is this bend towards open source that is … That’s where everything is headed. There’ll be companies that will exist that will continue to develop here and there. If you look at it over the next 10 years, that’s just where everything is headed. Then the question is, “What do you build on top on that?” Everybody moves up the stack in terms of where they add value and where they can provide value. It just makes a ton of sense.
Yeah, we’ll see how it evolves on the hardware side. Certainly, we’re seeing it, obviously, on the software side with things like OpenStack, Hadoop, and others that … Yeah, the train has left that station. It’s interesting as a venture investor, too. Ten years ago, even five years ago, there was a lot of venture investors that would tell me they were just very uncomfortable with open source. Can’t make money. There’s no models that make sense.
There’s still some that think that way, but I would probably tell them then they really shouldn’t be an infrastructure investor, because it’s just … They may find one or two companies here and there, but in general … If you’re delivering an application and you call it infrastructure, maybe that’s a little different, where it’s actually a SaaS application. If you’re actually delivering core infrastructure, all of that’s going open source.
Yeah, it’s great. It does. It presents opportunities for, yeah, things like support. How does that work? Who’s going to do it? What features do you build on top of it? How do you interact with the community? It’s great.
Jeff Dickey: When do you think that we will hit, or that open source will be the norm? Is it in the next two years? Are we still 10 years away from everything’s open source and everyone’s figured out how to make it work?
Ryan Floyd: I don’t know if I would say everything will be open source. What I mean is the foundational … It’ll be like Linux. People just, boom, that’s what they’re deploying. They may have other modules and other pieces that ride on top, but the core foundation is open source. For people that are deploying private clouds, yeah, certainly in 10 years. I don’t see how it’s not open source, because at that point, there will be enough people. The simplicity will be there. There wouldn’t be any reason not to use it.
The argument to go with VMware today, there’s many of them. It’s simple to use. You have trained engineers. You’ve already got VMware infrastructure up and running. It’s super high quality. You can trust it. Awesome support. There’s so many reasons to buy VMware today relative to OpenStack. I think with time, that’s going to shift. My bet is that VMware is going to shift with that. It’s a huge opportunity for VMware.
People always say, “OpenStack, it’s going to put VMware out of business.” I don’t think so. VMware’s got such footprint in all of these accounts. All it has to do is just slowly change its product to be able to basically deal with VMware, with OpenStack, with public cloud. It’s the vendor everybody’s going to want to work with. In a lot of ways, I almost think, with VMware, it’s theirs to lose. I think that’s where you’ll see all these companies evolve over time. They’ll just continue to move up the stack.
Jeff Dickey: We got a few minutes left. There’s, let’s see, a couple questions I wanted to get to. Actually, I think I’ve hit all of them. Do you have a favorite business model? Is that too much secret sauce? Is there something that you like?
Ryan Floyd: I have no secrets. I’m very open and transparent, both with what I’ve done well and the mistakes I’ve made. For the open source world, I think the business model that makes the most sense, and I’ve been talking about, is the one where you’re actually providing functionality on top of the open source project that is valuable to a customer set, that goes far beyond support. I think support’s table stakes.
Let’s maybe just talk about SwiftStack specifically. SwiftStack will support Swift, but the enterprise feature set that they have delivered, things like active directory integration, it’s like, “Ah, it’s boring.” I guess the open source community could do it, but who really wants to deal with that? Unless you’re an enterprise, you’re not figuring out how to do a better job with resiliency of your data store. It’s not the cool stuff with Swift. It’s these enterprise-y features that a lot of people don’t really care about. Those are great things for companies to build against, because enterprises will pay for them. Then in addition to that, obviously, you can support the Swift cluster that people are running.
I think those are the business models, to me, at least, that make the most sense today. I think part of what underpins that, too, is that whatever, whether it’s OpenStack or whether it’s Hadoop or it’s Cassandra, I think one of the things we’ve learned over the years is that in order to be successful as a commercial entity behind an open source project, there’s got to be sufficient complexity. If it’s really simple, it’s just tough, because the feature set you’re going to build on top of it, the support model, all the reasons that enterprises will want to deal with you, they’re not there. When you start talking about something like OpenStack or like Hadoop or like Cassandra, there’s so much complexity to it that there’s a good business there.
Jeff Dickey: Yeah. I think that’s why Cloudera got almost a billion dollars. There’s a lot of complexity there, and people will throw money at it.
Ryan Floyd: I think Cloudera got a lot of money because Intel knows that if it’s successful, it’ll sell a lot more processors. It’s a great business, too, but I think it was juiced up a little bit by the strategic interest there. Yeah.
Jeff Dickey: We’re just about out of time. The last question that we ask, and I forgot to tell you this upfront, but we … Oh, let me step back.
Ryan Floyd: [Crosstalk 00:51:08].
Jeff Dickey: First, are there any companies that you think that we should be looking at, in your own portfolio or things that are exciting you right now?
Ryan Floyd: You mean from an OpenStack standpoint?
Jeff Dickey: Yeah.
Ryan Floyd: You’ve already had SwiftStack on. I think it’s definitely worth getting someone from Metacloud on, like Sean Lynch or someone else from Cisco, but probably someone from Metacloud. Have Sean on the show, because I think while he probably can’t talk about specific customers, I think in terms of having insight into production use cases, there aren’t many companies that have more insight, I bet, than Metacloud today. Maybe Mirantis. Mirantis is probably the only other one that I could think of off the top of my head that has more insight there. I think that would definitely be a company to go talk to.
You could talk to the Nubeliu guys. That would be interesting, if you were looking to get an international perspective, what’s going on. We’re always so North-American-centric and what’s happening here. Nubeliu could give you a perspective on what’s happening in Latin America. The Solinea guys, it’s a services company, so Storm didn’t invest in it, but I’m a big supporter of them. They’re doing a lot of work in Asia, and they could give you a good perspective on Asia.
Jeff Dickey: That’d be great. Yeah, thank you. Where do people find you?
Ryan Floyd: I try to make myself really easy to find. I’m on Twitter. My handle is @RyanFloyd. Ryan@stormventures is email. I’m on LinkedIn. I’m very easy to find.
Jeff Dickey: Thank you so much for taking the time out and doing the show. Niki, we miss you. We can’t wait to have you back next week. Ryan, thanks. Really appreciate it. This is a …
Ryan Floyd: Yeah, hey, thanks. It was great. I love talking about it. I look forward to seeing you at the OpenStack summit in May.
Jeff Dickey: Yes. We’ll see you there.
Ryan Floyd: All right.
Jeff Dickey: All right, thanks, Ryan.
Ryan Floyd: [Crosstalk 00:53:11] take care.
Jeff Dickey: Bye.
Ryan Floyd: Bye.