 Live from San Francisco, California, extracting the signal from the noise. It's theCUBE, covering DockerCon 2015, brought to you by SiliconANGLE Media, with special thanks to Docker. Now your hosts, John Furrier and Jeff Frick. Hey, Jeff Frick here with theCUBE. We are live at DockerCon 2015 at downtown San Francisco. We've been here for two days, back-to-back coverage, going out to the events, extracting the signal from the noise, finding cool people with great stories that we can bring on the air and share their stories with you, and we're really excited about this next segment. I'm joined by my co-host, Brian Grasley, host of the Cloudcast, and our special guest all the way from the Netherlands, right? Not Amsterdam, the Netherlands. Is Eric Felixig, and he is the great title director of the Cloud Orchestra, not director of orchestration, of Nerdalize, welcome to theCUBE. Thanks, Jeff, thanks for having me. So you guys have a really interesting product, a really interesting story for the folks at Mr. Keynotes, tell us a little bit about what Nerdalize is all about. Yeah, let me explain you a few things about our business. We're doing two things at the moment. On one hand, we are offering free heating to residential homeowners, and on the other hand, at the same time, we're offering compute capacity, infrastructure as a service, below market prices. Two. Two commercial and researches. Users of infrastructure as a service that want to do compute and that care about the price performance ratio. So we're actually competing on price, which is actually very interesting because we can do this because of the combination of these two markets. So everyone now is jumping on the internet to try to figure out how do those two things tie together, and really it's a pretty ingenious bringing together of two very different things. You basically are piecing out data centers and sticking those pieces in homes. Is that accurate? So we're actually, we have built a radiator that can actually contain a Xeon-based server motherboards. So we're actually physically placing these motherboards in a heater, and we're distributing these heaters through the Netherlands, so they compose together a massive computational infrastructure. So it's a heater first with computation and store attached? Well, the thing is it's a heater for one side of the market, and it's a cooler for us, so we can actually offer computation to other parties. But is there actually a heat, if it's, if there's no computation going on, for some reason everybody's out tonight. Is there actually a heater, or are you just using the heating from the computer? No, we're using the processors as the heater. Okay, that's what I thought. The thing is like this, if you look on the heating side of the market, the regular way of heating is a giant waste. There is, at least in Europe, I don't know about the US, but in Europe 40% of the primary energy is spent on heating because it's a necessity. If you don't heat your house, you're going to die. But the way we generate this heat is just very polluting, and it's a waste. But if you see that the data center market is now using 5% of the worldwide energy from electricity, 5%, that's an incredible lot, right? And this is all 100% converted to heat. And this is in the way in the data centers. This is giving them big headache, and we're solving this problem by combining these models and cutting a lot back on infrastructural costs. Right, because they're also spending money on cooling it. Not only are they generating heat, but they're spending more in their heat dollars, on cooling. If you're talking about cooling, by the way, you got to know that the average power usage efficiency is 1.7 for data centers. This means 70% overhead on top of the regular energy you use actually for your server, and 70% is a lot. I have more innovative ways of doing that these days, and the market leaders are at 20%, maybe even lower, but we're at 6%, that's 1.06, and that is actually, that is very unique. Yeah, no, that's incredible, and that's lower than you see, you know, Facebook and Google talk about 1.2 or 1.3, like, that's 6% loss, that's incredibly efficient. So I mean, you guys in essence are, tell me if this analogy is wrong, I mean, you guys are kind of the Uber for taking heat out of data centers, and you're putting it in places where people need heat. You're using interesting resources, yeah. The thing, if you talk about Uber and Airbnb, for example, I think the main reason they're successful is that they are combining existing models, and they're more efficiently using existing assets. Yeah, right. And the stuff where we, you know, the stuff that costs a lot is assets in this world, right? You have services, you have assets, and we're reusing the roofs and the fiber optics infrastructure that is abundant in the Netherlands, and we're reusing the available electricity grid that is excellent, and we're using all these facilities to offer sort of the Airbnb for data centers. But then the other thing, because you often hear about the data centers, right? Google and Facebook that run big ones, they're constantly looking for ways to be more efficient, but in sticking it next to maybe a commercial building, you know, there's easy logic, but you're taking it to a whole other step, not putting a big data center next to a big commercial building that needs the heat, you're breaking it up into little bits and pieces. So managing this distributed infrastructure, a whole nother kind of layer of challenge that you guys have taken on. Yeah, it's super exciting. It's not for the timid, the game we're playing, and we want to be bold, and we want to do something radical. So if you look at the way we now distribute the heaters from a logistics perspective, we're partnering with Enaco, which is the second largest energy provider of the Netherlands, and they're doing the logistics for us. So that helps us a lot, and it's more efficient to put it actually in the heat, from a heat, or in the home, from a heating perspective, right? Because if you are actually transporting heat, you're going to be very inefficient. And that's the problem that the current approach is deal with is the inefficiency of transporting heat over long distances. Okay, so in essence, fundamentally, you guys are a software company. You're solving a distributed problem. That's got to be why you're here, is sort of a core customer use case for Docker, right? Distributed technology, distributed business model. The thing is, are we a software company, or are we a hardware company? For sure, a hardware company, because we're in the business of cooling hardware. And we are heavily investing in this technology, and we actually employ aerospace engineers because they can make all kinds of custom parts, from metal. We do a lot of stuff that if it falls, you hurt your foot, you know? That's the stuff we're in. And we have actually a rocket scientist because he's super knowledgeable about heat dissipation and modeling this stuff. So there's a lot of innovation we're doing in that field. And at the same point, we're also having partnerships because there is a lot of work to be done on getting in the right partners to pull this to the right level. And how did you get here? Which part of the problem were you trying to tackle? For DockerCon? No, no, no, just for your business journey. How did you get started? The thing is, we have a CEO, and he's going to hate me for this, but he's the guy who breaks the physical stuff, you know? So he once broke the thermostat in his house. He's a great guy, but he broke this. And he was like, oh, what do we do now? Because he was renovating his house, you know? And then our current CTO had, they had a very cold winter. And he had this laptop. So he was sort of hugging this laptop to stay warm and they had blankets and stuff. We've all been there. We had a laptop on our lap and he gets too hot. Yeah. Right, so it's there. But he was actually appreciating this heat, right? So he said, let's go to Dell.com. Let's order 100 of them. And we will have nerd heat, right? So that's where the name also come from. And this is how the idea came about. So you do some, you know, back of the envelope calculations seem to make sense. And later you'll learn it does actually make sense for more reasons than you even initially expected. And that's a surprising, that's a surprising, but really, really great thing. And then how big, just give people a feel for the scale of the company. How big are you? How long did you get around? How many of these devices are out in the field? Yeah, so we're three years in the running now. We're about 12 people. And we have recently closed to deal with Inneco. So they're one of the investors in the Netherlands. And we're placing device in the field in cooperation with them. Now, if you're in hardware, you got to be careful because there's sort of slower iteration cycle than with software. We have five in the field now operating very well. And the customers love it. So it's good. It's interesting, the last show I was on with John, we were talking about sort of where does hardware go? Because everyone talks about commodity hardware. I think in the data center, yeah, hardware's becoming commoditized, but essentially he's doing an internet of things type of application. There's opportunity to innovate around hardware because unique designs, unique design patterns, there's still innovation around hardware. You talked about that. There is a lot of innovation in cooling technology and the data center. And if you're looking at hardware as a commodity, I mean we have the hardware as a commodity since early days when we had these third generation languages and compiling stuff, this helped a lot because you wouldn't have to care that much anymore about where you exactly run it. We're out of the mainframe area. Then Google came about and said, we're not going to use, for example, SunSpark hardware anymore. Although it's slightly more reliable, it's still not going to fix our problems. We need massive scale, we need even better reliability. So we're going to solve it in software. And that was really innovative. And what is really great is that now that you see this whole Docker infrastructure coming about, they're making the technology that the big guys are using to manage their applications at scale and to do high availability and that sort of stuff, you see this becoming available to everybody. And now we're about taking the next step. So in order to cut costs even further, we say don't only use commodity hardware, but also software distribution to cut costs, but also distribute the physical servers. So I think there's a parallel here with the way Google decided to do distribution because it actually scales better, but it's also more cost effective. But somebody's going to go, I remember my first data center tour and it was all about the security, right? The lockdown and the fingerprint and the ice scan and all these kind of things. And of course I ripped open a door that just popped open in my hand, but that's another story. But so are there specific applications that work best in your situation? How do you answer the security and kind of the classic data center? Do you have a specific set of apps that you're going to market with or is it really just general purpose computing at a much better economics than you could get from AWS or some other cloud provider? How are you positioning? From a security perspective, it's important to know that at this point we're offering ephemeral storage and this is all encrypted. And we organize it such that the keys, we also forget them, it's only a memory. So from a data security perspective, there is not much concerns now from our customers. And actually at this point, we're offering ephemeral storage because it fits the model just easily at this point. Now if you look at the type of applications that we're offering, at this point we're doing mainly CPU intensive workloads. So it is sort of a natural fit because you're actually in the business of generating heat. But it's also very much because there's a lot of cost there for your 32 core instance at the alternative cloud provider is super expensive. So if you can actually do like CPU bound computationally intensive workloads like your protein folding bioinformatics is exploding, you have your machine learning very computationally intensive, you have your geoinformation systems and your video transcoding. There's a lot of stuff going on there that just needs massive compute, right? And yeah, we have seen in the last like 30 years that compute is just very price elastic product, right? As it becomes cheaper, people want more of it. And that's why it's good to be offering cheaper. Yeah, and I have to imagine in your business, you guys think about where it goes down the road. You've got the challenge of how do I distribute lots of heating units, but on the back end, you guys must be learning a ton about how people consume power, what patterns look like, that data, it becomes a big data problem. That becomes an interesting to municipalities potentially becomes interesting to power companies. So there's lots of business opportunities as you guys grow, I would assume. Yeah, there's actually, we're on the nexus of a number of industry trends and markets. So we're in the power industry, we're in the heating industry, we're in the glass fiber industry, we're touching with the distributed systems industry. Yeah, internet of things, all those things. And we're building partnerships such that also the environment can profit a lot. If it is about the power, for example, a lot of the people who work at the company have a background also in smart home and smart energy management. So we know a lot about this stuff. And there's also reasons, of course, that our partner in ECO is interested in this. Right, right, and orchestration's going to be a huge thing because what if it's a hot day and I need to run a big job? I mean, are you a power company first? Are you a computing company first? How do you balance those out? You have totally different customers, right? You have a different customer set. Yeah, yeah, yeah, there's two markets. And this is very confusing for a lot of people. And actually the business model is from the compute side. So because we're eliminating 60% of data center costs by not having this overhead, we can actually have a more affordable product for compute clients. So that's the real sort of business driver, right? And what we're doing in this space is we're creating products that are aimed at, for example, scientific computations where people don't really care where they run their software, but they just want a very good price performance. And that's what we can offer because of structural cost advantages. That's great. So we're here at DockerCon. Let's talk about DockerCon. Let's talk about Docker, you know, big show, a lot of buzz really at the front end of an exciting time. How do you guys use it? Why is it important to your business? So it's important for a number of reasons. Now, if you look at the talks here, it's really good talks, really smart people, and it's a lot about how to operate your software infrastructure smarter, faster, and how to accelerate and innovate here, right? And this helps us a lot. So we're using Docker internally. We're using it to build software, to ship software, to collaborate faster and smarter. And that's sort of the regular thing, right? Right. As an example, we're building Benchmarker where we have a method of reusing existing tools in the ecosystem. Like you have your InfluxDB, your Grafana, for monitoring, we're building stuff ourselves to compare benchmarking results from different providers. At the same moment, we're having Docker as a packaging format is great for us because we don't have to solve the problem of saying, you have AMI images, you have different standards, and now we have some different, neutralized format. We don't do that. We say, there is an industry standard, package your stuff in Docker. You can run it in the dark place. For your customers, right? Because they basically have your Docker package. And we can deploy their stuff at Amazon as well because Amazon supports the same format. So we can benchmark all these cloud providers and give the customer more transparency of how much their job actually costs at various providers. Right? And then turns out that we cut costs by 30 to 55%. But you can still run the job at another provider if for whatever reason it's a really hot day and you got to run it someplace else. So you're tapped out in capacity. You've got the ability to run it on alternate systems. Yeah, sure. If you don't like us, ditch us. I mean, we're in the business of offering a solution that is just more affordable. And we like transparency and we don't like locking, right? So it helps us a lot that people, if people have Docker packaged applications, they can migrate to us. And if they decide they want to move away, that's fine. Because we think we have sufficient grounds to compete on. If it comes to the heating, by the way, we actually solved the problem of the summer days, right? Because you've been mentioning this for now and it's a really interesting thing. Because it's an interesting matching. It's a different type of optimization kind of a problem. So I was really looking at, like we have this incredible challenge of scheduling this stuff, right? But the cool thing is that we're working with these physical engineers and they came to me and they say, well, we can actually transport the heat to another location. So the thing they solve this is they have the technology to create 50 degrees Celsius hot water. 50 to 55 degrees is sufficient to do very efficient heating. And this works for us, but we can also pump it to other places and, for example, ditch the heat in the summer if it's not needed or integrate with your traditional heating systems. There's heat pumps, cold storage technologies. So there's a lot of things we can do now that we can actually move the heat from the CPU to hot water. And therefore, basically the whole challenge of scheduling this is solved and it will be sort of an optimization later. But yeah, that's of later concerns because we're- We could go all day, right? It's fascinating. You come to the community-centric shows. When technology friction goes away, this kind of innovation, how it's fascinating. Right, right. So I'll give you the last one. We're getting the hook on time. And like I said, we could go all day. What's the message to the folks at home that didn't make it here to the show? What's the vibe? Should they come to DockerCon 2016? Well, DockerCon is great. The thing is that what we have seen in the last few days, Docker is really commoditizing the infrastructure for the cloud. Because they're being an abstraction layer on top. It's great for consumers. It's great for people who want to run their stuff because they have more choice. And what I see is that the cloud market is being more homogenized in the infrastructure layer. So people can actually pick on the non-functional aspects of their infrastructure. And it's great for them. And Docker is really helping in this. Great people, awesome atmosphere. Great. Again, thanks for stopping by and sharing your stories. Jeff Frick here with The Cube. Thank you so much. With Brian Graceley. We're at DockerCon 2015. We will be right back with our next segment after this short break. Thanks for watching.