Max Schulze, NBF | KubeCon 2018
>> From Seattle, Washington, it's 'theCUBE' Covering KubeCon and CloudNativeCon North America 2018, brought to you by 'redhat' The CloudNative computing foundation and it's ecosystem partners. (upbeat music) >> Hello everyone and welcome back to live CUBE coverage here at Seattle for KubeCon, CloudNativeCon2018. I'm John Furrier. Stu Miniman, breaking down all the action here for CloudNative, trend, a lot of ecosystem partners, a lot of new developers, a lot of great open-source action in the cubes here covering it. We've been there from the beginning, our next guest and user, Max Schulze, Advisor and Founder of NBF, welcome to the CUBE, thanks for coming on. >> Thank-you, thank-you for having me. >> So tell me about what you're working on. You are doing something pretty compelling with Kubernetes and CloudNative, take a minute to explain what you do. >> Yeah actually, we are advising a very large energy utility in the Nordics and what we're trying to do with Openshift and Kubernetes is actually to shift loads between different data centers based on power availability. So if you have wind and solar power, you know that you only get energy when the wind is blowing so you really need to be able to match that load of the data center with the actually energy production which is quite challenging to be honest. >> Max you have different take on 'Follow-the-sun' that we used to talk about in IT I'm guessing, yes? >> Yes >> Take us inside a little bit, the sustainability is really interesting and how some of the power, you know, usage and heat and everything and maybe you can explain that a little bit before we get into the data. >> Of course, so generally how we got to a sustainable data center source was that in the Nordics you see a big growth of data centers in general so all the hyperscalers: Google, Microsoft, AWS. They are all coming to build data centers in Nordics. It's cold, power is cheap, you have lots of renewable energy available and we started to think 'Okay, but they have two problems essentially.' They generate a lot of heat, which is just emitted into the atmosphere so it's wasted, and the second problem is that they want 100% reliable power and reliable power you only get from nuclear, you get from gas, coal fire power plants not from renewables. So we looked into this, and we started to think about okay can we maybe get the heat out? Can we extract the heat from a data center and inject it into district heating grids and actually heat homes? With a hyperscale data center from Microsoft, 300 megawatts you can heat about 150,000 homes, that's quite significant. >> Yeah and how are you doing that? I mean I talked to a company once that was like 'Oh well we're going to, you know, we'll just distribute the servers different places and there will be ambient heat off of it.' But you're extracting the heat and sharing it. Explain that a little bit more. >> So most existing data center projects, they extract the heat out of the air but that's really inefficient. You get to about 100 degrees Fahrenheit which is not uh high quality heat. So what we want is 140 degrees Fahrenheit, about 60 degrees celsius, which means that we have to use liquid. So we have to use water in this case and we use a cooling system that is quite ironic from a start up in Germany called Cloud & Heat that uses hot water to cool servers. So the water really flows at a very very high speed through the data center and on it's way picks up a very low amount of temperature and we get out the temperature, we get out the water at 140 degrees Fahrenheit and we put it in at 120 degrees Fahrenheit. So it's quite, not a big difference, but it flows at a very high speed. >> So it makes it work? Makes the numbers work. >> Exactly. And so what's the home count again you mentioned one hyperscale data center, like a Microsoft data center powers heat for how many homes? >> About 150,000 homes from 300 megawatts worth of data center. >> And you guys put this into a grid so that's, does the location of the homes need to be nearby, is there a co-location kind of map or? >> Yeah actually, in order to do this we have to move data centers closer to cities. But luckily, data centers actually want to be closer to cities because your closer to peering points and one of the reasons why they usually can't come closer to cities is because power is not available near a city. So we um try, we can give them both. Right, they can come closer to the city and we can give them power, and we get the heat in return. So, so everybody wins. >> Yeah so I mean, a lot of the discussion we've had is the interaction between software and my data center infrastructure. You've got a story of software, with you know, actual like city underneath the infrastructure. Maybe you got to help explain how that was built out, what tools you're using and walk us through this all. >> So we originally started with Openstack, which was the first test because we need, in order to do this heat extraction we need to also steer really the software, the workloads that run on the data center because you know a chip only gets hot when the server actually does something so we really had to figure this out. We started with Openstack and then we started looking into load shifting which immediately brought us to Kubernetes and then Openshift because you can use the internal scheduler to basically force loads across different locations. We connect it to our energy systems, to our forecasting systems and to our heat load management systems and then basically push workloads around. Right now we have two sites where we test this and it's not as easy as it sounds. And we basically want to move workloads, concentrate them where we want, we have heat. So um yeah, Redhat is helping us a lot doing this but still it's not that easy. >> Yeah yeah, it's interesting. You know, I think back you know, virtualization was about you know, how can we drive some utilization and get some out? You really want to you know, concentrate and run things hot. >> Yeah, exactly. >> Quite inter- Alright tell us about your involvement in this ecosystem, you know, what brings you to the show this week, what do you get out of coming to a show like this? >> Yeah, actually I came because Redhat invited us to talk at the Openshift gathering at the beginning of the conference. And generally, we don't really have a commercial interest in making data centers or data infrastructure sustainable, we, we don't gain anything from that, but we believe it's necessary. If you look at the growth curve of data centers you can really see that they will consume more and more power, and then the power they consume is not compatible with renewable energy. So we are hoping that we can influence people and we come here to tell people our story and we actually get great feedback from most of the nerds. >> Well it's a great story. It's one of those things where you're starting to see data centers trying to solve these problems. It's great with the renewable energy, having that kind of success story is really huge. Um, You mentioned that data centers want to be close to cities. I got to ask the question, in Europe, well you've lived around a lot of places. Is there a more cloud city oriented, like is it London, you got Paris, you got... I know Amazon's got data centers in Ireland. Is there certain cities that are more CloudNative culture? How would you break down the affinity towards CloudNative? If you had to map Europe, which major countries and cities would you think are advanced, cloud thinking vs. tire kickers or you know, people just kind of just trying it? >> In Europe there is a region called the FLAP region, that's Frankfurt, London, Amsterdam and Paris. Those are where you have the highest concentration of data centers, but it terms of CloudNative adoption, I would say that probably in the UK you have the most adoption rates and in the Netherlands. Germany is always, I am German so I can say this, we are always a bit behind in terms of cloud technology because we're a bit scared and we don't know- >> You'll watch everyone test it out and then you guys will make it go faster. (john laughs) >> Maybe, maybe, maybe a bit more efficient but uh, generally I think the cloud adoption rate in Germany is the lowest and the UK and the Netherlands is the highest I would say, yeah. >> Awesome, well thanks so much. Congratulations on your success, we'll keep following you and when we're in Europe we're going to come by and say hello. Thanks for coming and sharing the stories. The CUBE, breaking down all the action at KubeCon, CloudNativeCon. I'm John with Stu Miniman. Day 2, we got three days of wall to wall coverage. Thanks for watching. (upbeat techno music)
SUMMARY :
2018, brought to you by in the cubes here covering it. minute to explain what you do. load of the data center with some of the power, you know, and the second problem is Yeah and how are you doing that? So we have to use water in this case Makes the numbers work. you mentioned one hyperscale data center, of data center. the city and we can give them with you know, actual like So we originally started You really want to you know, and we actually get great How would you break down the in the UK you have the most it out and then you guys will Netherlands is the highest I would we'll keep following you
SENTIMENT ANALYSIS :
ENTITIES
Entity | Category | Confidence |
---|---|---|
Ireland | LOCATION | 0.99+ |
Europe | LOCATION | 0.99+ |
Max Schulze | PERSON | 0.99+ |
Amazon | ORGANIZATION | 0.99+ |
John Furrier | PERSON | 0.99+ |
Microsoft | ORGANIZATION | 0.99+ |
Germany | LOCATION | 0.99+ |
Stu Miniman | PERSON | 0.99+ |
ORGANIZATION | 0.99+ | |
100% | QUANTITY | 0.99+ |
AWS | ORGANIZATION | 0.99+ |
UK | LOCATION | 0.99+ |
Paris | LOCATION | 0.99+ |
two sites | QUANTITY | 0.99+ |
John | PERSON | 0.99+ |
Netherlands | LOCATION | 0.99+ |
Amsterdam | LOCATION | 0.99+ |
London | LOCATION | 0.99+ |
Frankfurt | LOCATION | 0.99+ |
300 megawatts | QUANTITY | 0.99+ |
120 degrees Fahrenheit | QUANTITY | 0.99+ |
second problem | QUANTITY | 0.99+ |
140 degrees Fahrenheit | QUANTITY | 0.99+ |
Nordics | LOCATION | 0.99+ |
two problems | QUANTITY | 0.99+ |
Redhat | ORGANIZATION | 0.99+ |
one | QUANTITY | 0.99+ |
Seattle | LOCATION | 0.99+ |
Seattle, Washington | LOCATION | 0.99+ |
KubeCon | EVENT | 0.99+ |
both | QUANTITY | 0.99+ |
three days | QUANTITY | 0.98+ |
CloudNative | ORGANIZATION | 0.98+ |
this week | DATE | 0.98+ |
john | PERSON | 0.98+ |
first test | QUANTITY | 0.98+ |
CloudNativeCon | EVENT | 0.98+ |
about 150,000 homes | QUANTITY | 0.98+ |
NBF | ORGANIZATION | 0.98+ |
about 100 degrees Fahrenheit | QUANTITY | 0.97+ |
CloudNativeCon2018 | EVENT | 0.97+ |
about 60 degrees celsius | QUANTITY | 0.97+ |
About 150,000 homes | QUANTITY | 0.97+ |
CloudNativeCon North America 2018 | EVENT | 0.96+ |
KubeCon 2018 | EVENT | 0.92+ |
Redhat | PERSON | 0.92+ |
Day 2 | QUANTITY | 0.89+ |
Openshift | EVENT | 0.87+ |
CUBE | ORGANIZATION | 0.86+ |
Openstack | ORGANIZATION | 0.84+ |
FLAP | LOCATION | 0.79+ |
Heat | ORGANIZATION | 0.74+ |
German | LOCATION | 0.72+ |
Kubernetes | ORGANIZATION | 0.7+ |
CloudNative | OTHER | 0.6+ |
Openshift | ORGANIZATION | 0.57+ |
Kubernetes | TITLE | 0.51+ |
data | QUANTITY | 0.5+ |