Interview with Philipp from EMPA


Philipp Heer shares his insights about the challenges of integrating the ECO-Qube project in data centers, the problems with sector coupling due to the different lifecycle times of energy sector buildings and data centers, and his perspective on the future potential once the project is finished.

You can watch the whole interview here, or you can read the transcript below.

Transcript of the interview

Mo 0:05

Welcome to this next interview in our series of interviews for the ECO-Qube project. I'm joined by Philipp, who's representing EMPA, one of our pilot data centers. So Philipp, could you introduce yourself and your role in this project?

Philipp 0:27

Sure, a pleasure. My name is Philipp Heer from EMPA. I'm the Deputy Head of the Urban Energy Systems Lab. We do research for urban energy systems, typically dealing around the question: what should sustainable energy systems look like? And if we have an energy system, how should they be operated to be as renewable or sustainable as possible? Within the ECO-Qube project, we host one of the data centers that has been installed in NEST, our demonstrator site. And we're curious to figure out how to operate it as sensible as possible.

Mo 1:04

And what is innovative or different about the ECO-Qube project?

Philipp 1:10

For us it is interesting, because we can pursue research questions around what happens when society and technology change. So having more edge data centers, which are introduced in various types of locations ― Do they fit? Where do they fit? And if so, what are the synergies between the data aspect and the energy aspect? This is very interesting for us and novel set of questions we can pursue.

Mo 1:43

Specifically, it's around the edge data center element. So what potential changes the edge style of data center might have, excellent. What were your main objectives and goals in the ECO-Qube project?

Philipp 2:04

Well, one is wrapping our heads around how to integrate the data center in NEST, because it is an operational district size energy system. And how do we add such a novel type of equipment in this setup? That's one. And two is then, do we have enough renewable energy on the electricity side to operate the thing? When it is needed? When is that, when it is needed? And can the waste heat be valorized? So do we have heating demand when the thing actually operates? Or how to triage these three aspects?

Mo 2:42

So integration being the major point. The biggest challenge for the ECO-Qube project is integrating a lot of different elements: the input energy, your waste energy, the cooling system, and the IT system. Really cool. And how is the project going? We’re one year into a three year project, what’s happened so far from EMPA’s perspective?

Philipp 3:11

We were mostly focusing on the data center integration part. So get physical infrastructure ready, in a sense that we have the interfaces to connect the thing, but also data accessibility, in terms of what are the virtual interfaces that you need to actually deploy computational jobs there, that we get the measurements from the system to actually calculate what's the efficiency, or in other words, the KPIs of the ECO-Qube to actually be able to calculate based on the data that we acquire. We spent lots of time around that. We are looking forward to the next phase, that will be controller development, how to operate the thing.

Mo 3:56

So setting up that and actually doing the integration, that's what we've been doing for the last 12 months. What were the biggest challenges that you came across in that last 12 month period?

Philipp 4:10

Well, it's quite a multifaceted project. So lots of different interests, in the sense of different stakeholders along the value chain have to valorize such an edge data center. We come from the research side and from the energy perspective. We learned a lot when it comes to what important aspects do we need to cover from the data side, such that sector coupling makes ― sector coupling in the sense not electrical and thermal energy, but value chains of electrical, thermal, and data stakeholders and there's quite some potential there, not just in the research project, but actually often out on the market as well.

Mo 4:56

So to summarize the challenges there were integration and complexity. It is a multifaceted project. So resolving that complexity was probably the biggest challenge that you coped with. And then you started to talk again about the future and the opportunities, which is sector coupling, which no one's really cracked. So I wonder if you would like to talk a little bit about how ECO-Qube starts to problem solve within the realm of sector coupling.

Philipp 5:33

There are different sectors that need to be coupled with technologies and that can be heating systems, from electrical, to thermal, and so forth. Compared to what we want to do at ECO-Qube, classical sector coupling would be easy. Because we add this dimension of having different locations with computational jobs that can be allocated on a geographical level, which has an impact on these different decentral energy systems. So not just having an additional stream of the data aspect of the sector coupling, but the decentral multi-location aspect of this as well. And when we face this challenge, I'm very confident that we can quantify what's the added value of that, once we deploy this in an operational system, or three in ECO-Qube to see where's this is heading. This is super interesting.

Mo 6:42

Maybe we can talk a little bit about that. In the first sense of the word sector coupling, we're talking about energy and data center infrastructure, then in the second sense, in the deepest sense, we're talking about load migration, and data center architecture, and energy architecture, across our society and how that might change. So maybe if you can dive in. Why might it change? And what are the opportunities of that change?

Philipp 7:18

The change is driven by new needs, in terms of computation. The need of computation increases on a societal level, with the assumption that this can be done centrally, but there's still an increased need to have this in a decentralized manner. So the argument is, we need to have computational power and there's heat related questions around that. It’s not, okay, we have heat related questions and we want to solve that with edge data centers. So decoupling where something is coming from, towards identifying where the synergies between the sectors are, that's quite a large step, especially when we have established sectors or established players in sectors. With ECO-Qube it's more or less new lands that we can discover. And to settle on this new land with approaches that we find bring added value that in a classical sense would never have been identified is the actual potential that we have here. And to think not just in one installation, one implementation, but multiple ones, which can be coordinated.

Mo 8:44

So yeah, so that is a coordinated new architecture. And it's fundamentally different to how we do digital architecture, digital infrastructure currently, and what you're saying there was that a lot of the other sectors ― heat, energy, are very mature. And, the opportunity exists with the data center industry because it is still very young, it can make a turn more easily.

Philipp 9:17

That is true. One of the reasons for that is when we consider buildings to have a half lifetime of 50 years, and the infrastructure of data centers has a half lifetime of like three, four or five years at the best. So there are generations in between when we try to combine these different aspects. So the built environment exists from the data perspective. And now the challenge is how to leverage and bring these sectors together with fundamentally different approaches, so the established sectors and new technologies. It's an interesting playing field coming from the research perspective.

Mo 9:58

Absolutely. I think that's probably one of the biggest juxtapositions. The energy industry that works on a 50 year plus timeline, and the data center industry which is working on an increasingly short timeline, where probably three years is a decade. So we've touched on how the ECO-Qube project is really the forerunner of a potential change in architecture, and that also a change in perspective, right? It's how we provision digital and energy infrastructure, because it allows us to do it in an integrated decentral way, where previously we've been doing it in a non integrated, centralized manner. So there's a cross cutting, even within these sectors there's a number of different things that we're talking about, because we're talking about the trend in edge data centers, which is closer to people. We're also talking from the energy perspective about a trend towards decentral energy sources, namely, wind and solar farms. They're not gigawatt power stations like they used to be. And then we're talking about the integration of not just the decentral asset into the energy system, or the edge data center into the digital system, but actually, across each other. So it's really layers of complexity. I want to talk now about what could come next. So the next two months in this project? Why don’t we start there, what's next in the next two years?

Philipp 11:54

What's next after having commissioned our data center, or the data center that we have here, we're eager to start collecting data and get a feel of what's the operational efficiency of computing in such an edge data center with an attached cooling system and identify the potential ― How can this operational efficiency be improved on the scope of the data center and cooling system itself, and then expanding the scope within the energy system of the district to identify first potentials and energetic flexibility to be able to shift operational loads. So to start data collection, to identify based on the data what's the potential, and to derive methods for operational improvements from there. This will be the main focus of the next three to six months, which won't be done by EMPA alone, but also by partners of Work Packages three, four and six. So, a lot of outreach will be necessary there.

Mo 13:04

Excellent. And then what happens at the end of the project? It's a hard question because we don't know what will happen in two years time at the end of the project. But if everything goes well and ECO-Qube demonstrates how you can make a data center smart and integrate it into its environment, it could be pretty revolutionary. What would you see? So, you have a license to have your opinion now. What do you see could come next? Because at the moment EMPA is a quasi-research data center ― Is it pilots? Is it like you said from a regional perspective, and then we would build out? What do you think would happen? I give you freedom to really have an opinion here.

Philipp 15:18

So, in an ideal case, at the end of the project, I would say that we will be able to show what the potential of such a cooling system and edge data center integrated in an energy system is, in the locations that we have in ECO-Qube project. But, also calculated locations to provide the potential if installed in new locations, new geographic areas, new energy loads, new load compositions, new compositions of data centers, and what the potential of such a system is, to be actually capable to quantify the added value of such a system.

Mo 16:04

So if I understand you correctly, what you're talking about is mapping the opportunities. And in order for that to happen, what you're really talking about is mainstreaming the ECO-Qube approach, because you're basically saying energy utilities, DSOs, TSOs, data centers, telecommunications, will all be considering how to integrate, which is considering the ECO-Qube approach. So what you're really saying is, by mainstreaming the ECO-Qube approach, you see a lot of these synergies happening. And that is what you think is next, is that right?

Philipp 16:46

In that sense yes, mapping ― while the project Horizon is finite, so the amount of potential locations that we can quantify will be limited. But starting this mapping process is an initial major step towards this mainstreaming process.

Mo 17:05

Excellent. And last question, Philipp, it's just an opportunity to say some final words before we end the interview. Describe the ECO-Qube project in three words.

Philipp 17:31

What is it? It's four: high risk, high potential. Because personally, I've never touched a data center before in my life. I come from energy research. And to be able to not just assume and simulate, but to actually have such a thing locally here, to test it, to benchmark it, and to be able to show for a sustainable energy system, how valuable what we have is. And what added value can it provide to society? That is really awesome to me. And not just you know, that we have it here in our basement but rather having this multifaceted project with multiple people involved from different perspectives, to actually get something tangible out of it and not just something a researcher can play with. That is quite exciting to me.

Mo 18:28

Probably quite satisfying as well. Brilliant. Thank you, Philipp. Thanks for the interview. Great to get your insights from a pilot data center's perspective. So these are the guys actually testing and implementing the ECO-Qube solutions. So Philipp, thank you for joining us today.

Philipp 18:48

It was a pleasure. Thank you very much.

Back to all News