Green IO

Gaël DUEZ
undefined
Sep 17, 2024 • 1h 2min

#45 - Assessing digital sustainability’s maturity with Aiste Rugeviciute and Rob Price

In the early 2020s, companies started facing a big question: how could they be more responsible in the digital world? Could something similar to CSR exist for this virtual and yet highly materialized world? Corporate Digital Responsibility (CDR) was coined to offer some much-needed guidance.🎙️ To explore its ramification, Gaël DUEZ chats with two renowned experts in CDR: Aiste Rugeviciute, co-author of “B.A.-BA du Numérique Responsable” and now pursuing a PhD in the socio-ecological impacts of CDR strategies, and Rob Price, a key player in developing an international CDR framework. Rob also hosts the “A New Responsibility” podcast, diving deep into CDR's role in business.Some Takeaways:    🔑 the CDR framework in a nutshell,   🌿 the importance of embracing a balanced approach to CDR in most companies, and   🛠️ a sneak peek to the newly-released CDR maturity model.❤️ Subscribe, follow, like, ... stay connected the way you want to never miss our episode, every two Tuesday!📧 Once a month, you get carefully curated news on digital sustainability packed with exclusive Green IO contents, subscribe to the Green IO newsletter here.  📣 Green IO Paris is on December 4th and 5th 2024 --> use the voucher GREENIOVIP to get a free ticket! Learn more about our guest and connect: Aiste's LinkedInRob’s LinkedInGreen IO website Gaël's website 📧 You can also send us an email at contact@greenio.com to share your feedback and suggest future guests or topics.   Aiste and Rob's sources and other references mentioned in this episode:CDR ManifestoB.A.-BA du Numérique ResponsableA New ResponsibilityMaturity ModelEthos FoundationGlobal action plan for a sustainable planet in the digital ageTranscriptIntro 00:00To change, we have to think about how organizations or governments incentivize change in an economy and how that helps businesses to do the things that need to be done better.Gael Duez 00:25Hello everyone, welcome to Green IO with Gael Duez - that’s me! In this podcast we empower responsible technologists to build a greener digital world, one byte at a time. Twice a month on a Tuesday, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the Tech sector and beyond, to boost Digital Sustainability.One of my more esteemed peers in Digital Sustainability, and a good friend, who’s now CSO of a large digital tech company used to work before in the garment industry. Sometimes, well quite often, she goes ballistic about the infant level of the digital industry in sustainability compared to other industries. The very idea that digital technology produces products as does the garment industry with clothes and that these products have a footprint both environmental and societal is not that widespread. And when Tech executives become aware of it they often lack the frameworks, the best practices and the metrics to steer the sustainability angle of their company. And answer this pivotal question in an industry, which has just a touch of ego and hubris: how good am I compared with others? And eventually, how good am I with keeping our planet habitable for the human race. Am I being too sarcastic here? Well let’s go back to a more action-drive mindset then. Since the beginning of the 20’s, a concept has started to emerge embracing these questions and providing some framework for companies with a significant use of digital technology: the idea of their CDR. Full disclosure, I am using this approach with clients when I do consulting gigs so I might be a bit biased. As usual, question and double check everything that is said in this episode. All the references will be put in the show notes on greenio.tech and on your favorite podcast platform. Transparency and Accessibility remain in the DNA of the Green IO podcast. To deep dive in CDR, I have the pleasure to welcome 2 of the best experts we can find on the topic. Aiste Rugeviciute who graduated in both computer sciences and Sustainable Development and Social Business and has worked for several years on the intersection of Tech and sustainability, she co-wrote the hands-on book "B.A.-BA du Numérique Responsable" in French and she is now doing a PhD on Corporate Digital Responsibility (CDR) strategies and their socio-ecological impacts. Rob Price is one of the core members of the international group of academics, corporate practitioners and published authors who collaborated in 2021 to aggregate their existing body of work into a single, international definition of the set of principles supporting Corporate Digital Responsibility. He’s also a fellow podcaster with the “A new responsibility” podcast which covered during 5 seasons the use of CDR in companies. He gave a much listened talk to Green IO London last year. And btw this year Aiste will be the one giving a talk on her CDR Maturity Matrix. So make sure to get your tickets for this great gathering of responsible technologists on September 19th. And without further notice, Welcome Aiste and Rob. Thanks a lot for joining Green IO today.Rob Price 03:55Thank you.Aiste Rugeviciute 03:56Hi, thanks for having us.Gael Duez 03:59You're more than welcome, both of you. I have a terribly complicated question to start. Could you define CDR in two sentences maximum?Aiste Rugeviciute 04:12Well, it is a complicated question because there is actually no one definition, at least between the academics. Everybody comes up with their own definitions. But in a nutshell, I would say there are two main ideas. So the first idea is, of course, about responsibility. So what are companies' responsibilities with respect to digitalization, or the way they use digitalization? And the second one is trying to maximize the positive effects of digitalization while reducing the negative ones. So that's kind of the two main ideas combined. Rob, what do you think?Rob Price 04:57Well, I mean, it goes back to the introduction, really, doesn't it? So in 2021, that was the question we asked ourselves, and we wanted one sentence that described CDR in the context of all the definitions that the variety of us had created at the time. So, I mean, reading that corporate digital responsibility is a set of practices and behaviors that help an organization use data and digital technologies in ways that are perceived as socially, economically, and environmentally responsible. I think the key thing for us was, at the time, trying to think of something that provided a framework and guidance to help organizations be aware of the consequential impact of the things that they were doing. And no doubt, through the conversation, we'll talk more about finding the right balance in terms of framework versus measurement criteria versus a method, if you like, in terms of organizations doing those things.Aiste Rugeviciute 05:57I think also one of the key things, what Rob just said, it's about the perceived value. What's in line with the so-called reference points in the society. That's kind of a key thing when corporations, organizations think in general about their responsibilities, and when it comes to digitalization, aligning with societal social expectations and all the idea, you know, if it's contributing to reducing climate change or any other negative impact, it's all about in line with what right? In line with what expectations and whose expectations. That's where the question of responsibility comes in place.Rob Price 06:48And can I build on that very briefly, which is, I'm very conscious of talking about CDR in multiple countries around the world. I think it's very difficult to be really tightly defined, because you have to be conscious that different countries, different political systems, different cultures, there would be pushback if you kind of introduced CDR as a mechanism that you had to follow. Literally. It's important to be conscious of the environment in which you're asking organizations and governments to think about the way in which they operate. Even if they take some of those aspects, that's better than all of those or none of those.Gael Duez 07:31So it's really a question of contextualizing the approach, and still we reach some kind of agreement around the world and I guess some principles, some guidelines, as both of you mentioned. Could you maybe elaborate what are the main guidelines, the main items that have reached some sort of consensus that, yes, we should pay attention to this aspect and this aspect and this aspect.Rob Price 08:02So, in a sense, the blank sheet of paper that we did start with in the past was probably between about 2016 and 2020. A number of definitions of CDR did appear. Some of those were written from business points of view, or government points of view, or academic points of view. But when we did an analysis of them, I would say that 80% of them were pretty similar. Maybe some were focused more around sustainability, some were more focused around trust and inclusion, but 80% was common. And the purpose of the work that we did at that time, with all of those parties bringing together those definitions, was to try and find a common framework that actually was inclusive of everything. So we defined a framework of three intersecting circles, which gave us seven principles. If you think of each of those sections, at the heart of everything was trust and purpose was beginning to be talked about more commonly. So, purpose and trust, fair and equitable access for all. So that's around equity, diversity, inclusion, and more societal well being. Thinking about the impact on people, one of the things that I almost enjoy is when I'm talking to organizations asking if their products and services are predicated on addiction to drive advertising rates, it's always an uncomfortable question. But nonetheless, thinking about the impact on people, considering some of the economic impact, I think is interesting. The fourth principle, economic and societal impact, and probably one of the hard ones. But it's beginning to think about some of the things around algorithms, fair share of outcomes, of benefits, the way in which you value things that you're delivering through digital technologies. Talk about the impact economy in principle five. So, thinking around, or beginning to think about that intersection with the economic and the sustainable side of things, and goes into more detail about supply chain and green tech, and some of the ways in which you can use digital services to directly impact, innovate around sustainability. And the final two, very much focused around sustainability. I'll start with seven. Seven, reducing the tech impact on the climate. You're thinking about what is happening with my data centers. Am I using renewable energy? Is it really renewable energy? Etcetera? And then six, we talked very briefly at the start around innovation, thinking about how I can use these technologies to innovate and solve the world's biggest problems, especially around sustainability. So I think this is a really important one. I want to explore it later. We talk a lot around impact of AI, for example, which clearly has massive energy use, massive water use, in terms of data centers. That's something to measure principle seven. But it's really important to think about principle six in terms of how I can use that technology to create positive impact and benefit on those very things that I'm worried about negatively impacting. And it's a balanced scale. So those are the seven principles. They go into far more detail through the manifesto. There have been other definitions that have emerged since, but I can track them all back to those seven principles. I haven't yet seen something appear that's new, a new concept that doesn't fit into that framework. And of course, we do continue to look at those definitions to determine how we continue to evolve, because I don't think any of us said when we created this back in 21, right. That's it. We've cracked it. Nobody could ever improve on that. It's always evolving, always a living thing, and then it's how people bring that to life, which is probably a perfect position to hand over to Aiste, in terms of the work that you've been doing.Aiste Rugeviciute 12:05Yeah. If I might just add or more generalize. So those seven principles, they are a little bit in more detail, but if we take a step back, basically, we can summarize that it's about the environmental and social side. So we have both sides, and that's what Rob was talking about. Right. So there are social, societal questions and there are environmental questions at the same place. Why I'm underlying it, because sometimes especially, I don't know why, especially in Germany, when they used to speak about corporate digital responsibilities, they completely ignored, or they tend to ignore the environmental side. So they tend to look only on the social, societal side and especially concentrate on the data questions. Now, it's shifting a little bit, but at the beginning, I remember when I was reading it, I was surprised to find it because it was completely the other way around. In France, everybody, whenever they were talking, they were talking only about environmental questions. And the social societal questions were put aside and that was being dealt with by more, way more engaged people or very niche, niche people in certain NGOs. Currently, I can see that the discourse is becoming more balanced, which is nice. So it takes into account two sides. So as I said, the environmental and social, societal side. But it's also the question of technologies perceived as a solution or as a problem. So that's we have both technology or digitalization. So we have to generalize two dimensions. So there is one environmental, social, societal side, and then is it good or bad, or again, it should be balanced discourse, but that's the attitude as well.Gael Duez 14:09And I have one small question just to clarify. The principle number seven, which is the impact on. Rob, you mentioned impact on climate. Is it only climate or is it all kind of environmental footprint?Rob Price 14:25Yeah, no, it's reduced tech. Impact on climate and the environment is the actual words that it says. So it is that consequential impact on everything that you have directly or in the supply chain, on the climate and the environment, planet, etcetera. And I think I wanted to just add something to the thing, the description ack, because it is absolutely about people, communities, society, planet in its entirety. I don't want to forget what I think is probably the hardest part of it to think about sometimes, which is the economic kind of angle, if it changes. We have to think about how organizations or governments incentivize change in an economy, in their geography, in their region of responsibility, if you like, and how that helps businesses to do the things that need to be done better, whether through incentive or whether through regulation. So it's that intersection of all aspects of that. I think it's complex. It's not easy. I mean, we've seen that over the past few years. Many organizations that I speak to, they're not trying to solve everything. At the same time, that's not realistic. It's about understanding the consequence of the things they're doing and honing in on some things that would make a difference, that are important because it positively impacts what they're doing or the way in which they're seen in the market, the respect they're given, if you like, the ability to better place product or service because of those things that they're doing. So it is a finely tuned thing engine that you need to kind of understand how we can nudge it in the right direction to be better in terms of the impact it has on the environment and everything around that. If we were talking seven for example.Gael Duez 16:26And now that we kind of lay the ground with the main concepts and the subtle equilibrium between them in the intersection and the kind of systemic interaction that we can see between the seven of them. If I'm listening to this episode, and I'm convinced that, ooh, this framework sounds very interesting, how do I go from the academic and conceptual approach that helps me to better understand what I should pay attention to, to something a bit more concrete for me or actionable for me? And my main question actually will be, Aiste, do you believe that the CDF principles can become actionable?Aiste Rugeviciute 17:12I would very much hope so.Aiste Rugeviciute 17:17There are different steps before they become concretely actionable. Right. So the first question is, I think to ask, why do you want to do that? And then to ask, what do you want to do? So why? I think it's an important question, which, especially in the organization corporation environment, sometimes it's forgotten by the practitioners and then they are being asked by their managers, you know, and they're like, yeah, but why? What's the purpose? Right. So the question why is very important because most of the time the response is money, is to make profit, to respond to the stakeholders. Right, sorry, shareholders actually. But beyond that, we have other requirements. And that's where I come back to the same notion of responsibility. So the why could be because there are legal requirements, there are risks associated, there are opportunities, there are myriad of stakeholders who suddenly have shifted expectations. So your direct clients, it might be your B2B clients, right? It might be the societal legitimacy for you to stay active in a region, in a state, in a country. It might be simple because you want to benchmark yourself with others, so there are reputational damages, et cetera, et cetera. So the first question is, actually why do you want to do that? Is it because you want to go back to reduce risks, take some opportunity, go beyond your legal requirements and so on. Once you kind of clarify the idea why you want to engage in the CDR, it's going to be way easier going forward to structure your actions and approach towards these guidelines. So let's imagine that you are B Corp, for example, and your idea that you want to do CDR because you want actually to contribute to creating a better society. And you consider that yourself as an organization corporation. That's one of your responsibilities.Gael Duez 19:55By B Corp, you mean the label that some companies get.Aiste Rugeviciute 20:00Yeah, sorry, I put the label because I know that in the US, when you get a B Corp label, it's much easier to act in line with societal expectations, and not only with your shareholders, because in the US there's a lot of shareholder pressure to get profits, monetary profits. But once you have a B Corp label, they change status in their corporation, legal, something with legal things that in that case, corporations are expected or are allowed way more freedom in order to act in accordance with societal expectations.Gael Duez 20:48Thanks for the clarification. And let's go back to what you were saying previously.Aiste Rugeviciute 20:55Yes. So once the company decides why they want to do that, the next stage is to understand where they are now, right? So maybe you are already doing very well and your corporate digital responsibilities. So once you clarify your responsibilities, the question is, okay, what is our current state? And once we evaluate the current state, once we reflect on what we are doing now, then we can look at how we can improve ourselves, where we can go in order to go to our objectives as well. A question is, objectives, is it a continuous objective? Is it very concrete that you want to, I don't know, reduce your CO2 consumption by your information systems, or you want to reduce the digital divide in the society? So digital divide in the society, you know, you can have very concrete milestones, but it's a continuous action, right? It's not going to be at some point completely achieved over. So that's basically the three steps. So the first question is why you want to do it. Second one is analysis, understanding where you are now. And the third one is putting the milestones, objectives, priorities. Because as Rob mentioned, you can't do everything at the same time, especially as an organization, you have so many things on your plate. So, you know, you have those seven principles which Rob mentioned, but between the seven principles, maybe some of them have higher priority because of where you work, because of your industry, because of your stakeholders, because of the whole context.Rob Price 22:42Completely agree with that. And just, it's a really important point, I don't think, when everyone who fed into that work was never saying there was an equal balance, or needed to be equal investment to address and mitigate each of those separate principles. It's a framework, and the framework is then applied to the organization. I think very few times has an organization ever come to me and said, we've decided we want to implement CDR, how do we do that? Most times it will be they're worried about reputational damage, or they know that they need to improve something around their impact on the environment, or how do they get trust, because they're worried about reputational damage. So it's more around applying it as a framework to help that organization resolve the concern. Or maybe there's been an incident that's highlighted a problem that is perceived across that organization, and it's how they begin to find a pathway to make an improvement that improves the thing that they're concerned about.Aiste Rugeviciute 23:52Always some kind of entry point towards that CDR, as Rob said. So it might be reputational, it might be very often a regulation, currently at least what we observe in France. So you have regulations which ask you to reduce your electricity consumption, energy consumption, and reduce your CO2. So that's an entry point. So companies say, oh, I need to reduce my CO2 somehow. So let's try to find all possible leverages. And then when they start looking and they say, okay, well, I might reduce my CO2, for example, or environmental footprint by reducing the number of screens we have per collaborator. So my overall footprint, environmental footprint is going to go reduced, but then that's going to affect productivity levels of my collaborators. So wait a second, that's kind of related to something else, right? And then it might still mean, oh, it's very complicated, the digitalization as a problem itself, it's a systematic problem, and there are different impacts which are very much linked with each other and they're not necessarily going in the same direction. If you reduce one, it can be at the cost of the other. And I think that's where the structured approach of CDR, reflecting on it, allows organizations to take into consideration more things which are more important to them and their stakeholders and the challenges they are trying to respond to.Gael Duez 25:29Rob Aiste just shared a very concrete example on this kind of trade off and multidimensional approach. Could you share an example of a client? You can name it or not. You might be under an NDA that started with you. A journey, I would say, toward CDR, and to see the few steps that were taken and how a bit more concretely than some actions were taken.Rob Price 25:58I'll use a couple of examples that are not my clients, but they're well discussed, well documented, and in the public domain, if that works, which is probably, and they're both, well, both international organizations. So the first one is Merck and the work of Dr. Jean Enno Charton, who's head of bioethics and digital ethics at Merck. And I think, why do I start with that? A large organization, I think 60,000 people, something of that order around the world, but working in a space that's clearly got strong insight around regulation and controls because of the nature of what they do already. So there's almost a lot they would be, I think, one of the first types of organizations to start thinking about broader CDR and digital ethics in the wider round. And therefore they've focused in and put people and built a small team to build out a framework that takes the best in terms of some of the guidance that's around and early thinking around digital ethics boards and mechanisms and frameworks, and created something that they believe is right in their organization. And I've seen that build and progress. I've presented with Jean Enno in the past, and it is a very thorough process that they've gone through. The last conversation I had with him was interesting because having got that framework in place, that's right for them, it's then how do you communicate it across multiple countries, thousands of people, and effectively get that complete change on boarded and maybe enabling people to say, hey, this, this thing that I've spotted over here, the other side of the world, is something that's different. And maybe that's my introduction to the leader and the work that they did, which again, is one of the early pieces of work, as far as I was concerned, around CDR and led by Jakob Versner. And again, he put in place an explicit CDR. They had CDR principles that they defined for them. I think there were 14 separate areas that they identified. It was driven around digital transformation. So their starting point was every project that has a business case to progress anywhere across the business, we're going to interject, and this is the framework by which we'll analyze the consequential impact and look across the supply chain, very focused around sustainability and local community impact in terms of the societies, the farms, kind of in whichever countries, in whichever parts of the supply chain. And one of the things that I loved about that was that whistleblowing aspect. It was saying, look, there is no problem with being able to say, hey, I've spotted something we need to look at this. And I think that's such an important part of culture and maybe going back to the early conversation of transparency, giving people the confidence in an organization to feel that they have a voice. And I know we've put in this kind of marketing kind of brochure that we're net zero and compliant, but actually kind of, actually the proof points that we think we're quoting aren't quite right. I think it's important for people to be able to say that. So, I mean, there are many other organizations and much work that's been done, but I think both of those two examples are the ones that I'd give as well established on their journey using CDR and digital principles and continuing to evolve and build out over the last several years, rather than somebody just starting today.Aiste Rugeviciute 30:05If you are a smaller company, if you do not have 40,000 people and a lot of experts who can contribute with their knowledge and plan how CDR is relevant for you, that's a different challenge, right? For the smaller companies when they want to do something, but they don't know how to get going there. And currently, unfortunately, we are lacking tools or there's not that many frameworks which can provide. I emphasize a structured approach because let's be honest, people like to be structured. You know, they like to think in terms of some kind of systems and being clear about the steps which they need to do in order to get to the next stage.Gael Duez 30:54As you said, in a big company, you can indulge yourself to have a dedicated team working on this topic and elaborating fine tuned principles, thinking about how to deploy it, and bringing the leaders. But if you're a medium sized company, mostly it's a one person show most of the time and it's not even a full time job. So having a structured approach will definitely help. And this is exactly why I was very much interested in your maturity model. So the floor is entirely yours now to talk about it.Aiste Rugeviciute 31:32Yes. So CDR maturity models. So it's a proposition. It's not a normative thing, but it's a proposition which is a result of academic collaboration and work with different people, different organizations. And thank you, Gael and Rob, you were part of that for all the disclosure, as much as there were others around. In total, I would say 30 people who were involved in contributing with their knowledge, ideas, giving feedback at the different stages. So the maturity model, what I'm talking about, it's fully available online. There's an academic paper I still need to write, actually a more down to earth paper, practical explaining, because I understand that academic papers are not very easily readable by most people, especially those who want to go just quickly and apply it. But the idea of this CDR maturity model is actually very simple. It's a model which allows at the current stage, to look into the scope or help to define the scope at which an organization should look in order to analyze their CDR strategy or analyze their actions, currently, actions being taken inside of the organization. The model, as I said, currently is descriptive, so it's not very normative. It does have six maturity levels. So the idea is that you go from nothing at all. There are no actions being taken up to the six, which if you're exceeding, you are the innovator. You are at the front of this whole CDR movement. There are five dimensions, and within five dimensions we have 18 sub dimensions or 18 focus areas. So we have CDR governance that kind of stands itself clear and underlying there are three sub dimensions. And to be honest, CDR governance, it was one of the most easier dimensions to get consensus on. Almost everybody, all the participants, you know, they were, we do need that. We need actions, rules against the CDR, governments and strategy and all the planning. So that's the first dimension. The second dimension is the ecosystems, which means that as an organization, you should be doing something to contribute to the wider discussions of CDR or in general, digital sustainability. It could be through your relationships with your suppliers, it could be through participation in different external think tanks, groups, open source projects, etcetera. Then we have a workplace and culture. So that digitalization is implying, well, very rarely we will have an organization currently without technology in place, without having their collaborators being exposed to technology, and collaborators can be empowered with the technology. So one of the examples, you know, generative AI, it has a lot of potential to empower people, but it also comes with risks. So that's the whole coming back to the principles that technology can bring positive and negative impact. So I have a workplace and culture, you know, at what point people recognize the risks and opportunities and do they know how to act upon them? Then the fourth one is digital services or solutions. So it's all about dev teams. So internally, what kind of applications we're using, how we're developing, what are the processes in place, if it's relevant for a given organization. And then the fifth one, which is the most difficult, I would say, and this one still needs to be, I think, further developed, is infrastructure or IT assets. So everybody agrees that we do need to think about how we are dealing with data centers. What about our equipment? So IT assets are very much thinking about concrete hardware. But, and that's where the more complicated discussions. Cloud, for example, where does that fit in? Is it fitting in your infrastructure? Is it a more digital solution? Or is it an ecosystem? Because it's actually your provider? Right? So currently it's put in the model. It's under infrastructure, but the model is supposed to be a living model. So the model is there helping, hopefully helping practitioners, helping academics to structure the conversations about CDR. What's the scope? How do we evaluate the current state of our actions, in what areas? So they can be called capability areas. But we don't really give in to that model. We don't really give advice on how to reach the next level. So that would be a more normative model. So I hope that somebody will take it, you know, and as I said, it's a living model. It's supposed to be criticized, it's supposed to be, you know, people should hopefully engage with that model, propose next steps, evolutions, how we can apply it in order to plan the strategies more effectively.Gael Duez 37:45Okay, thanks a lot. And I must admit that I'm happy that the cloud was kept in infrastructure because I was a strong advocate of it and I had the opportunity to use it with a client. And actually I was kind of exactly what you've described. It's not necessarily super normative, but it was at the very early stage of consulting on green IT strategy. That was how it was labeled. And I remember I needed something to provide them structured feedback on the, I think it was 20 or 30 interviews I ran with their key actors. And it was pretty useful just to drop the visual representation and say, you know, as far as I understood with those discussions, here is where the main impacts are today. So it makes total sense because you're building digital services to pay attention here and here and that, etcetera. And it was a very loose assessment. It was not super structured with the metrics, etcetera. But still, it went straight to the point with my clients and the executives, like, okay, we got the big picture and then let's investigate a bit more on it. Which actually leads me to another aspect that both of you, you've been advocating quite a lot, is talking versus acting. I think it's very easy to get lost with a high level of conceptualization. And the CDR approach can help actually to get lost somehow if you deep dive into the details and you say, okay, and we should structure like this, and then create a substructure and then another substructure, and for each substructure we should get metrics and goals, etc, etcetera. And like ten years later, you've got a beautiful paper, maybe a white paper or an academic paper, but not that much action being done. And I know that both of you have got strong opinions about it. Maybe, Rob, you wanted to start, like, how do you use the CDR principles and the CDR approach to literally drive actions rather than discussion?Rob Price 39:48I think it's fascinating to actually kind of look at what is different, what is happening around the world and to understand relatively kind of the way in which change happens. I mean, the thing that I would say is organizations love to measure things. I mean, can they love to measure things from a financial point of view. But actually just over the last 20 years, I think many organizations, unless you can measure it, you can't possibly then do anything to change it. So I think it's really important to have some sort of model that enables organizations to assess where they might be. I just don't think that the answer is that everything needs to be 100. It comes back to choosing where to make the impact, because you've seen an assessment and a chart that suggests that there's weaknesses in certain areas and then for what you do about them, if I can talk briefly about the approach that I see in Switzerland. So Ethos, who were kind of involved in one of the early definitions back in 2020, I mean, they started measuring the top hundred or whatever number it is of Swiss businesses on a CDR index that they've defined. And of course, the first time they did it, everyone scored appallingly because it was a new concept. It wasn't the way that they'd been managing their business, if you like, but when somebody with a voice, and in the Ethos case, because of their involvement in pension funds and they've got a financial voice, if you like, in Switzerland, then organizations then think, well, we don't want to be scoring so badly next time, so how can we do some things to ensure that the score is better? So the first point is, measurement's key because that's what organizations get. Understanding the influence for change, whether it's government regulation, whether it's consumer behavior, whether it's somebody who's got an involvement in the investment decisions, in your long term sustainability as an organization, Ethos, as an example, there are different patterns as to why people choose to change, but then the question is, what do you do? And the answer in terms of what you do is in many cases do the things that are in the discrete area. I mean, if we're talking about the technology in a data center or in cloud CDR isn't inventing some different path that changes what you do? I mean, that's straightforward digital transformation. It's just trying to shine a light on is it being done in the right way? And I'm sure if, I mean, Gael, I remember Mark's session at the conference last year where he was talking about what is your true energy source? So it's about shining a light on you might have thought you were doing things in a way, but had you considered X, Y and Z? Because they're negative, consequential impacts that you just weren't thinking about, because you were just looking at it through the lens of head of IT, or head of compliance or GDPR, legal perspective, whichever one it is. But in each of those areas, there are things that can be done. So I almost see CDR as a pane of glass that you put across your organization that enables you to think about all that in connectivity between the things that are being done. Because what I see in organizations is people don't do that. People focus on the thing that they are responsible for, they know about, their passion is about, rather than that holistic view.Aiste Rugeviciute 43:35And I think that holistic view is also coming to the obsession with the measurement or the lack of holistic view is obsession with a measurement. Well, you gave a lot of different cultural examples, Rob. I come from France, or my current experience is very much in France, and here the measurement is very much obsessed, being a thing of people involved in IT. And I think it's almost a characteristic trait, you know, oh, we can measure how much cpu usage is this, we can measure how much this code is more effective than the other code. And all that measurement ends up being some kind of performance indicator, right? We can express the measurement in terms of money, we can express measurement in terms of energy consumption, in terms of maybe how many people we trained on CDR issues. But what actual value do those performances generate? That's the question which is often missing from the discussion. And I think that's exactly when you talk about the holistic view, we often forget that measurement or performance is not equal to the value. And what exactly we're talking to whom, you know, the value. So the value to the collaborators, the value to the future generations, value to your clients or your shareholders is not necessarily the same either.Rob Price 45:11I had a conversation a couple of years ago with an initiative that had put a framework and measurement in place. It wasn't a CDR, but it was very close to it. And I asked the question of how it is actually measured then? And the answer was a four month audit. And an organization is not going to do that as in what we're trying to do here. We're trying to impact the planet. We want every business, every organization, every government to be conscious about the way in which they use digital technologies. It can't be complex for every organization if we want everybody to do it, to do it. So how can we use it? I mean, I'm going to go back, maybe I am wearing my technologist hat here at this point, but finding ways to automate the measurement of those things that we're looking to assess. So that something can be done almost immediately with confidence as opposed to kind of misreporting, I think is critical to any of these succeeding.Gael Duez 46:20And is it also a question of accepting proxy metrics instead of final metrics?Aiste Rugeviciute 46:27This is the key, you know, to what granularity you need to measure things, right? Because you can get so obsessed and so you know what you're going to do with 400 KPI's. At the end of the day, you can measure very precisely how, how much of the resources your code uses. But if you do know that your code is creating, I don't know, negative impacts or creating addiction on a larger scale, well, all those metrics, they're not going to help you to actually make an impact or achieve something different, right? So I think sometimes we forget that experts or people working in the company can very quickly and for a pretty low price, I would say evaluate or estimate the impacts themselves. You know, they can. You just, you do need to provide some certain framework, you do need some certain performance indicators because otherwise people love it, you know, and that's how they measure the success of their company. So it could be a proxy used as money is often used as a proxy for any kind of indicators. But in order to act, well, you need to know priorities, you need to know the magnitude of the problem, but you don't need to know the exact numbers. If you know that you're using thousands of computers, but you don't develop or, you know, your website is just very rarely used, but clearly that the impact is more of your hardware, of the usage of the computers, instead of concentrating on eco design of your website. If it's not used by anybody or if it doesn't bring a lot of value, neither to your collaborators nor to your clients. And you don't need a very precise indicator for these to take action in this respect.Gael Duez 48:30That's why I love using the OKR framework because I think it sets two different levels and I use it almost all the time in sustainability consulting. Remember, what are the goals? Like the kind of medium term aspirational goals and this is really what you want to do. And then, yeah, well, you've got key results, but you can try to track them over a longer period to see if you're aiming in the right direction. But first have several of them, because unambitious goals usually should be measured in several different ways. And not, please do not use a north star, that is the worst possible thing to do, like having just one metric to understand the world, how stupid this is. And then and then also acknowledge, exactly as you did Aiste that, yeah, at some point we might drop a metric to keep another one and not end up with 20 or 30 different key results or KPI, to try to understand actually the real true goal that we have. We could be reducing carbons, we could be reducing water consumption, we could be having a positive impact on some kind of very significant social issues, etcetera, etcetera. But yeah, the question of measuring, and I know that it's especially true in France, but to be honest, it's quite also a very strong cultural trait in the US as well. And they love measurement. Just watch a basketball game and you will have statistics on everything. I'd like to drift a bit away from the implementation itself of CDR and to put on the table the topic of innovation. Because if we go back to the definitions that both of you gave at the very beginning, it was really the definition of CDR, I mean, it was really this idea of trade off that we need to know how much we impact the world, to know how much good net we bring to the world, and not just saying we're saving the world. So let us build ten data centers per week. But that's connected also with the fact that most of the time, with digitalization comes some sort of innovation. So what would be your stance on innovation and sustainability? Are they opposite most of the time opposite or are they complementary or is it a bit more complex?Rob Price 51:03I think something that a number of us were involved in, which was the Codes Initiative and the 22 report global action plan for a sustainable planet in the digital age. And the very simple part of that was it talked about three shifts that were important. I think the thing that was so important to me in the code conversations was that there was such synergy between everything that people were talking about there and CDR. It wasn't called CDR. I mean, CDR is in the report, but actually the conversations were the same and the three shifts put the enablers in place. So collect the data, collect the information, understand the impacts that we're going to be having. Protecting. Second, protect from harms, mitigate negative effects. Third one was to innovate to solve the world's biggest problems. In essence, those were the three things. Now, why have I said that, organizations, in my experience, don't like being told not to do something, and when they are told not to do something, then they don't want to spend any money on not doing that thing or to do something in a way that they're supposed to, as opposed to the way that they do do. And it comes back to if your model is based on addiction to maximize advertising revenue, because that's your entire kind of way in which you operate and live, then you're not going to suddenly stop that, are you? I mean, it's just not going to be there. So I think one of the things I hear a lot at the moment, and I'm firmly at the heart of the generative AI world and using those technologies is people focused around obsessing about quite recently harms, native facts. It is just doing things not very helpful or impacting people's jobs, etcetera, but forgetting about, because they obsess about that, forgetting about the potential power of those technologies, generative AI, AI machine learning, data science, etc, to solve the world's biggest problem. And we need that. The world has lots of problems at the moment. I mean, it's not just climate change, but finding a way to, I mean, just aging demographics is going to be a problem in terms of long term sustainability of organizations. I mean, sustainability in the largest sense. People can't get access to the people they need, etcetera. So we have to innovate. We need to innovate with a very conscious focus around minimizing negative effects and being aware of harms and all those things. But let's try and focus on positive impact. And I know from experience, impact measuring impact on people and communities is notoriously difficult to do. But let's be conscious of those things and absolutely focus on innovating because we have to. But with that consciousness about doing it in the right way, consistent with regulations, consistent in terms of, or thinking about the impact on people. But I am very firmly, I mean, you mentioned, I think at one point, Gael, that you've seen maybe an evolution in my thinking over the last two years, and I have, I am firmly in the, we have to use these technologies to innovate to solve some of the biggest problems that we've got, but with a very conscious awareness of some of the harms that could be caused. And yes, to think about how we kind of mitigate those.Aiste Rugeviciute 54:40I think this is the biggest challenge, how to innovate thinking about the system, about the global system, about the whole other potential impacts. Because if you look at the current strategist policy strategies or even within companies, right. All the strategies about sustainability, sustainable development, right. They are made in one line and then digitalization strategies are made in the other line and they're not really talking with each other. If you look, honestly, if even, I think the newest UNEP report on digitalization, they all talk how digital technology is going to empower and allow to achieve the digital development, sustainable development goals for sustainable development, you know, reduce some problems and so on, but they do not really talk about the potential negative effects that come with a digitalization. So currently in a lot of discourse, and that's what was really underlined in a digital reset paper report, which was published, I think, at last, in 2022. The policies do not talk with each other, and we need them to talk with each other. We need to ensure that digital transformations are aligned with sustainable development goals of sustainable development trajectories. And one of the things they are talking about in their report is exactly that. System innovations. That's what Rob was mentioning. So, innovations, first of all, I don't think that innovation should replace the word progress, which currently often is used instead of the progress we are using innovation. We used to talk a lot about advancing some progress and so on within the companies, and now it somehow got replaced by innovation, which is not exactly the same thing. Right? Whereas innovation, if we're using it as a definition, in order to come up with something new, in order to respond to the global challenges and global crisis, and we have a lot of those crisis currently. And I do believe as well, that technology has a lot of power, but it does have a lot of risks, associated risks. And we need to develop that capacity on different levels, on politics, on corporations, on individual levels, recognizing that technology is not neutral. We've been told, you know, oh, the technology is not neutral. No, it's not. Technology comes by itself with a negative environmental impact. I think one of the good examples is that because of the big data, because of the development in the algorithms, we have way more accelerated our knowledge of climate change. We actually can model things. We can understand it better because of the advances in it. So that's a positive thing. But now we know the problem is there. Now we understand very clearly where we are going. So it's time to shift and to use it differently and not to continue running huge calculations on something which we're not able or we don't have capacities to act upon. So I think that it's a very difficult change in mindset, because we don't like to think on a global scale. Right. It's difficult, I think, for us to think about positive and negative impacts at the same time.Rob Price 58:3920 25 years, nearly ago, I did some research around disruptive innovation, and I seem to remember a definition from that time that innovation was invention through commercialization. So innovation was nothing unless you found a way to make money from the thing that was being invented. And maybe we need to add with purpose or with an eye on doing it the right way to that definition. So it's not just invention through commercialization, but it's invention through commercialization with purpose.Gael Duez 59:13That's a beautiful definition. Well, thanks a lot, both of you. That was a very, very interesting discussion, especially at the end around innovation. I think we could have done another episode entirely on this topic, but I need to let you go to your regular life. So thanks a lot for joining. Thanks a lot for highlighting this very multiple aspect of a CDR. So thanks a lot and hope to see you in London in a few weeks.Aiste Rugeviciute 59:45Thank you.Rob Price 59:46Great. Thanks.Gael Duez 59:49Thank you for listening to this Green IO episode. If you enjoyed it, please share it and give us five stars on Apple or Spotify. We are an independent media relying solely on you to get more listeners. In our next episode we will talk about SEO, Search Engine Optimization and Sustainability. Does it sound like a long stretch? But I discovered Jean-Christophe Chouinard’s work 2 years ago and more recently the Green SEO meetup - a fringe event of the huge global convention BrightonSEO which seems to be to the SEO industry what Cannes festival is to the movie industry. It must be something related to the seashore… Anyway, I jumped on the occasion and invited 2 pillars of this Green SEO community Stuart Davies and Nathalie Arney. And I got great insights on how SEO practitioners should incorporate sustainability in their day to day job. Green IO is a podcast and much more. So visit greenio.tech to subscribe to our free monthly newsletter and share the conferences we organize across the globe. The next one is this Thursday, September 19th in London. The 250 available tickets are sold out. Well, except for late green IO listeners who can still get a few remaining free ones using the voucher GREENIOVIP.Roxanne 1:01:20Build a greener digital world one bite at your time.❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.  📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents. 
undefined
Sep 3, 2024 • 56min

#44 - Can the data center industry become circular? with Deborah Andrews

💭Fueled by the ongoing artificial intelligence boom, data centers are popping up around the world like mushrooms after a good rain raising serious sustainability concerns. Hence a pressing question: can the data center industry become circular? 🎙️To get some answers, Gaël Duez welcomes a veteran in Circular Economy and Life Cycle Sustainability Assessment, Prof. Deborah Andrews, from London South Bank University, the founder and academic lead for CEDaCI. Some Takeaways:💡 how the CEDaCI Compass tool can help data center being equipped more sustainably,♻️ current inadequacies in recycling infrastructure for electronic waste, and⚡ concerns about the rapid development of AI and its energy demands❤️ Subscribe, follow, like, ... stay connected the way you want to never miss our episode, every two Tuesday!📧 Once a month, you get carefully curated news on digital sustainability packed with exclusive Green IO contents, subscribe to the Green IO newsletter here.  📣 Green IO London is on September 19th 2024 --> use the voucher GREENIOVIP to get a free ticket! Learn more about our guest and connect: Deborah's LinkedInGreen IO website Gaël's website 📧 You can also send us an email at contact@greenio.com to share your feedback and suggest future guests or topics.   Deborah's sources and other references mentioned in this episode:CEDaCI Project CEDaCI CompassInvestment in clean energy this year is set to be twice the amount going to fossil fuelsDeforestation-free supply chainsLife Cycle Assessment (LCA) – Everything you need to knowTantalum is mined in Central Africa, mainly in DRCInterreg NWEEuropean Regional Development FundTranscriptIntro 00:00There's a real need to shift thinking and business models and possibly, rather than selling it, equipment actually to sell services. So companies own equipment, so then they're responsible for maintenance and for what happens, either for extending life, which would be very much to their advantage, or for recycling, ending life.Gael Duez 00:37Hello everyone, welcome to Green IO with Gael Duez - that’s me! In this podcast we empower responsible technologists to build a greener digital world, one byte at a time. Twice a month on a Tuesday, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the Tech sector and beyond, to boost Digital Sustainability.And because accessible and transparent information is in the DNA of Green IO, all the references mentioned in this episode, as well as the transcript, will be in the show notes. You can find these notes on your favorite podcast platform, and, of course, on our website greenio.tech.Can the data center industry become circular? Simple question, very complex answers.Fueled by the ongoing artificial intelligence boom, data centers are popping up around the world like mushrooms after a good rain. For an industry which is more and more under scrutiny due to its environmental footprint, the sustainability angle cannot be overlooked anymore. Still, the main hurdle remains data, as many guests already stated in this podcast. Hence, my wish today to get insights from the program leader of an initiative which has managed to build high quality primary data on the data center industry. The CEDaCI project sounds like a Dan Brown book title, CEDaCI code to unveil all the mysterious power beneath our almighty data center industry. And to some extent, it's quite the plot. But I will let Deborah Andrews, Professor at London South Bank University and a two decades long veteran in Circular Economy and Life Cycle Sustainability Assessment reveal all of it. Hi, Deborah. Thanks a lot for joining Green IO today.Deborah Andrews 02:32Thank you very much for inviting me.Gael Duez 02:35It's a pleasure. It's a pleasure. And I have a very direct question to ask you to kick start our discussion. Did you start the CEDaCI project out of frustration somehow?Deborah Andrews 02:49Okay, well, first of all, thank you for calling CEDaCI. That's the Italian interpretation. We normally describe it as CEDaCI, but hey ho, it's an acronym for the Circular Economy for the Data Center Industry. And yes, there was an element of frustration. I had worked with researchers and operators and so forth in the data center industry on a number of research projects from about 2010, 2012, and was acutely aware that the sector was very fragmented. There was the most phenomenal amount of expertise in the sector, but people worked in silos and didn't connect with each other. And consequently, there wasn't a sort of whole systems approach to the challenge of sustainability. And the industry experts in each of the sectors were doing the best that they could for their particular sub sector, but there was no consideration of the impact that those actions had in their sector. What the impact on other parts of the industry were? So it was absolutely apparent that there was a need for a whole systems approach. And it was very timely in, you know, having spoken to people who subsequently became partners in the project, that they were also acutely aware of this challenge. But being in academia, I was very lucky, being sort of slightly outside the industry, to be able to bring various representatives, stakeholders, etcetera, from the sub sectors together without having any bias.Gael Duez 04:38And Deborah just could you illustrate, maybe with a few examples, what silos are you referring to?Deborah Andrews 04:45So, one of the first things that we did in the project was to carry out a very critical appraisal of the state of the art. In other words, we did a scoping review of what was going on in the industry, and we identified eleven key sub sectors, key players within the industry, starting with suppliers and then going through design, manufacturing, etcetera. But what we did was linked the silos, or found evidence that these silos actually were related to all the different lifecycle stages of data center equipment. Now, we focused on electrical and electronic equipment, because based on prior studies with a very extensive PhD that we ran in conjunction with HP, we found that the hotspot, if you like, the environmental hotspot, and this was looking at a whole data center. So the building or the services, M&E, etcetera. The key area of environmental impact was IT equipment, partly because of the embodied materials, the energy consumption, but also the short life of products, and in particular of servers. We found that they have ordered center equipment, and this was also reflected in a big EU report that informed lot nine that servers had the highest environmental impact. So that was the focus of the CEDaCI project. But coming back to the life cycle stages and so forth, and the various silos, then we have the installation phase. And use, of course, is incredibly important. Operational energy transport, taking stuff to and from data centers, perhaps taking to secondary market operators or recycling plants. Then we have data destruction, which could be through mechanical means, shredding, etcetera, or it could be more digital with software. And then ultimately we move on to the end of life. And we could say end of first Life, which leads to secondary market and reuse. This includes refurbishment and remanufacture. Ultimately, though, whether you send your equipment to secondary market suppliers or straight on for recycling, eventually all equipment ends up with end of life processes. Now, ideally as much of the product should be recycled as possible. But what tends to happen is that the low hanging fruit, things like the casings and so forth, as their steel and obviously external, those recycled and the majority of the PCBs aren't, they end up in landfill. There's a real need for a shift in thinking practice to manage end of life equipment. In this CEDaCI project, one of the key things, one of our USPs, was to bring together representatives, stakeholders from these various subsectors. But one of the points that came up, because we organized co-creation workshops as part of the project, to identify what stakeholders felt they wanted, whether what they wanted was in line with what we felt would benefit them, which was a tool to aid decision making about to help with the sustainability profiling companies and so forth. But one of the USPs brought together these people, and then invariably they commented how much they had learned through the co-creation workshops, because they didn't talk to people who worked in other sub sectors. So this, again, you know, from a sort of academic or life cycle thinking perspective, you need an absolutely whole systems approach to the challenge of whatever industry you're working in to develop a circular economy. Because every action at every life cycle stage, etcetera, has caused an effect. It has an impact on all other life cycle stages.Gael Duez 09:27Okay, so there are so many things to unpack here from systemic thinking and your systemic approach, like the benefits of co-creation, and also what you've mentioned on e-waste. And if you, if you indulge me, I'd like to deep dive about this end of life differentiation that you just made. The first end of life, and then the second and eventually the final end of life, which is when the electronic equipment becomes e-waste. It's very important because I see today in sustainability criteria applied for RFP or bid, etcetera, or auditing the data center industry, or auditing hyperscalers, that more and more they kind of tick the box of all servers are donated to charity. All servers go to the second hand market. And so we're good, job done. I think it's a bit more complicated than that, because if the server is donated, but one year later, it ends up in a landfill somewhere in Africa, Southeast Asia, South America, or even Europe, that's not really a solution to the problem. So my question is, how important is this distinction between the first end of life and second end of life? And is it today enough taken into consideration on how to assess the sustainability of the data center industry?Deborah Andrews 11:02Okay, so let's think about equipment, and in particular, servers, which I've said have the highest embodied impact of data or data center equipment. And the technology, obviously, whether it's memory, processing speed, whatever, is developing incredibly rapidly. We won't even mention AI at the minute, but the sub-technical change means that an awful lot of equipment is replaced while it is still working. You know, it's functioning very, very well. Can understand, in a way, why hyperscalers and other bodies may want to have the latest, fastest, best memory, etc, equipment. But what on earth happens to the stuff that comes out of hyperscale and other centers when it's still working? Why would you want to recycle anything, take something out of service, recycle it when it still functions very well? So there's obviously a very key concern is memory and disks, be they hard drives or solid state drives. What happens to the data? We'll come back to that in a minute. But there are initiatives to encourage use of second life products, products that are still very serviceable. I mean, that's, I suppose, quite a nice analogy, in a way is, and let's assume they're all electric vehicles rather than fossil fuel driven. But, you know, maybe the hyperscalers want Ferraris, when actually many other industries would… something like Fiat Cinquecento or VW Polo will meet their transport needs. Okay, so there are drives. And certainly, I know in the UK, government and organizations like the NHS are really encouraged to use secondary market products, which one of our partners, Techbuyer, a company that is linked to them, interactive, they did a massive amount of research for their own business, but also the research fed into the CEDaCI project, they found that there were a lot of myths around the performance, particularly to do with operational energy, the performance of old, within three years old, and brand new products. If the equipment is set up correctly, then the difference in energy consumption is negligible. Obviously, it depends what compute activities you're engaged in, but like for like, energy consumption, the level is negligible. The OEMs, of course, want everybody to buy new equipment, so they're always saying, “Oh, it's better, you know, it's faster, it's more efficient, blah, blah, blah.” So coming back to the car analogy, an awful lot of public bodies in the UK, and probably in the EU as well, are being encouraged to use second life products which meet their technical needs without any problem whatsoever. But the good thing about this is the secondary market, we should have the products there, and this is a key element within the circular economy as well, of course, because it's not just about recycling at the end of life, it's about extending product life. As I said previously, why would you want to take a product out of service that is still functioning and can fulfill somebody's requirements? I think the key statement around reuse is keeping something in service for as long as technically and economically viable. That's really, really important. So the secondary market is various companies have been set up to collaborate with big hyperscale operators or smaller operators and are promoting good practice, I think, in terms of circularity and economics and resource efficiency, there are some challenges there to do with legislation and ensuring that second life products have the same warranties and so forth, and also to do with data security. Ideally, we should simply wipe drives, be they solid state or hard disk whatever, and reuse them. But there's a lot of anxiety about data security. So organizations like banks, for example, insist that drives are shredded on premise. They don't leave the bank once they're there, they come in as new products and they end up leaving as bags of tiny little bits of metal. They're shredded on site to ensure data security. But I think, and again, there are a number of really good research projects ongoing looking at data security and performance of either software wiping, whatever, so that hardware, HDDs and SDDs, SSDs can be reused.Gael Duez 16:25Are you optimistic about the fact that even for highly secured working environments like banks or even the military, software wiping will prevail at some point?Deborah Andrews 16:38I'd like to think so. I mean, I'm not an expert in this, but I would like to think so. And I think the more really robust and empirical research that can be carried out to reassure end users that software wiping or equivalent is safe as shredding, the better, really. So, yeah, fingers crossed.Gael Duez 17:05And Deborah, you were mentioning that there are a number of companies that have been set up to meet this new market. Could you share with us some trends? Is it still marginal, or are we witnessing a boom in the second hand market of professional IT equipment?Deborah Andrews 17:23It's really interesting. I think it depends very much on time, place, circumstance, etcetera. So it was interesting. During COVID for example, there was a marked increase in demand for second life products, partly because of the increase in demand for data center industry services due to homeschooling and home working, for example, but also because there were supply chain issues. And if you remember, there were all sorts of problems with chips, etcetera. So once we got over Covid and the industry really kicked off again, back to normal business as usual practices. I'll give you a little anecdote here. Okay. There was a bit of an issue around the secondary market because some manufacturers had lots of new equipment in stock before COVID but they hadn't, for various reasons sold it on.So when there was a sort of shift in business practice generally around the world, a lot of the big manufacturers suddenly decided to flood the market with new equipment. And the price difference between that and reused equipment was negligible. So of course people wanted all purchasers, procurement teams wanted new equipment. So that had a really adverse impact on the secondary market for some companies. I'm not saying all, and it ist anecdotal, but I think it looks as though new and secondary markets are very subject to influence from external factors. You know, they're not as consistent, as stable as certainly the secondary market as we would like it to be.Gael Duez 19:28Did you explore with the CEDaCI project, the potential lifespan of servers if the second hand market were to be a very well functioning market? Because I've seen some colocation service providers starting to claim, and I congratulate them for this, that they keep their server for seven, sometimes eight, sometimes even nine years. But can we do better? Can we imagine a world where a server would last for 20 or 30 years? Or would it make any sense?Deborah Andrews 20:04First of all, if your server, and it's like, you know, your car, your Polo, your Fiat Cinquecento, whatever, even your Ferrari, if it's working well, then you should be able to keep it in service, in life or operation for a long time. The big challenge is when things, parts start to wear out or break, for whatever reason, they fail, you need to replace them. And a lot of manufacturers don't keep parts. In fact, I think legislation at the moment stipulates that manufacturers only have to supply parts for up to eight years. So if you have a ten year old server, where are you going to get the parts? It could be that you go to, say, a secondary market supplier and they have parts, but it becomes increasingly difficult over time to replace components. So that's a big challenge. The other thing is looking at changes in compute capability. We know that, for instance, there's all sorts of issues around whether you go for air cooling, to go for liquid cooling, is liquid cooling more efficient, etcetera. And AI increases the operating temperature of components. So that may well mean that we have to redesign service to manage factors like that. As compute changes, we actually need to reconfigure either layout, or if you are going for air cooling, then the number and type of fans that you have or you switch to liquid, whatever is most appropriate. So in theory, if we had modular servers that you could take out certain components of and replace with upgraded components that had common connectors and, you know, the box, the space required to house them was the same, then it seems you should be able to keep servers. You know, even if you end up with everything, all the internals, the guts of the server being replaced, the chassis, you should in theory, be able to keep reusing that for perpetuity. But whether that's feasible or not, because of changes in shift from expert liquid cooling, whatever is, we need to think about that.Gael Duez 22:40And before talking about the main results of the CEDaCI project, and as you mentioned, what all the stakeholders learned from each other, I've got one final clarification question. Because you were very assertive that the embodied environmental footprint within equipment is by far the biggest share of the overall environmental footprint of a data center. And if this position is not a debate at all when it comes to end user equipment or devices at home, whether it's laptop, smartphone, etcetera, etcetera, I heard some different opinions where it's more like 50-50, because professional equipment lasts longer. Some people advocate that actually the use phase, and especially the energy consumption during the use phase is far from being negligible. And for some of them, it's even the majority of the environmental footprint. So maybe could you clarify whether you were mentioning only the carbon or other environmental impacts, or even maybe the carbon. Your calculations make it clear that embodied carbon is even bigger than energy consumption, GHG emission. I was looking for a bit of a clarification here where you stand on this debate.Deborah Andrews 23:59Okay. So when we said that in our calculations that the largest impact in the data center was the IT equipment, when we said that the largest impact was the IT equipment, we based our model on a data center that lasted for 60 years. The actual infrastructure, the building, some of the M&E, was replaced after 20 years, etcetera. But the IT product life was based on three to five years, okay? And that's for hyperscalers, that's a long, long time. Although Google now says that they're going to keep servers in life for five years, we'll see about that. So coming back to... So i if you decide to pull your data center down after 20 years and build a new one, then that ratio of impact from IT to building will change. But just be clear about that, okay. The second thing to be really, really clear about is the kind of metrics you use when you are looking at impact. So carbon, obviously, and carbon equivalent, it could be methane or other hydrocarbons. Carbon is only one metric, carbon and carbon equivalents. And you exclude a whole array in fact, thousands of other impacts and outputs, inputs, etc. And impacts, when you only look at carbon and carbon equivalent. So you're not thinking about the impact of water, the impact of gold mining, for example, which is incredibly toxic, the tailings can be, may have one of the highest environmental impacts, etcetera. And the impact is not just environmental impact, it can be very detrimental to ecosystems. So toxic substances, mercury, arsenic, etcetera, used in mining processes will obviously have an impact on people living in the area if they get into the water supply and into soil and so forth. So if you're looking at carbon, you exclude all of those factors. I'm not 100% against carbon assessment. I just think we have to think about it in relation to everything else. It was really, it was the original sort of metric linked to life cycle assessment dating from the 1960s. So when LCA first began, it just considered energy, be it operational or embodied. And that's where the linked carbon assessment comes from. It's incredibly important when we're thinking about climate change. We can't underestimate its significance. However, we do need to think about all the other impacts as well. So we carried out some studies of exactly the same piece of equipment, one looking at carbon, operational and embodied carbon, and the other looking at a comprehensive life cycle assessment, which looked at and included thousands of inputs and outputs. But just on a carbon study, we found that operational energy, when you're looking at lifespan, let's say five years, operational energy accounted for 85% of impact, or the carbon in the operational energy, 85% of impact, and embodied in energy was only 15. Okay, so with the newer equipment, that ratio shifted a bit and the sort of makeup shifted to 20% for embodied impact, and 80% for operational impact over a five year life. So it's no wonder that when the data center industry is being guided by, you know, the need to reduce carbon, be it embodied, operational, it's no wonder that you focus, the industry focused on operational energy initially, and also I think it's easier to manage to make change. So now we see, you know, more use of renewables, etcetera. And also improved operating efficiency of equipment itself. When we look at comprehensive life cycle assessment, we've got a couple of surprises. We didn't think that the difference between carbon and full LCA would be so significant. So the same piece of equipment or two pieces of equipment. The first was, as I said, a really old server, and the ratio of operational to embodied impact was about 80:20 for carbon assessment, and again, it was lower. It was, I can't remember something like 75:25. But what, this is the really, really, really key finding. When we looked at the comparatively new piece of equipment, which was from about 2017, we found that the embodied and operational impacts were about the same. It was about a 50:50 split. So that was really, that means, that was a real surprise. And it really highlights the necessity of examining, measuring, monitoring, you know, improving the physical resource efficiency, increasing use of recycled materials, increasing recycling processes, and building a decent infrastructure to do that, changing practice with things like encouraging, if you can't reuse a whole piece of equipment, can we harvest components and reuse a individual components, etcetera. So that was a very revelationary, very revealing study.Gael Duez 30:15This shift from 80:20 to 50:50 was mostly due to energy savings, energy optimization of newer equipment, or was it also, as you said, because of lower cost of building equipment, thanks to the use of recycling materials, etcetera, etcetera.Deborah Andrews 30:36Okay, so this 50:50 split, it was a lot to do with improved operational energy. When you're looking at the full life cycle assessment, you are considering things like what happens during mining processes. It's not just the energy or the toxicity, etcetera, and also things like what percentage. Well, now we should be thinking about the percentage of recycled materials included, that, in theory, should reduce environmental impact.Gael Duez 31:14So now that we laid the ground for a better understanding of all this environmental footprint, etcetera, let's go back to the actual findings of the CEDaCI project, both for the stakeholders, but also for an average, I would say, data center operator. What are the findings he or she should be aware of? And what about this tool that you developed called the compass?Deborah Andrews 31:38The CEDaCI Compass, the circular data center compass. That was a key output from the CEDaCI project. And the majority of the research that we did underpinned the development of this digital tool. It's free to use. You can find it, access it via the project website at cedaci.org. The idea was that we wanted to help people working in the industry to make informed decisions about how to support their transition to sustainability and circularity. And one of the things that we are very keen to do well, apart from being absolutely objective, it's completely non judgmental. We don't offer any, you know, this is right, this is wrong. We just present the results. One of the things that we did as well was to separate out the three sorts of tenets or pillars of sustainability. So you put your information into the tool, or you select various criteria, and then when you see the results, comparing two different servers, you can compare the environmental impact, the social impact, the economic impact, so that, you know, it could be, as an operator, you're more concerned about social factors than environmental, so you base your decision on that. Or I would imagine most people think about economics as their priority driver. But, you know, being aware of the other criteria is really important. If you lump together, you know, these three key tenets of sustainability, you get slightly inaccurate results, you know, you don't know what I mean. It could be the social impact of one server is very high, but its environmental, adverse environmental impact is very low, or the other way around. So that's why we separated those points out. The other thing that we included was a criticality indicator. Now the EU and, well, UK to a certain extent, but certainly the EU has become increasingly conscious of resource efficiency and has identified, now, 30 materials that are what they define as critical. This is because of the amount of resource that is as yet unmined in the surface, the amount of material that is currently recycled. And I would like to include as well the possibility of substitution. But the other really, really important is the geopolitical factors, where on earth is the material located? Because that has a very significant impact on availability. So the critical raw materials are basically defined as those that are of major technical and economic significance to the EU and UK. So we included a criticality indicator to raise awareness of these materials. The use of these materials in data center equipment, all electronic equipment, uses some critical raw materials of some type. We cannot make, for example, mobile phones without tantalum, which is essential for capacitors. Tantalum is mined in Central Africa, mainly in DRC, Congo, where mining practices are eye wateringly horrific. They are environmentally and socially damaging. So we wanted to raise awareness of those sort of issues as well, to encourage better practice. So the compass was developed. As I said, it's a free to access online tool. And it was basically to inform potential end users about the impact of their choices. The other thing that was really, really important at the moment, we don't have anything like the scale of recycling infrastructure that we need to manage all the equipment that's currently in circulation, let alone all the electrical and electronic equipment that will come into the waste stream imminently. We've got a huge problem with this. There's a collection globally and it does depend where you are based globally and as does reuse. Incidentally, people with less expendable income tend to be more frugal by necessity rather than intent, I think or wish. Overall, the collection rate of e-waste is less than 20% globally. And that includes consumer as well as commercial products. But we don't have anything like the infrastructure that we need to recycle this anyway. The recycling processes at the moment, they focus on anything with iron in it. So steel, copper, aluminum, and gold. And there's not very much gold, you know, by mass, it's comparatively little gold in electronic equipment. But of course, its inherent economic value makes it attractive to recycle it. So there are masses and masses of materials, many of which are critical on the critical raw materials list, which aren't recycled. Unless we get our act together and develop a proper recycling infrastructure, there's significant potential for disruption to supply chains.Gael Duez 38:54But do you believe it is possible? My understanding of the chipmaking industry, or even slightly less complicated part of the IT industry, the design itself is so complicated, is so… the different metals are melted together to create alloy. You've got ceramic, etcetera, that I don't even know if it's feasible to recycle the way I would say the average John do understands it, which is we will extract to reuse it the same way. And it's more down cycling or it's even. I mean, I honestly wonder if recycling is really the way forward. So don't get me wrong, instead of being able to reuse for super long period of time, as you previously mentioned, having components that are interoperable on open standard etcetera, and saying, okay, you know, this memory card might be 20 years old, so it's ridiculous the amount of data you've got. But hey, I've still got half a billion of them. And if I've put them in some racks, it's still a decent, you know, decent enough, or whatever. But my point is, I think we are fighting an uphill battle if we really want to recycle, like extracting the tantalum you were mentioning, or the cobalt or whatever, rather than redesigning our industrial process and making also sure that the warranty period is so big that so long, that actually we shift the burden of recycling to producer, which are eventually, ultimately responsible for putting things on the market that are absolutely not recyclable. And that would end up being e-waste in a matter of years rather than decades. And when you see the environmental footprint of everything that you describe, we should talk in decades rather than in years. But that's a personal opinion. Sorry, but my point. What do you think about the feasibility of recycling? Or are we talking about a slightly different approach in the recycling industry for the IT equipment?Deborah Andrews 41:05I think we need to have all of those things, really. I don't think there's not one size fits all solution. I think there are some really massive challenges with electronic components, for example, because of the way that they function. You're looking at atomic levels, the way that if you're creating signals, etcetera, the way that atoms and the subatomic particles behave. So that is obviously going to limit the way in which components are designed and manufactured at the moment. Maybe in the future they will discover different approaches to data transfer or signaling, switching, whatever it may be. I don't know too much about quantum, and I don't know how this is going to change things, if at all, if we exchange one set of problems for another. I don't know, but I'd certainly like to find out an awful lot more about it. But coming back to your question about end of life and so forth, I absolutely agree that we need a different approach to the manufacture of many products. I think we need to really focus on the things that can be upgraded, swapped, repaired, etcetera. Again, most or many electronic components, it's nothing. You know, they're so tiny, it's impossible to repair them. So focus on the stuff that we can repair and upgrade and keep in service for as long as possible. The other stuff, we need to certainly develop better recycling capability. But there's an argument that's put forward that some of the materials, the economic value is fine if you have a kilo of stuff, but by mass per component, the mass in individual components and in servers as a whole is very low. So the economic value of any particular materials reclaimed from the server will be low. So we need critical mass to make development of recycling and reclamation technologies for particular materials economically viable. That's going to be the driver. The other big challenge we've got with recycling is sometimes the processes, and we're not. You know, there are lots of ways of doing this. One of the partners in this CEDaCI project, a company called TND or Terra Nova Developments, they developed some new recycling processes to reclaim materials that aren't commonly reclaimed. And they use a mix of thermal and chemical processes. And because you're using more than one process, of course, that increases cost. But, you know, if you have critical mass of stuff coming through the system, it does become economically viable eventually.Deborah Andrews 44:23One of the big challenges is very often, but if you have a printed circuit board with a huge number of materials, embodied materials, you're processing to reclaim one or two or three, those processes can damage other materials, and so you can't reclaim those. So, you know, I think at the moment, it's impossible to reclaim all of the materials in PCBs, which are the biggest challenge for recycling. So I think we need to think about when we're designing, not just designing for here now and in use to design, thinking about how can these things be easily disassembled to facilitate, if not, you know, chucking the material into a smelter to recycle it, but actually being able to reuse the materials in the… as soon as we take them off one product, we can put them on another.Gael Duez 45:22Do IT equipment manufacturers today start to tackle the issue? Or are they still mostly in the business as usual approach and not at all incorporating, as you said, a repairability aspect when the design sinks, even a bit of a recycling aspect?Deborah Andrews 45:45I think there's, and I'm not going to highlight particular companies, but I think there's quite a lot of smoke and mirrors and greenwashing. And this, again, it's anecdotal from personal experience of, you know, I worked through this CEDaCI project. They all still want to sell new equipment, and certain companies say that they have a kind of closed loop, but it's a very open closed loop, shall we say? They're not responsible. They do take equipment, but then they sell it onto secondary market agencies. They're no longer responsible for that equipment once they've sold it on. There's a real need to shift thinking and business models and possibly, rather than selling IT equipment actually to sell services. So, you know, you rent, I don't know whether you do it by compute capability or operating time or whatever. So then they're responsible for maintenance and for what happens, either for extending life, which would be very much to their advantage, or for recycling at the end of life.Gael Duez 47:00Which is what is slightly happening with the big three hyperscalers because they've started designing their own servers. And I could bet that they thought about the fact that even from a financial perspective, the longer you keep this equipment, which is on your own cost base, the better it is for your bottom line. And I have a final question for you, because you're privileged witness of the data center industry for almost a decade, two decades almost now, and there is a bit of an elephant in the room that you actually, you teased us several times during the interview about AI and more generally, I would say, about the trends. And I would like to share with you an anecdote while I was recording this very enlightening episode with Professor PS Lee in Singapore, who's one of the best experts in the water cooling techniques and more generally on building energy efficient data centers, especially in tropical climates. We had a really fruitful discussion, and he's a big advocate of technical optimization, and he knows a lot about these topics, etcetera. And at some point in our discussion, he posed, and he had this kind of overwhelmed moment, you know, when you've got just too much weight on your shoulders. He was like, but you know, Gael, at some point we will have to ask also the question of the level of consumption of compute that we want in our society. Because despite all the efforts that I'm doing to reduce energy consumption, the current trends, and especially the AI boom, I don't see how I can make it. And it was this kind of face, like, I simply don't know how I will be able to manage such an exponential consumption in energy, even if overnight all data centers in the world would switch to water cooling, super efficient water cooling techniques and whatnot. And it kind of struck me like even someone within the tech industry and such a strong advocate of technical optimization saying, wow, but the trend is really worrisome. Is it something that you're aligned on, or are you a bit more optimistic about the current trend in the data center industry?Deborah Andrews 49:29I think it's really scary. Forget about the ethics and what AI can and can't do. Forget about that. I think it's really scary because it seems that the industry is racing, racing forwards to develop either new data centers, new equipment to manage AI, etcetera. But my feeling is that some of the, what we've learned, good practice, etcetera, is forgotten in that race. So it's a business as usual approach. Build, install, run, replace, run. It's very worrying because the physical resources and energy required to manage AI operate, even the simplest operation, the demand for energy, is astronomical. So there are arguments saying, oh, well, AI is going to measure this, that and the other, and improve, you know, this, that and the other, but, or improve the operational efficiency of this, that and the other. But I don't think we've done any kind of calculations at all to see whether the benefits of running an AI software or, you know, program activity, whatever, to assess resource efficiency, whether the benefits achieved through the resource efficiency or more significant than the impact of running the AI operation. We need to really, and it's not going to happen, but it would be great to pause for a year and just to sort of examine some of these factors and to see where the benefits of AI really lie. So if we do carry on business as usual, and I know there are various regulations coming in from, you know, digital sustainability and so forth in the EU, but I don't think they're going to have a massive impact on, certainly the speed of development of AI is far faster than the implementation of these new regulations. But the other thing that's worrying, if there are constraints of operating in the EU, for example, does it simply mean that providers will go elsewhere, they'll build in locations where the regulations don't apply. And so we move our problems to another part of the world. In effect, you know, more buildings in Africa, Asia, South America.Gael Duez 52:05And from a cold financial perspective, do you think that the current trend of building data centers everywhere in increasing compute capacity is sustainable? Or do you foresee some bottleneck or even some forced pose because of resource exhaustion?Deborah Andrews 52:25There may be a quantity crunch, unless we think about extending product life, as we've already mentioned, and recycling, reclaiming more materials, changing business practice, business thinking. I think that there's potential for quantity crunch, but whether operators will start to charge for access to digital services. I mean, at the moment, you know, you buy your phone, you have a package, and you can contact anywhere in the world whenever you want. Okay, you pay for apps and so forth, but actually you're not really paying very much for the digital services that enable those apps to function. And a lot of things are free anyway. So will we have to pay for digital services? Is that one way of monitoring or constraining digital activity or not? That's one question. But the other thing is, is it ethical? You know, if we think we're in a luxurious position in Western Europe in terms of economics, although there are people in digital poverty, but, you know, generally, as a child or a university student, if you don't have computing equipment at home, you can go to your study institution and access digital technology there. But that's not the case in many populations in developing countries where digital tech is a luxury for the upper echelons of society. And yet we can see how access to a phone, not smartphone, just an ordinary old fashioned phone, has empowered women, for example, helping them to set up businesses and so forth. Is it right to charge them for data or do we make data charging… is it sort of socially stratified or according to income bracket? I don't know. I think we need to be a lot more visionary, look forwards and be proactive and anticipate problems and design them out before they happen. Now, whether that's… I think that's possible with some equipment, but whether it's possible for, you know, in terms of human behaviour and so forth, there's a whole other matter altogether.Gael Duez 54:18It makes total sense. Thanks a lot, Deborah, for joining. That was very enlightening and a unique perspective on the data center industry, environmental footprint and what could be done to reduce it. So thanks a lot. Once again, it was a pleasure to see you and hope to see you to Green IO London as well.Deborah Andrews 54:36Thank you very much. It's been an absolute pleasure. Thank you.Outro 54:42Thank you for listening to this Green IO episode. If you enjoyed it, please share it via email, on LinkedIn, on Facebook, on Twitter, if you are still there. We are an independent media and word of mouth is the only way to get more listeners. I don't ask you to rate it five stars on Spotify or Apple Podcasts because of course you already did it, didnt you?It's time for you to grab a book or enjoy a good article. And guess what? You can find many ideas in the latest edition of the Green IO monthly newsletter. And don't forget to book your ticket for the next Green IO conference in London on September 19. Good news for your holiday budget. You can still get a free ticket using the voucher GREENIOVIP. Just make sure to have one before they're all gone. I'm looking forward to meeting you back in six weeks to help you, fellow responsible technologists build a queener digital world.Roxanne 55:40One byte at a time.❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.  📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents. 
undefined
Jul 16, 2024 • 50min

#43 - Digital sustainability in a Tech behemoth: Japan with Trista Bridges and Paul Beddie

Trista Bridges and Paul Beddie discuss IT sustainability in Japan, highlighting stakeholder-oriented society, emerging startups, and regulatory roles. Insights on the country's journey towards net zero emissions and the challenges of transitioning to renewable energy and reducing plastic waste. Stay connected for more sustainability discussions and upcoming Green IO conferences.
undefined
Jul 2, 2024 • 56min

#42 - Decarbonizing Tech: 2 CTO share their paths with Ludi Akue and Owen Rogers

💭Decarbonizing tech stacks and platforms? This is the challenge many CTO face these days. 🎙️In our 42th episode, 2 CTOs with different backgrounds and operating in different lines of business cross share their experience in reducing the environmental footprint of their operations. Ludi Akue, former Lunii’s CTO, and Owen Rogers, .eco’s CTO meet Gael Duez to share their insights and reflect on each other journeys in sustainability.Some Takeaways:🌐 pro and con of migrating infrastructure to a lower carbon region,⚙️ how to leverage technical debt to boost decarbonization,💰 when does optimization pay for decarbonization,⚠️ the pitfall of creating a separate program to steer sustainability,and much more!❤️ Subscribe, follow, like, ... stay connected the way you want to never miss our episode, every two Tuesday!📧 Once a month, you get carefully curated news on digital sustainability packed with exclusive Green IO contents, subscribe to the Green IO newsletter here. 📣 Green IO London is on September 19th 2024 --> use the voucher GREENIOVIP to get a free ticket! Learn more about our guest and connect: Ludi's LinkedInOwen’s LinkedInGaël's website Green IO website 📧 You can also send us an email at contact@greenio.com to share your feedback and suggest future guests or topics.   Ludi and Owen's sources and other references mentioned in this episode:Putting our websites on a low carbon diet Lunii.ecoCarboSustainable Development GoalsUN Global CompactNot the End of the WorldLucie 26000Transcript Intro 00:00With regard to organizations prioritizing sustainability goals, a big part of that, in my opinion, is ensuring that those goals are publicly stated so that there is an element of accountability around that.Gael Duez 00:24Hello everyone, welcome to Green IO with Gael Duez - that’s me! - Green IO is the podcast for responsible technologists building a greener digital world, one byte at a time. Twice a month on a Tuesday, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the Tech sector and beyond, to boost Digital Sustainability.And because accessible and transparent information is in the DNA of Green IO, all the references mentioned in this episode, as well as the transcript, will be in the show notes. You can find these notes on your favorite podcast platform, and, of course, on our website greenio.tech. Decarbonizing a tech stack, or a platform, or a full information system. This is the challenge many CTO face these days. And beyond carbon, more environmental indicators are taken into consideration. Far from being just a few technical choices to make, it often implies multiple decisions at multiple levels with multiple stakeholders to succeed. Hence, this episode where I wanted CTOs to share and exchange about the strategy they choose to achieve the sustainability goals their company committed to. 2 CTO with different backgrounds and operating in different lines of business are welcome today. Ludi was CTO at Lunii for three years, a highly successful toy maker in France, and she's now digital CTO at Bpifrance. But she's also part of a vibrant community, tech.rocks which gathers French CTO in Paris and beyond. Owen is CTO at Dot ECo, a company operating the registry ECO and has successfully run decarbonization of its tech stack. And he's also part of the climate action tech community. And this is where I've had the pleasure to meet him. Ludi Owen, thanks a lot for joining Green IO today. It's a pleasure to have both of you in different time zones making the effort to be in the show at the very same time. So thanks a lot and welcome to the show.Ludi Akue 02:41Happy to be here. Thanks for the invitation, Gael.Owen Rogers 02:44Thanks for the intro, Gael.Gael Duez 02:48So without further notice, I would like to ask both of you the same question, and it's a highly complicated one. What was the impact, or in another word, what were the goals you were aiming to achieve when you started to reduce the environmental footprint of your technical operations or your tech stack? Maybe, Ludi, you want to start?Ludi Akue 03:16Yeah. Thank you for the question. So the goal was pretty big. It was, we need to change the way we produce everything. So it's a big, hairy goal. Not clear at all. That was the goal I had in the beginning, knowing that the company Lunii was a toy maker. So as you said, a hardware toy maker. And I managed both the hardware teams and the software teams. So the challenge was just clear. Very clear that we need to change things. We need to be impactful, but we don't have a clue about the KPI's to get there.Gael Duez 03:59It was kind of top down. It was like a big strategic enterprise goal and in a very sort of OKR approach, like we've got something very ambitious to achieve, I guess not in a single year. Make something with it. Am I correct here?Ludi Akue 04:19Yeah, I think it's much more subtle than that. There were some key results, let's say more on the supply and the manufacturing side. The first model of the toy was made in China and brought back the manufacturing to France. For example, you have big goals but after those goals, when coming to the tech stack, everything about the tech operations, architecture, there was a blank page. Like, I needed to build a plan with my teams, of course, to measure and to reach the goal of changing, be more responsible.Gael Duez 05:05I guess this is a moment where I have to admit that, a) my daughter has bot histoire, b) actually, I know this because the one my daughter has has been made in France on top of it, but the one my nephew has, which is one year older as still made in China. So I was like, oh, my God, they made it. Congratulations.Ludi Akue 05:31But, yeah, I just want to add something very quick. I was really impressed by the vision of the founders. It's like they wanted to change things as they are growing a growing company. They now have a working business model. I can share some figures. At that time, there were about more than half a million customers, more than a million of toys sold all over the world, in France, in Europe, North America. So they had this strategic vision to say, we need to measure and change our impact. I was really impressed by that.Gael Duez 06:15And what about you, Owen? What were the goals when you started this sustainably journey at .eco?Owen Rogers 06:23So I would say for us it was pretty different in as much as it was more bottom up instead of top down. And we are purely virtual. We just deal with digital products, so there's nothing physical for us to have to worry about optimizing. So my job, I think, was probably significantly easier than Ludi's in that regard. So for us, we operate the .eco top level domain in partnership with the global environmental community in the face of increasingly dire climate effects that we are seeing in the world around us. We were feeling like there had to be something more that we could do some way that we could have an impact ourselves. But it wasn't totally clear how we could do that. Being a company that focuses strictly on digital products, we started to learn about the climate effects of technology and realizing that there was in fact, a lot that we could start to do to reduce our impact. So we wanted to understand what was our carbon footprint. None of us had any expertise in doing carbon life cycle analysis and carbon footprint calculations. So we went and spoke to a number of auditors that will evaluate the carbon footprint of organizations. And the response that we got for most of them was, we're super busy right now, and perhaps there's some work that you can do on your own in order to understand your own footprint, compile the data, and then perhaps we can come and assist you later. So we started to do that ourselves. We started to work with our vendors, with our partners, and pull together whatever data we could to understand our own carbon footprint and realize pretty quickly that this is not complicated. Doing basic carbon calculations is not a hard thing. It's certainly something that you're doing some pretty rudimentary mathematics, and it's something that most software engineers should have no trouble with. And so that enabled us to calculate the carbon footprint of our company, and that served as a basis for optimization. So we did our first carbon footprint calculations in 2020, and we established a goal of reducing our overall, that's our Scope 1, Scope 2, and Scope 3, carbon footprint by 4% for our 2022 fiscal year. And a lot of that was through optimizing our digital services.Gael Duez 09:23But just for clarification, Owen, your goal was to reduce by 4% in two years.Owen Rogers 09:29Within one year. So that was for our 2022 fiscal year. We were quite close to achieving that objective. We managed to reduce our overall carbon footprint by 3.7%. And we have a similar target for our 2023 fiscal year. We actually have no Scope 1 and Scope 2 emissions because of the fact that we're a purely virtual organization. So during COVID we gave up our office space, so we work entirely remotely. We also don't ourselves run any data centers. We don't have any office space. So everything for us is Scope 3 emissions. So we've tried to be fairly broad in terms of including as much as we can within our Scope 3 footprint. So one of the things that we account for is 100% of the energy used in the homes of our employees within our footprint, which actually is quite a significant portion of our footprint. But the recognition being that's where work is happening right now. And so we should account for that footprint. And our employees are not otherwise attempting to account for their home. Any emissions associated with energy use in their home.Gael Duez 11:09Fair point. By the way, Ludi, did you have the same approach? And what can you tell us about the carbon footprint of a toy maker?Ludi Akue 11:23Yeah, that's a very interesting question. Yes. We had the approach of measuring the carbon footprint. And for Lunii at that time, we have the three contributors that we have. The first one is everything related to supply logistics. So transportation, production, and manufacturing. So it's big. If we need to address only one point, we need to address that. So the first contributor, I think it's something about maybe 40%, I think. I don't have my figures. I don't have the figures here. And the second one is the operations like customer service. And the third contributor was the tech. So they take about 6%. I remember the figure because it was 6% of the whole. So as a hardware company, the real job is to tame the production before the other one.Gael Duez 12:32Right. That's true that you're a CTO, but you're a CTO not only in the way you were a CTO at Lunii, but not only in the way we think about it in the digital sector or in the IT industry, because you are in charge of the IT infrastructure, obviously, and we will talk about it. But you were also in charge of the entire designing and building of the toys itself.Ludi Akue 13:00So the first thing is to measure. So we have these carbon footprints measurements with the leader in France. It's Carbo, hellocarbo. So we found out that our… So we have pretty much more Scope 1 than everything else. Even if we needed to help our suppliers to comply with our mission, our goal said. So we had a supplier code of conduct to do that. And the key results were bottom up. So we had this high goal. We have this big result, like relocating in France, co-designing the product, and for the rest… So the operational teams come with the key results they need to address, they need to tackle. So you had all the production team, the production is supply and everything. The real job here is to tackle the Scope 1 carbon emissions. So the production teams managed to has the carbon emissions. So it was a great win that we can have carbon emission.Gael Duez 14:18That's very interesting that you say that actually the IT itself was only 6%, because that's, I guess, a challenge that many CTO and CIO, they face, that they don't weigh that much within an organization, but if we gather all of us together, we weight sometimes as much as the ROT freight transport worldwide. So that's quite significant. And that's what I call the green IT curse, which is you're small within an organization, but eventually you're super big at that global level. So if we focus on this more IT side of things, Ludi, you mentioned at the beginning that you had a very aspirational goal. How did you manage for IT operation and also the firmware, as you mentioned, to make it something tangible, something actionable.Ludi Akue 15:13The first thing to do is once we have the measurements, like the carbon footprint, we have something to reduce. You have a goal, you have something to reduce, like a salesperson, let's go differently. And how we tackle that is to identify where we are not very good. Like for example, something that I want to stress here is imagine it's a fastly growing company. It's a scale up. So the scale up, as every scale up has some technical depths. And so here the technical depth is an opportunity to think differently. We need to rebuild something, so how can we rebuild it? For example, let's go mobile first to reduce, to optimize the web pages. Some indices were like, let's move into the cloud. So I decided to move into the cloud and to use managed services and to benefit from the calculation. Even if the calculators are not really reliable, they can show you some trends. So that's interesting. You can measure that. Let's be simple in the way we move to the cloud too. Taming technical labs is when we had an existing microservice architecture, but we don't really need a microservice architecture. So it's like, it's to say like the vision, technical vision is to go back to a modular monolith, for example. It was a strong choice. Our job is pretty simple. So for the scale, everything was homemade. Like you have an e-commerce platform that is homemade on a microservice architecture. Lunii also has a mobile application for the usage of parents and kids can recall their own stories as well. And you also have an application, a usage application, I will call that, that you can download stories, you can buy stories, you can download stories, manage your toy. Right? So the technical vision here is, okay, everything was built for growth. Right now, what is the company’s job? The company's job is to make toys and audio stories. So we don't need to continue to build an in house e-commerce platform. So a strong decision was like, we need to stop building the e-commerce platforms and buy an e-commerce platform like Shopify. This was a strong decision, so we don't, don't need the whole infrastructure that comes with building an e-commerce website.Gael Duez 18:13And that's super interesting, because to make sure I understand it well, your main point here is leverage technical debt, because you were in a scale up company, so you had the opportunity to redo things and you, like, technical debt becomes an opportunity to rethink everything and rebuild everything in a more sustainable way. And the way you did it was a bit counter intuitive regarding state of the art architecture, etcetera. It was more… So, obviously, you had like, let's try to eco-design, let's try to move to the cloud. I will have one question about it, but it's maybe we were too ambitious with our architecture and a good old fashioned modular monolith. And buying more off the shelf stuff will actually help us to reduce environmental footprint.Ludi Akue 19:05Yes, absolutely. It's like, it's just technical debt was the leverage we needed to do things differently. It is simplicity [and] optimization, simplicity and automation.Gael Duez 19:21And just for the sake of understanding this e-commerce platform that you've chosen, was the sustainability criteria taken into consideration?Ludi Akue 19:33I think that, no, we didn't include that. Maybe we did, because we have a supplier code of conduct. I'm sure we have those because we have a companion wide supplier code of conduct. It was a matter of simplicity. So I think the first thing is simplicity and automation. Yeah.Gael Duez 19:54And what about you, Owen? What were the technological choices you made to achieve the -4% reduction? Was it cloud migration or anything related or a very different approach?Owen Rogers 20:10So, like Ludi, our digital services accounted for only 10% of our overall Scope 3 emissions. So not too far off from the 6% at Lunii. The 4% reduction target was going to be entirely achieved by optimizing the efficiency of our digital services. So really what that meant was we were looking to reduce the carbon footprint of our digital services by upwards of 40%. So for us, we are running entirely in the cloud. So it didn't entail a migration from physical data centers to a cloud environment. We were already there. But one of the key things that we did was moving our infrastructure from one region to a lower carbon region. So we ended up migrating our entire stack to Quebec, Canada. Electricity in Quebec is entirely generated through hydropower. It has one of the lowest carbon emission intensities in the world. So that was definitely a positive move in reducing our footprint. Otherwise, we did quite a bit of work in terms of optimizing the footprint of our website and services similar to what Ludi was talking about doing work around optimizing images, JavaScript, CSS. Our sites are currently already static. Not entirely, but most of them are already static. And that's definitely a huge win because you're not running any servers or we're not running any servers on the back end. But my approach architecturally is actually a little bit different from the solution that Ludi was suggesting. I can definitely understand why with a microservices architecture you may end up with a very large fleet of instances depending on how you distribute that work, cross compute. But our focus is more on trying to eliminate servers altogether and ensure that everything that we're building can run through serverless. Serverless is, I think, the current end state or optimum for virtualization within cloud providers. It enables them to maximize multi-tenancy so that you can have the most compute being run through the fewest possible instances. So that's the direction that we're moving, at least for the parts of the tech stack that we have under our control. We were already using serverless and we're continuing to use serverless, but in terms of architectural direction, that's where we're headed. Anything new that we're building, we're trying to make it serverless first and move away from having to deal with any infrastructure that we need to manage ourselves.Gael Duez 23:10Did you use a specific tool to measure the carbon footprint of your tech stack or how did you tool up? And maybe the answer is not at all. We didn't need to.Ludi Akue 23:22We used a tool, Carbo. Carbo is a French tool. I think it's hellocarbo.com. It's a bit of calculation and declarative, also measure. We also use the cloud calculators. So you have a lot of calculators with the major cloud providers so you can have an estimation of your carbon footprint. Something that works great at Lunii is we build external accountability. How do you do that? Is like we sign the global compact. So the Sustainable Development Goals of the UN we signed that we needed to provide every year some reporting on how we are advancing the mission. We are also signed with a collaborating company, Lucie. So Lucie 26000, we needed to have a clear goal in our OKR to have a Lucy label by 2022. So we build external accountability to make it work like it's not just us with our OKR, but we need to report to the UN that we measure everything that we build this, we build that, so this helps us move forward.Gael Duez 24:48But as far as I know, hellocarbo is a generalistic carbon accounting tool. So you didn't have the need to tool up specifically to measure anything regarding your IT stack except the data provided by the cloud providers, as you mentioned. And what about you, Owen? Did you use any specific tool or not that much?Owen Rogers 25:14So for us, we followed the sustainable web design methodology as the foundation for our carbon footprint calculation. For our digital services, that methodology uses data transfer as a proxy for calculating energy consumption, and then on the basis of that can be converted into carbon. It certainly is not perfect, and unfortunately, there aren't really perfect measures at this point in time in the cases where you don't own the infrastructure. But it does have the benefit of relying on data, data transfer that is relatively easy to obtain. So what we did for our digital services was look at data transfer, whether that be transiting via the content delivery networks that we were using, or load balancers, etcetera, and estimated our footprint on that basis. So that's the tooling that we used. It's not a specific third party tool. A lot of this data was just aggregated in spreadsheets, and then the calculations were being performed there.Gael Duez 26:33And there is nothing wrong intrinsically with using proxy metrics, as long as we don't claim that it helps us understand absolute values. And for anyone following Gerry McGovern's work, I think using data as a proxy, it's far from perfect. There is a question of embedded carbon, etcetera, etcetera, but it doesn't lead people in the wrong direction. That is simply not true. Maybe we will have better metrics, better tools, as you say, Owen, but a company strongly engaged in reducing the amount of data produced they manipulate the exchange is not going in the wrong direction when it comes to reducing our environmental impact. Now I'm fully convinced of it.Owen Rogers 27:23Well, in general, data transfer is correlated with energy consumption. But if you look at most organizations in terms of how they're calculating their carbon footprint, especially for their Scope 3 emissions, it's almost entirely based on spend. So whatever their expenses are, they're using some sort of a general metric that is converting dollars spent for this category of product into carbon intensity. And personally, I think the data transfer is going to, at least within the technology space, be better correlated with energy use than spend. And spend also gives me nothing. I mean, I can try and optimize my spend, but it may point me in the wrong direction, like what we were talking about before. I could move all my infrastructure to a high carbon intensity region, reducing my cloud spend and therefore, under that basis of calculation, reduce my overall carbon footprint. I mean, it's just crazy. It leads to the wrong types of outcome, at least with data transfer. Plus, one of the nice things about the sustainable web design methodology is that it is somewhat component oriented. In cases like for my infrastructure, I have a better way to calculate the footprint because I know the number of instances that are running and I can estimate the infrastructure and I know what the carbon intensity is of where that infrastructure is running, I can get a better estimate for that component. But if I want to have a broad scope that includes emissions associated with the use of our products, so the energy drain on end users devices as a result of using our digital services, I can still estimate that under the umbrella of the same methodology and use data transfer as a proxy for that. I'm hoping there's going to be a revision to the sustainable web design methodology. The better accounts for what we're learning now about the energy use associated with the various components involved in the end to end journey for digital services, so as to better discount the network component. But I do want to underscore the importance of something that Ludi was talking about, which is with regard to organizations prioritizing sustainability goals. A big part of that, in my opinion, is ensuring that those goals are publicly stated so that there is an element of accountability around that. So Ludi was talking about the UN Global Compact as establishing a requirement for a public report around sustainability. Similarly for us, we publish our Annual Impact Report and part of our impact report is setting targets for the upcoming fiscal year around sustainability and impact. And so that means that you have an element of accountability there to your customers, to the community you serve, to the general public. And I recognize that that is hard, especially in cases where we're relying on data that we know isn't perfect like there… I know that the sustainable web design methodology has flaws associated with it and that I'm sure in a couple of years we will come up with some sort of more comprehensive methodology that will enable us to produce more accurate carbon footprints. But I don't want to let the imperfect be the enemy of the good. I think it's much better for us to try and get in front of this and be transparent about what we're doing with the data that's available. Recognizing that at some point later we're going to have to say, okay, we're going to need to revise these estimates of our footprint and our impact, because this is where the state of the art is now.Gael Duez 31:35Ludi, Owen mentioned not the cloud migration, like you had to do at Lunii, but relocating within the cloud operators in Quebec, which has tremendously low carbon electricity, even if we include embedded carbon from the hydropower. When you talked about the migration you did to the cloud, did you consider the region that you wanted your services to be hosted to make sure that actually you might not migrate from a very old data center with a PUE of 2 to a brand shiny cloud providers with a PUE of 1.2, even a bit less if you listen to their marketing communications, but hosted in an area where the average carbon intensity of the electricity will be at time five or six higher, which is very classical case with a very famous three letters hyperscaler. But I will stop here. So my question, sorry, I'm rambling a bit, but my question is, did you have the opportunity to consider the intensity, the carbon intensity of the electricity grid from the region where you relocated your cloud activities.Ludi Akue 32:56Not directly. We were on a pass before. We didn't pass. Like I said, we had this microservice architecture, and it was really ineffective in terms of management, in terms of billing, bills were very expensive and ineffective. Like, we cannot really measure things in fine-grained speaking. So the first thing is we consider green goals when moving to the cloud. This was really clear. We work with a provider that helps us migrate because they're green objectives while moving to the cloud. We also consider GDPR goals when moving to the club. So we were more like, we want to stay close to our customers. And at the same time, our customers were kind of everywhere. But our big, you know, huge base of customers were located in Europe, and we chose to locate our clouds in Europe. So privacy was first, the closeness to our customer base was second, and then ended in green objectives.Owen Rogers 34:24For us, we migrated from Ohio to Quebec. And so Ohio, the grid in Ohio is largely powered by coal and by gas, and so there's a 200 times difference in carbon intensity between Ohio and Quebec, so it's significant. And Ohio is not the worst in terms of carbon intensity within the United States. So there are much worse places to be and many of those places operate large data centers. So I think it's absolutely something that technologists should be taking into account when they're considering a cloud migration or when they're looking at ways to reduce their carbon footprint.Gael Duez 35:14And just a side note, that's something really to pay attention to, because unfortunately, sometimes carbon incentives and financial incentives are not well aligned in the hyperscaler industry. And that's really a shame, because obviously, the dirtiest place should be the most expensive. And that's not always the case. And sometimes a CTO faces a very difficult challenge to explain that he wants an increase in 5, 10, 15% of his or her bill, because that is to achieve a target of decarbonization, and that shouldn't be the case. But that's okay. So that's something to pay a lot of attention to.Owen Rogers 35:50Well, it did increase our bill by moving to Quebec. It is a more expensive region for us to operate in, but not significantly so. So that was a price that we felt like we were quite comfortable to pay. But to both your point and to Ludi's point, as a CTO, it's essential to find alignment between overall business objectives and your environmental objectives. Now, fortunately, in many cases there is alignment. There are ancillary benefits that can be achieved as a result of reducing carbon. So specifically, a low carbon intensity website is generally going to be a higher performing website. So you're able to pitch this type of an initiative under the guise of performance improvement, which will boost user experience, hopefully boost retention in sales. So there are these ancillary benefits. Similarly, a low carbon intensity infrastructure, so relying on serverless or higher levels of virtualization, is generally going to be more cost effective. So often there is alignment between an infrastructure cost reduction program and reducing carbon intensity. At the very least, you can find it if you look for it. It's not to say that reducing spend within the cloud is always going to lead to lower carbon, but you can find and make the case for where those two things are going to be aligned, and then that gives you a lever to push on.Gael Duez 37:30And what a wonderful transition. Because I wanted to ask both of you the question of process and strategy, like how you connect, how you align with the business strategy, and what were the process, what was the organization that you set up to achieve these goals? Because we discussed quite a lot the technical side of things, but it's more a human thing than a technical thing. So maybe, Ludi, you want to share also how you actually created something from the wide statement. And of course, how did you manage to align them, as Owen just said, with the overall business goals?Ludi Akue 38:11My context was easy, because you have these big goals for the company, so everyone is aware it's a strategic goal. And I recommend that green initiatives should be taken at a strategic level for everybody to be aligned. Right. So once you have that, the fact is, as you are a growing company, you scale up, don't have time, do not really have time… the tech team didn't have really the time to build the product, ship the product, take a technical debt and also take into account something to reduce the carbon footprint. So for a long time, it was kind of a failure to add some green initiatives to the backlog. Something I learned is adding some green initiatives to the backlog is the failing mode. You should be able to initiative into the existing backlog. And the first thing is to work with the teams is to come up with the optimizations. So optimizations pay for the business. That's the first thing. So if you have optimizations, if you have your website that respects the providers, for example, it's good for the SEO, it's good for the user conversion. So you can show the business your optimization strategies. Link to the ARR of the subscription number of the user that will convert it to buy a product, to use your service more so you can showcase that your optimization is aligned with the business. Right. When taking the example of the cloud migration, for example, is to show that the existing is very complicated, it is costing us. So the hidden cost that we have with the existing systems and to show the return of investment with the migration to the cloud. So it is the same thing. And it is to make the business case of when you can make the business case and for the rest, as it's the whole company initiative, to show that you are working towards the goal of reducing carbon footprint.Gael Duez 40:42If I understand well here, there are two main takeaways of what you've just shared with us. One is optimization will pay for decarbonization. And I would love to comment on this one. And the second one was do not try to create a separate track, a separate program about sustainability. Just add it as a new criteria in the definition of dawn or a new type of task or sprints in your backlog. But do not try to run something separately, which is, as far as I know, kind of now a state of the art approach. Also, when it comes to cybersecurity or accessibility, for instance, don't try to create things on their own. So same applies to sustainability, am I right?Ludi Akue 41:33Yeah, you are correct. You need to shift left on sustainability.Gael Duez 41:39When Ludi mentioned in our previous talk that there is a sustainability manager, or at least a sustainably reference at Lunii, was it the way you were organized at .eco? Did you have someone kind of running the show? Or did you have some sustainable reference in the different teams? How did you make sure that the key results, we're not talking about the overall strategy, but the day to day operations embedded in the sustainability goals.Owen Rogers 42:13We're a small organization, so there's not the same level of challenge associated with communicating and disseminating these types of objectives throughout the organization. In general, we have a set of KPIs around impact that are pretty well broadly understood. We review them on a monthly basis, along with how we're tracking towards our other goals for the organization. So I think it's pretty front and center for our team.Gael Duez 42:45Would you be able to share this KPI? Not the numbers, but what they are?Owen Rogers 42:50Well, I think the main one in terms of climate impact is our overall carbon footprint reduction target for our fiscal year. So again, we're aiming for another 4% reduction from our baseline target. We also have an objective to register through SBTI. And one thing that we have achieved recently, which is not directly related to climate, but all kinds of plays together, is that we are now B Corp Certified. We will be certified as a B Corp. And that touches on a lot of different pillars around sustainability.Gael Duez 43:29So if I understand it well, it was a very bottom up approach. You're a small team, so the communication was pretty straightforward the moment you had the right goals and not that much of an issue on how to incentivize people. Ludi, if I get it right, it was more a top down approach and your organization was scaling at the time where you were trying to reduce your environmental footprint. So what about the human side of this? Did you face challenges to get the sustainable angle adopted? How did you convince people to join?Ludi Akue 44:10Yeah, so we have these goals, these top down goals, but I think in the same test, even if you have 100 people, you have the people who were at the company at that moment were really aligned in terms of their own values. So you have these top down goals, but you also have a bottom of approach. Like the teams were responsible for their key results and their key initiatives that should align with the goal, like, because they are doing the work, so they know where to look at. And something that really works, in my opinion, is having a sustainability manager reference. I think it's something really powerful because she did a lot of training. So first, the whole organization has champions into the teams. So every department has sustainable champions. And the champions help spread the world, spread the culture, spread the training, the information. The challenge was not on the human side, the challenge was on the workload side. Picture, scale up, growing fast. We need to build 10,000 things in parallel. And we want teams to focus on sustainability. That's not going to happen.Gael Duez 45:41And how did you manage to get some positive arbitrage towards sustainability.Ludi Akue 45:47We learned to make it right at the executive committee level. We needed to come up with priorities first to tell the teams that, okay, now we don't have 10,000 priorities to chase. You have five ones. And sustainability is one of them. So make room for that. So when we started to have those ok in place and to have these priority likes, every executive committee member has the priority to make sustainability happen.Gael Duez 46:22I've got one last question. What about the trend that you see in our industry? And for that, it's very interesting to see that you've got two different perspectives. Because, Ludi, now you don't work for Lunii anymore, you're at Bpifrance and you see dozens of different startups, scale ups, companies, and you've still got this sustainable young angle at heart. So I'd like to ask you the question first. What is the trend that you see today regarding digital sustainability?Ludi Akue 46:55What I'm saying is sustainability is not anymore an option. It's not. I can build my company and I will figure out that later.Gael Duez 47:08Can I borrow your statement that sustainability is no longer an option? To make it the title of the episode?Ludi Akue 47:18With great pleasure.Gael Duez 47:21And what about you, Owen? You're in the heart of the web hosting industry in particular. More generally, is the digital technology sector sustainably an option?Owen Rogers 47:35I think that for the time being, it is an important consideration for anybody looking to get into the technology space, whether they're launching a new business or an organization, is to think about how they can do so in the most sustainable way possible. At a certain point in time, I don't think it's going to matter. I think that with the rapid decarbonization of our electricity grid that we're not too far away from a scenario where most data centers are going to be powered using renewable energy, or at least low carbon energy. But I think that we have a role to accelerate that transition as much as possible by helping to educate the general public that their choices matter and that there are options out there that will reduce their climate impact and it won't cost the earth that there are cost effective solutions out there. One of the KPIs that I didn't talk about earlier that we have established for ourselves is the proportion of people using eco domains that are using sustainable web hosting. So on an annual basis we go and evaluate every .eco website and run it against the green web, check from Green Web Foundation to see whether or not they're using sustainable hosting. And what we're seeing is year over year growth pretty significantly. I believe that last year we saw an increase from 33% to 40%. Now, some of that is hopefully due to our education, but I think a lot of it is actually due to the changes that are happening in the overall industry where more and more hosts and data centers are working on ensuring that they offer a sustainable solution to their customers.Gael Duez 49:43So that's excellent news, because once we will have achieved decarbonizing the data centers on the energy consumption angle, we will have to open the even more complex and complicated question of embedded carbon and making sure that we use equipment as long as we can and not change them every five or six years. But that's a different debate.Ludi Akue 50:10Yeah, but it is the... I think that it's... When we are looking at the statistics, it's the huge problem ahead. It's a huge problem you need to tackle. Like, we went talking about community, we also have. We are also responsible not to like having heavy websites, heaven applications that will need people to change their mobile phone, for example. So we have many things to do on that side too. And for me, the end device is where the problem meets the core problem. It is something we need to fix what we can fix, the decarbonization, like you are talking about, Owen, to move to the real problem. The real problem is there. The end devices and the raw materials also. I mean, together.Gael Duez 51:16Yeah, agree. And that's true that with your angle of being someone who also had to produce hardware, you've got this, I would say some awareness about the fact that end user equipment still accounts for the largest part of the overall digital sector, carbon footprint. It's true that with an episode with a CTO, we tend to focus on how to decarbonize a tech stack, and how to actually reduce the environmental footprint of the tech stack. But if you look at the entire digital sector, the data centers and networks are not the places where the majority of the impacts happen. It's still a very significant chunk of these impacts. But the end user equipment is even worse, that's for sure.Owen Rogers 52:08I'm optimistic that that's what's coming next is that we'll see a real revolution in terms of e-waste management and like a cradle to grave approach with regard to electronics manufacturing because of the fact that they're. They're full of valuable minerals and it's... At a certain point it's going to be a lot more cost effective to reuse those minerals than it is to continue to extract it. I'm on a bit of an optimistic kick right now because I've been reading Not the End of the World by Hannah Ritchie, who I think would be a fantastic person for you to invite on the podcast. I know she's been doing a bit of the podcast tour lately, but I think it is important for us to ensure that we're also telling positive and optimistic stories about our future because it's too easy to get drawn into some of the fear and negativity around climate change.Gael Duez 53:11So two fun facts when I'm seriously considering, but I don't know if I can afford her, but to have Hannah Ritchie as a keynote speaker, or at least as maybe a keynote speaker. I would like to thank you again, both of you, to join, to make the effort to find a way to accommodate our different time zones and to be that precise when it comes to the goals, what you did and what you didn't do, and the process and the the human side of things as well. So thanks a lot for joining. It was a pleasure to have you on the show.Ludi Akue 53:51Thank you, Gael.Owen Rogers 53:52Thank you and thank you both. Thank you, Ludi. Great to meet you.Ludi Akue 53:55Yeah, great to meet you too.Gael Duez 53:59Thank you for listening to this Green IO episode! If you enjoyed it, please share it on LinkedIn, on Facebook, on Twitter (if you are still there)! We are an independent media and word of mouth is the only way to get more listeners. I don’t ask you to rate it 5 stars on Spotify or Apple podcast because of course you already did it, didn’t you?  In our next episode we will talk about Japan, not because the country used to be super dear to my heart. Yes, before the K-pop wave, another generation - mine - discovered Japanese animation and was crazy about it. Anyway, we won’t talk about manga but how a super techno-centric country is now considering sustainability and the potential limits that the ongoing ecological crises might create to our digital world. A fascinating discussion with 2 experts who have been living there for ages: Paul Beddie and Trista Bridges.Stay tuned.  BTW, Green IO is a podcast and much more, so visit greenio.tech to subscribe to our free monthly newsletter, read the latest articles on our blog, and check the conferences we organize across the globe. The next one is in London on September 19th. Early bird tickets are not available anymore but … you can still get one of the 90 free tickets for Green IO listeners using the voucher GREENIOVIP. I’m looking forward to meeting you there, to help you - fellow responsible technologists - build a greener digital world, one byte at a time [Roxane’s voice]❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.  📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents. 
undefined
33 snips
Jun 18, 2024 • 1h 1min

#41 - Decarbonizing AWS with Adrian Cockcroft

Former AWS sustainability leader Adrian Cockcroft discusses decarbonization in the cloud industry. Talks about measuring carbon for software, challenges faced by Cloud providers, and the need for sustainability in architecture. Highlights include the case for PPA, market-based approaches, and the importance of transparency in AWS sustainability efforts.
undefined
Jun 4, 2024 • 53min

#40 - Triggering action in green software: a Nordic perspective with Satu Heikinheimo and Janne Kalliola

Explore how Janne Kalliola and Satu Heikinheimo discuss triggering action in green software, including the trend of employees resigning over conflicting values, importance of collaboration for sustainable products, concept of carbon-neutral software, and Nordic approach to fostering sustainable initiatives.
undefined
May 21, 2024 • 54min

#39 - European Regulations in Tech: some insider perspectives with Kim Van Sparrentak and Max Schulze

European regulations in the tech sector are explored with Kim Van Sparrentak and Max Schulze. They discuss the Artificial Intelligence Act, environmental transparency, and the influence of big tech companies through lobbying. The podcast delves into the need for sustainability in regulations and the right to repair directive in technology. It also touches on the impact of lobbyists on tech laws in Europe and collaborative actions for sustainability.
undefined
May 7, 2024 • 43min

#38 - Building Green Software with Sara Bergman

Seasoned software engineer Sara Bergman discusses the angles covered in 'Building Green Software' such as AI, temperature-aware computing, cloud providers, and carbon vs grid-aware computing. They delve into the responsibility of software people in utilizing hardware and the importance of major cloud providers. The conversation also touches on the concept of temperature-aware computing and the environmental impact of hardware in software development.
undefined
Apr 23, 2024 • 53min

#37 - How to be a climate-conscious product manager with Antonia Landi and François Burra

💡In product management, there is a strong culture of adapting to change, testing, and rapid learning. Talking about change, we have a big big one ahead of us: climate change! How can product leaders include climate considerations into the build and run of digital services and scale sustainable practices? 🎧In this episode, Gael interviewed two hands-on experts and daily practitioners in product management on how to become a climate-conscious PMs. Antonia Landi, based in Germany, is a strong voice in ProductOps, and François Burra, based in Canada, is the co-author of the Climate Product Management Playbook.What product manager should consider?🌱 Environmental impact throughout the product life cycle,💻 Minimalistic approach on user experience, 🎯 Climate-related OKRs and reducing costs through intentional data storage.❤️ Subscribe, follow, like, ... stay connected the way you want to never miss an episode!📧 Once a month, we deliver carefully curated news on digital sustainability packed with exclusive Green IO contents in your mailbox, subscribe to the Green IO newsletter here. Learn more about our guest and connect: Antonia Landi’s LinkedInFrançois Burra’s LinkedInGaël's website Green IO website 📧 You can also send us an email at contact@greenio.com to share your feedback and suggest future guests or topics.   Antonia and François' sources and other references mentioned in this episode:Climate Product Management PlaybookDigital Product DecarbonizationPaper beats plastic? How to rethink environmental folkloreTranscript[00:00:00] Gaël DuezHello everyone. Welcome to Green IO with Gaël Duez. That's me. Green IO is the podcast for responsible technologists building a greener digital world. One bite at a time, twice a month on a Tuesday, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the tech sector and beyond to boost digital sustainability. And because accessible and transparent information is in the DNA of Green IO, all the references mentioned in this episode, as well as the transcript, will be in the show notes. You can find these notes on your favorite podcast platform and, of course, on the website greenio.tech Product managers have always lived in a world of contradictory injunctions. They are at the converging point of all requirements, business requirements, of course, and all the others. Let me illustrate with some sentences which might be familiar to many of you. We cannot launch feature X because I need some time to deal with my technical debt. Make sure your product is accessible and secure. We need to launch this proof of concept to test our hypothesis next week, but beware of being pixel perfect because of the risk for our brand image. Accessibility is key in our values. Such as innovation. So let's not miss the latest trend in virtual reality, et cetera. And now after cybersecurity, accessibility, performance, et cetera, a new wave is coming. Sustainability, with sentences like beware of your carbon budget. How much will this new feature impact our CSRD reporting? Watch out for greenwashing with these product claims. Did you check the W3C sustainability guidelines for your product? And product managers in charge of back end products are not off the hook. How does it perform with our API green score? Our cloud operations are too carbon intensive. Can we delay some process to enable carbon aware computing, or even better, grid aware computing? To be honest, I'm not that worried about the product management world succeeding eventually to overcome these new challenges. There is a great culture of adaptation to change, testing and fast learning in the product management line of work. The main questions are how and when, because time takes time with climate change. And many sustainability tools for product managers are still missing. Hence, a question for this episode is how to become a climate active product manager today, not tomorrow.And to try answering it, I asked two hands-on experts and daily practitioners in product management to join our show today. Antonia Landi, a former Aviv Group colleague, has become one of the top voices in European product operations. Being a community person at heart, I can testify, and loving bringing people together, she leads Germany's only meetup for product operations in Berlin, where she's based. You can easily find her work online. She has written for publications like LeadDev, Product Alliance, and MindTheProduct. François Burra is a Canada based product lead and UX consultant, mentor, and speaker with 13 years of experience working with startups and agencies in North America. After taking almost a full year off following climate courses, he pivoted his career to fight the climate crisis. He now helps digital companies and product teams reduce their digital emissions via consulting, measurement, and training. And talking about training, I had the opportunity to meet him via the wonderful climate action tech community when he was presenting his newly released climate product management playbook, which was well received in this community and others. So welcome Antonia. Welcome François . It's a pleasure to have both of you on the show. [00:03:56] Antonia Landi It's a pleasure to be here. [00:03:58] François Burra Thank you for the invite, Gaël.[00:04:00] Gaël Duez It's my pleasure. Thanks a lot for joining to set the stage. I would like to ask a general question about the context and the momentum. Why do you believe both of you that product managers have a unique opportunity and responsibility right now to embrace sustainability? Maybe François , you want to give the first shot? [00:04:23] François Burra Sure. I think product managers have a unique opportunity because their role, the sense of what they do is based on influencing. So whether it's influencing their team or even beyond their team, they collaborate with all departments across the organization, they tie the strategy with the execution. So they have these posts, like the way I see it visually, like a vertical and horizontal influence. And because of that, if they integrate sustainability as part of their vocabulary and part of their processes. Then suddenly it can shine throughout the whole organization. So that's why for me, they really have a unique opportunity to play a big role in making their job, a climate job and their company, a climate company.[00:05:08] Antonia LandiYeah. I think to add to that, for me, PMs have a unique responsibility because we'll actually have, or should feel a unique responsibility, right? To make sustainability the forefront of what we do, and especially looking at the climate impact of tech off the everyday things we use, right? How? How sustainable is this? You're actually working on and how much could you actually meaningfully change. And if we all individually take small steps to take sustainability to the next level in our digital products, then together we can make a very big impact. So to me, that's exactly why we should be talking about this.[00:05:53] Gaël DuezAre more people talking about sustainability in product management these days than before? I mean, both of you, you're privileged witnesses in your industry. So is it still a happy fuse or do you see something a bit more general happening? [00:06:11] Antonia LandiYeah, I think for me, honestly, there's still not enough. I think there's definitely more than there was even a year ago, right? We are starting to think about sustainability in product management. We're starting to right shift left with sustainability, not do it after the fact, after a product has been created, but really as we create the product. But to me, it's still not nearly enough, honestly, like we are scratching the surface only just barely, which is also why I was so excited to have this happen to have this recording where we uniquely focus on what PMs can do. [00:06:51] François BurraI 100% agree with Antonia. And thank you so much, Gaël, for creating that space because we need more person like you and podcasts like you to raise awareness, not only on the topic of digital sustainability or green IT overall because being French, I have ties in Europe, but living in Canada and North America, I can see the gap that exists between those two markets. And I usually joke, or it's not necessarily funny, but that there's maybe a three, five, seven years gap in terms of awareness between the two continents. And when it comes to product management or climate positive best practices, we even behind I think he started more with designers and engineers, you know, this movement but product managers still are lagging behind those two roles. And that's why I got passionate about trying to do something about it because when I was in a position with a startup. That was borderline doing greenwashing without bad intentions, right? The leveraging of the marketing that would come with it and the benefits that comes with this. I was faced with a sense of not being able to properly address those concern or I was not educated and didn't have the tools to influence the company to do a better job and avoid implementing those practices that were not necessarily transparent, which reminded me of a company, a startup that I used to work with a few years back before I made my pivot to work as solely on the climate crisis as a PM. I didn't have the tools to properly see through the marketing strategy that we had, and basically we were making claims that we were environmentally friendly companies or aiming towards net-zero and so on and so forth. But purely basing on the whole strategy on offsets on carbon offsets without any other actions or awareness of what it means to be in a zero or try to fight climate change. And we were making those big claims to our users that if they were using our app, they would do a good gesture for the environment. And I always knew that matter was super high. Unfortunately, I didn't have the tool to properly educate people or push back against it. So it ended up being part of the strategy and released and everyone was leveraging it in our social network and so on. And that's also why I felt this frustration, this not necessarily anger, but this desire to start educating myself and, and with that, try to spread the message and raise your awareness for my peers. [00:09:31] Gaël DuezIt reminds me of a very personal story of a guy who had a very significant over several hundred millions of budget in a big company as a CTO or group CTO, you name it the way you want and didn't do that much because he just thought that it was all about asking AWS to be greener. I must admit this guy was me. And yet, François , it's interesting what you've described because it's really what Harriet Kingaby in a previous episode on greenwashing, she referred to it as unintentional greenwashing, like greenwashing was the best intention. It's just that you don't have a, like a master plan and with people with dark hoodies saying, "Oh, we're going to screw our customers." No, it's just. You believe this is the right stuff to do because you don't have the information and the basic information regarding offsetting is sorry dude but the civil aviation took all the available land on earth to plant trees to offset their only activities and there is no room left for any others so that's just the basic issue with offsetting but I still love this example.[00:10:35] François BurraAgainst planting trees. Right. [00:10:37] Gaël DuezBut yeah, it's super important. Yeah. It's super important to plant trees for sure, but it doesn't offset anything. It's just that it will help us to rebuild the carbon sink that we need so badly. And what about you, Antonia? [00:10:53] Antonia LandiYeah. I mean, just listening to you, François , like I worked at the sustainability startup for a little while. Like, our whole thing is it was a reusable shipping box, right? You could reuse hundreds and hundreds of times. It had some smart trackers and everything. Obviously there was an app, right? So even though our core mission was sustainability, recycling was enabling the circular economy by, I don't know, packing your old clothes in the box or dropping it off at the center, having a community around all that. It's stopped at that product. Right? Even back then, we weren't savvy enough to think about where our data is hosted? How much of this data do we actually need to host, right? How many iOS versions or Android versions are we supporting, right? Things that nowadays, after my own research, after sort of embedding myself more into these topics, seem almost like common sense to me, but even in such a climate active, I suppose, environment. We weren't even thinking about that because tech was always only seen as an enabler and not as one of the root causes, right?[00:12:17] François BurraI think there's something that I observed over the years, which is that people especially climate tech or sustainable companies because their whole mission is to deliver impact based on their mission, they forget that the why is not allowed to align with the way they do it. And they could implement the worst practice in terms of, like, digital, right? And yet the target, like, claims that they're doing a good thing for the word, but, it's kind of like this duality that is not well understood. And what I could observe is that the reason why the company is created and the how, which is how they build their software from a digital standpoint, because they implement the worst practices at the end of the day, backlash the whole reason why they were created in the first place, which is something that obviously. No one is perfect, right? So the intention is to try to do good in all areas of the company. And there's parts that would be more advanced than others, but it's important and that's what today is important to raise awareness about, like how you can build software and digital products because that could impact. I have a great impact and you need to be aware of it to lessen the negative impact that you can have while building and maintaining what is created to solve this valuable problem, the society.[00:13:43] Gaël DuezI think it's very valuable feedback that you share about this discrepancy between what we want to achieve, what we are truly doing, et cetera. But both of you are daily practitioners, so let's get our hands dirty. How can we become a climate active product manager today and what is sustainable product management?[00:14:08] Antonia LandiI think for me, sustainable product management really is, so I have this super basic definition of what product management is, right? It's delivering value to the user in a way that benefits our business. And to me, sustainable product management is in a way that benefits our business and our planet. I think ultimately sustainable product management starts from the moment you are identifying problems and scoping solutions, but it can run the whole gamut of, and this is something that people will not want to hear, right? But how much data are we storing and how much of that data are we actually looking at? And then, I mean, I addressed this beforehand, right? Like the implications of our product choices. How many mobile phone versions are we going to support with our app, right? And it becomes more and more narrow every single year. And the consequence of that is that we are excluding ourselves from people who want to have older and perfectly working phones. I think there are just many small knock on effects that we just need to play through in our minds. We just need to think through the ramifications of small choices like that.[00:15:32] François BurraYeah, I love it. I think, as you said, like, we need to, as PM, consider the environment and sustainability as we used to think about the triptych technology design or user and business. And for me, we can see sustainability as the fourth pillar, or we can see it another way, which is kind of underlying conditions. So, if you apply those best practices will create a better experience, will create better technology because you learn faster, but also will drive better business results. So I kind of see this triangle like 3D shapes in my head right now, which may not make sense, but I think it's, we can really see it as something that will make your whole business and product better, but I think what is missing that we were referring to earlier is that because the awareness is is is lacking on the topic, we don't know the impact that we have. And if we start to line things together from an energy standpoint, carbon emission, water consumption and so on, then suddenly, we would've another lens to see the problem and other tools or reinforce the desire to implement the tools and the best practices that we all know that are good to implement. Right. So for me, that's where there is a huge opportunity for the PM. [00:16:49] Gaël DuezCan you indulge us in a structured way to deep dive into what are the big bunks of stuff or actions or tasks that you would advise as an overview to pay attention to when you're a product manager, because Antonia mentioned the data. She, and that's very rare in our industry, mentioned the hardware and making sure that we are not part of the planned or even not planned, but just technical obsolescence problem that our industry is rigged by. So, wow. Big kudos for that. And François , how would you structure everything because this is exactly what you've done in your playbook.[00:17:27] François BurraTo double down on what Antonia said, like the whole manufacturing, hardware, distribution, end of life and so on and so forth. That represents about two thirds of image emissions of the digital industry, right? So we cannot state enough in this podcast and in our lifetime, how big our impact, how big digital is physical. So to answer your question, Gaël, with a playbook that we co-wrote with Antoine Cabot, with a friend of mine working at Salesforce in BC, Canada. We structure it around five chapters, but maybe we can focus on some of them for the sake of the discussion. The first one is to, as a PM, embed climate or sustainability as part of your rituals and it's all about how you choose the right metrics, how you track your digital footprint, how you include your planet into your product requirements documents, your briefs and so on and so forth. So that's the first chapter. The second chapter is how you build in a more mindful way. So thinking about the planet as a Persona, animals as a Persona. How you try to tie your strategy to sustainability, sustainable development goals and so on and so forth and how you avoid obsolescence tactics and how you can leverage AI in a sustainable way because. As much as we want to make sustainability a trend for PMs, right now, AI is like leading the race by far. So we have to see what use from this momentum that they have to do the same for sustainable topics for PMs. So anyway, that was the second chapter. The third chapter is like how to apply frugal minimalist best practices. So coming back to what Antonia was saying, like how you minimize the amount of data that is being transferred. How you simplify and build a straightforward user journey that avoids all the fluff, remove product loads, you remove features that are not used, you remove data that are not used, you remove scripts and components and so on and so forth and don't forget to kill features. The fourth chapter is about how to be more carbon aware and or read aware, but just as a concept that normally we should make sure that we consume the least amount of energy, but also the low carbon energy as much as possible. And we migrate your server or your hosting provider to choose a sustainable one and things like that. And the fifth pillar, and I'm going to stop here because there's so many things that we can talk about about each of those topics is to leverage your influence. As we said earlier, we have a huge influence internally, we have a huge influence in our ecosystem, in our value chain, the way we choose our partner whether it's to transport our e-commerce website, let's say who are we partnering to deliver our product? Are we choosing a provider with sustainable practices or who uses electric vehicles or like bikes even for delivery. So like all of that is kind of how we can use our influence internally and externally. And I think this is a power that usually we have, but we don't realize that we have it. And there's a big miss opportunity if we don't leverage it to the full extent. So those are the five chapters. [00:21:03] Gaël DuezAnd thanks a lot for structuring it because now I'd like to play a little game with both of you, which is let's go for each chapter and pick one example that you would like to see widely adopted for each of these chapters. And of course, feel free to pick the chapters in whatever order you want. And Antonia, I might bet some money on which chapter will come first, but please feel free to start. [00:21:34] Antonia LandiI mean, honestly, there were so many things that resonated with me as you were talking, but I think first and foremost, and this is something that I stress over and over again, right? Sustainable product management is also just good product management, right? Don't build things nobody needs, right? Get rid of that bloat. I love that you mentioned minimalism. But I think to me, this really comes in two different forms. One of them is really actually how we structure the user experience of our products. Like, Netflix, if you let it, it will just keep playing. Like, YouTube will just keep streaming. Same with Spotify, right? They are built to keep streaming high quality, expensive data, right? Because that's how they make their money. But that has a massive, massive climate impact. [00:22:58] Gaël DuezAnd, I really love Antonia's idea of letting the user choose. And especially when you look at all these experiences with, as you mentioned, Netflix, YouTube, et cetera. In their values most of them is like do no harm provide the best possible experience when you say that one of the worst health crisis that the word is actually under and it's completely under the radar is a sleep crisis that we are literally destroying our health because with so much stimulated that we don't sleep enough in almost every country compared to the minimum requirement of our body. You could seriously challenge it if it's that much sustainable, offering people nonstop video and nonstop audio, at least without a message in some kind of action. I say, “Hey, by the way, you've just spent three hours.” It's 2 a. m. in the morning. Most of the people will go to bed at that time. Do you still want to like a bit like on the cigarette packet, but what was it your idea, Antonia? [00:24:07] Antonia LandiSo first of all, this also broadens the scope of sustainability, right? Is mental wellbeing part of sustainability as well? I'd like to think that's a part of it as well. But the second one, like now you're going into the... so I have made exactly this case of the climate impact of watching Netflix on your phone. Like, you download that episode, you're on a flight, whatever, watching Netflix on your phone. And two things usually happen, right? People are shocked at the impact of something so small and so mundane. And then people despair, right? Because it's like, well, I do this all day, every day. If we're in zoom calls, we're streaming data to each other just a whole entire time, right? What is the impact of that? But I think it's also realizing that yes, you might be contributing to the problem, right? Currently. And yes, there is this initial shock you have to go through to understand just how big the impact of our everyday actions with tech are. But all of this information, like all of us talking about this, isn't to make you feel bad about your daily habits. It's just to allow you to be more conscious in the future, to have that maybe sense of responsibility to have that just extra second where you think, well, hang on, do we really need this? Right. Can we give users an option? [00:25:48] Gaël DuezSo enable the user to pose and give them an option. I think it's a beautiful statement. François, do you want to go for the second round? What would be a very actionable idea or best practices that you'd like to share and where does it refer to in your framework, in your five chapters framework? [00:26:11] François BurraYeah, I actually wanted to react to what you were saying, because it was so interesting. I think Gaël, you were referring to Netflix and the need to sleep, but I think it was a Netflix CEO that actually said that their biggest competitor is sleep. So I think it shows how problematic it is and how we can talk about sustainability and responsibility, but we do have an impact in many different ways. And that we need to embody that in a more responsible way. And then when we talked about do every new employee needs to have a brand new computer, and what do we do with existing IT fleet and equipment that we handle? For companies that have hundreds or thousands of employees, just thinking about the lifespan of their equipment, if they push, like sometimes there are internal policies that you may have seen every three years you have to change your computer because you have to be the best performer. And, we want employees to be productive. So we're going to change it every three years, right? If you push that to maybe four years or five years, the environmental impact of that measure, that's kind of a simple quote unquote measure. Could be huge in the environment, considering what we talked about earlier about how much it costs the environment and the planet to build new devices and equipment. So that's 1 example that you can take as you could be the CTO in charge of that. There could be many people responsible for that type of decision, but this is super critical. And coming back to the next example, again, about autoplay, which is an environmental disaster, and only because of the example we just talked about, there's two practical ways you can fight against it. And I saw it with a client that I'm working with at the moment for the past few months, they are a media platform. And basically there's two ways you can approach it. Either you ban autoplay or you avoid it as much as possible, which is not necessarily the most acceptable way. But another way to look at it, and by no means it's perfect, right, is to instead of launching the video, the autoplay after three seconds, where the user basically doesn't have a choice because it goes so fast that it just happened, right? Maybe can you extend it by 5 seconds or 10 seconds, 15 seconds? So at least you give the user the control, the power. To reflect on that and maybe there is a chance also to educate the user. Hey, is it time to sleep? Or hey, do you know the environmental toll that streaming has? So the intention here is not to blame users and put the responsibility on them. It's us as digital creators and builders that need to build by default something that is green, right? Or sustainable. But there's so many ways that when you start understanding the problems and become an expert of the problem, and that's all about being a PM, right? You have to become an expert of the problem because before you become. Expert of the solution. So when you add sustainability into your framework, then suddenly you can have new creative solutions to embed into your product. And that's what I find fascinating. It's not like something more that you have to add to your day that is kind of bothering you and you have to make it fit. It's something that can unlock new possibilities and make your product better. [00:29:25] Gaël DuezIf I'm allowed to bounce back on what you say, François, you use the word framework, and I reckon an entire chapter is dedicated to culture, rituals, etc. And I know for you and Antonia, how pivotal this culture question is starting with you, François , what would be a good idea, a good practice to entrench the sustainability mindset in product management?[00:29:55] François BurraThere's different ways you can surface climate into your product. And we can start with OKRs. I think we were talking about OKRs before the show started. You can take those as opportunities, right? Either you have a company that is mature enough that you can be sustainable. Or climate related OKRs, like we want to reduce our footprint by 10 percent by the end of the year and then you break it down and you kind of see how you can reach that goal or you have non-climate OKRs, let's say, increased revenue or reduced costs, but when you talk about that, like when you break it down, reduced costs, probably like if you're a big tech company, cloud expenditure, it's probably a big center cost, right? Cost center, sorry. And why is that? Because you store a lot of data and oh, okay, so how can we reduce that data stored and, and see how we can be more intentional in the way we store what we need and what we, and not store what we don't need? And then suddenly like you become more aware of the lifecycle of data and then you can remove things that are not needed because as Antonia said earlier, like you don't need all the information about users that churn five years ago about what they did on a specific day, right? And so when you become a bit more intentional and minimalist and like more specific about why we are storing this? And why are we building this? And why do we keep this feature that serves no one suddenly like you will trim down your product, make it lighter, maybe make it faster, remove the fluff, trim images, compress and so on and so forth. And then, oh, well, we just reduce the cost of our cloud infrastructure by 5, 10, 20 percent and that feeds this. Okay. This is about the bottom line and we didn't talk about sustainability, right? We just did. What is what makes sense? But at the end of the day, we have a climate impact. So I think either you need a conversation with environmental topics, or you just find the opportunities within existing strategy, given that most of I don't want to overgeneralize most of sustainable and climate best practices for PM or for digital folks there. Is good for the business and the product, and there's other ways you can do it, but you can set carbon budget when you release a new feature to make sure that you track this you track and you make assumption and projection about how much a feature could, could wait in, in term of carbon and when it's released, like making sure that it doesn't exceed the sets, an amount of carbon emitted. So you stay within a range of an acceptable amount of carbon emitted. You can look at page weight. There's a page weight budget as well. There's tools out there that can help you to build pages that are not exciting. Let's say two meg or three meg per page. And that's a way that you can embed. The discussion with your folks, like definition of done and acceptance criteria with the designers and developers and start to make them part of the discussion and challenge them with them with this. And that should be something that excites them as well.[00:32:58] Antonia LandiYeah, I think for my side to add to that, like honestly, if you're really serious about this, make it a prioritization metric. Just like everything else, right? You, we talk about complexity, we talk about the size of the problem space, we talk about the projected ARR, right? If you've identified a really exciting new feature that might make you a million in ARR, and was terrible for the planet, are you still committing to doing that? And I think it is. We do have to retain a certain pragmatic approach in that everything's a trade off. Some things you might identify as an organization. We do not want to compromise on the specific part of the user experience, but we're going to do our best to do everything else everywhere else, right? It is right. It's not about despairing. It's not about we're doing everything wrong. It's really just about educating yourself, understanding your trade offs and then making them a reality, right? Like embedding them in the product management process, right? This is the future we've planned. This is how it's going to positively or negatively affect our climate bottom line.[00:34:14] Gaël DuezWe mentioned OKR. We mentioned how it needs to be ritualized in the definition of dawn, etcetera. I have a low ball for you because I'm still struggling with this one. And don't get me wrong. It's already super hard to get a key results around carbon budget, page weight, carbon footprint, you name it, whatever, but we know that if you're a frontend, I would say product manager and I'm struggling to find a proper key result that could be connected to an overall sustainably objective about how we make sure that we do not contribute to technical obsolescence. [00:35:02] Antonia LandiI think the perfect one is going to be difficult, but off the top of my head I would probably more likely look at the percentage of people included or excluded if you move up, let's say an iOS version, or if you move up a certain OS requirement and then it becomes about inclusivity as well, right? It becomes about how many people on the planet can use your product. And at that point, I don't know why people aren't thinking about this more, right? Like I used to work at the startup here in Berlin that had a fitness app and they purposefully wanted to exclude older versions because that was like a more premium thing now. Our app is only for people with new phones, right? But I think I would rather think about the percentage, right? Set a target percentage for the cement of people on the planet who can use our app rather than set specific targets for near specific versions, for example. [00:36:09] François BurraYeah, we tend to agree. I think we can look at versions being like, Oh, I covered the five latest versions of this OS and maybe I could go to five or like six or seven or eight, ten, right? Or we can just be a bit more diligent and thoughtful about, okay, who are our users? Like, how are they distributed amongst OS and versions and so on and so forth. And then, with this kind of idea, like percentage in mind and distribution in mind. Then we can, if we are delivering a good experience and we can be more specific about defining good for each company. But if we only cover a good experience for like 90% of our users, how can we make sure that the 10% others are also, even if they're run on old OS or equipment, then they can have the best version of the app. And if they can, because the device itself cannot support the most flashy features that you offer to the latest iPhone, then how you degrade or adapt rather the experience so that they could do everything that they need to do, but maybe in a more streamlined and minimalist way again, and I will go even beyond and that may be in the sense that by designing our software, we exclude people from the get go. So how can we go even by making our app or product even more accessible? We can increase our target market, right? Because suddenly if let's say we support a bank application, support transactions by SMS or by text messages, right? How much more we can unlock new customers and acquire a new target audience, right? So I think we also have to see the opportunities and not just be constrained by existing users. Let's think about non users as well. [00:38:02] Gaël DuezI'm very relieved François because I was afraid I was once again about to make a lecture on the survival bias and the fact that countlessly people were telling me this is our audiences is what we should focus on and I'm what was always like, you know, the survival by years, the plane, blah, blah, blah, World War II anyway, is there any one last darling best practices that you love to share, whether it's about culture, whether it's about onboarding your CEO. You name it.[00:38:30] Antonia LandiI mean, for me, like again, it's this ecosystem thinking, right? Like it's not just the digital products we build, it's how we build them. And, this is such a stupid example, but do you have oat milk or do you have almond milk in your office? Right? Like it takes a huge amount of water to produce an almond, which you then squeeze desperately trying to get something akin to milk for your vegan and lactose intolerant friends. What do you have? Reusable cups? Or do you have mugs? Where does the recycling go? Like even at this workshop I was giving a talk on sustainability and there was nowhere to recycle my paper cup. And that was awful to me. And it's just these little details that make up the entire environment of how we produce digital products that matters. [00:39:31] François BurraYeah, absolutely. As for me, I think as we stated at the beginning of this talk we have an issue right now that this is still the best kept secret, right? Like the impact of sustainability of digital products on the sustainability of the environment and the planet. So let's get people talking. There's plenty of people that care about this topic, and the more you dive in, the more experts you discover, and there's experts on sustainability everywhere around the world. So please invite some local folks to do talks at your company, and do a workshop about the environment like there is the digital collage that probably has folks around you that can come and speak about the interconnection between climate and sustainable products, digital products. Sorry. So I really started making this topic more broadly talked about. And maybe one thing that I really want to put a focus on by making your product emitting less emissions by using the least amount of energy you will get a product that will be faster that would perform better if you have an e-commerce platform, it would convert better as well, because the experience would be so smooth, regardless of the design, the device and the user satisfaction would be higher. Retention would be higher. And so suddenly with those product benefits that we're kind of starting to line up together, then they materialize into business benefits because your product will be differentiated. You would increase your margin. You will increase revenue. You would acquire more users, as we just said before. And considering the newest generation like the Gen Z and so on and so forth, they were raised and they are being raised with eco anxiety, so if your brand doesn't care about climate or he's using leveraging greenwashing tactics, it's likely that they probably some of them at least would not like to work for you. So from a talent attraction and retention standpoint as well as customer traction and retention, like climate needs to be part of your strategy. Like all jobs is a climate job and all companies are climate companies because you don't live in a bubble. We live on the same earth, right? So that's really something that I would love people to understand. We're not just like a pure environmentalist, you know, eating grass and feeding yourself with sun, right? There are ways that you can have the right value and serve your career, serve your business, serve your product and your users better. [00:42:02] Gaël DuezOne last question before we close the podcast. So we talked a lot about tactics at product management levels. We talked about OKR. We talked about the carbon budget. We talked about using some key results, the weight making sure that you can use your apps or your products on multiple equipment platforms. There is still something that I believe is related to what you said about leveraging your influence because product managers tend to have quite a lot of influence within a company. And it's how do you pitch your boss? How do you pitch your CEO? And, you've stated both of you that the sustainability momentum is getting traction, but it's not that big yet. And let's be honest, it's usually not among top business people that it's the biggest. And when you start traveling outside some places, you discover places where it's just simply not there at all. So how do you pitch your boss and do we always have this fight or this competition between sustainability in business and here I would like to set apart an unsustainable business model. Obviously, if you are drilling for more oil and that you have no transition plan and that you want 2 100 to do exactly the same business. You can build the cleanest possible digital products IT stack, et cetera. You've got a bigger issue than just this very specific part of your activity, but for, I would say regular economic activities, is it that much tension between sustainability of business? [00:43:41] Antonia LandiI mean, I think I'd go back to what François just said, right? Like it's actually just good business practice. You don't have to lead with sustainability, lead with lower costs, right? Client computing, especially like that, is always a huge bill. Lead with improved UX, right? Lead with more people who can use our product. And then by the way, this also means that we're lowering our footprint. By the way, this also means that we are now more mindful of what we store and how much we store and where we store it. And I've had to do this a lot doing, like, trying to get traction for mission important things under the guise of something else. But it's really just learning to speak the language that will resonate the most with your CEO, with your CTO, right? Getting that buy in and you can do it on such a small scale. You can do it for one feature, right? You can do it for, I don't know, the next AWS migration, because it's going to come sooner or later. I think almost using these valid business benefits as a Trojan horse to get people into the habit of thinking this way. And then honestly, before you know it, it is part of your culture. [00:45:03] François BurraAntonia said it perfectly. I don't know what I can add to it, but choose the right metrics that can resonate with them. Because like, as we know sustainable product best practices can really feed into many areas and avoid moralizing. Like no one wants to be on the wrong side of an argument and create for more. There's so many examples that we didn't share today, but like big companies and smaller ones that are doing great things in that area. So look at what's happening around them and create for more, because maybe your competitor is actually starting to implement those practices. And suddenly, people will start to freak out because they're like, okay, they're doing it. So we should do it. And also waves a bit of the flag of some climate risks that come with it. There's a lot of regulation in Europe. There's not that many in North America and we need those to come like GDPR. It just came in a wave and suddenly everyone adapted to it. And so there's also that part of the equation that maybe it would resonate better with some people. And we don't want just negative doom and gloom, but you know, you need to approach benefits, opportunities, and risks to be part of this question. So I think we have many tools basically to be able to convey that point. And as you said, Antonia, if sustainability isn't even part of the discussion, be my guest that's also perfectly fine. [00:46:21] Gaël DuezSo thanks François , because there were a lot of tips. Fear of missing out is pretty cool. Climate risk and fear of regulation is pretty cool as well. I think finding the right allies is so true. I mean, it really depends on some companies. That's so true that your CFO will be your best friend or the ESG director will be or sustainably director will be your best friend. So that's a very good point. My last question before asking the traditional closing question will be, do you need that much permission to become a climate active product manager? [00:46:54] François BurraIf I may start on this one, Antonia, it's kind of interesting because I'm giving training at the moment. And the 2 assignments that I give to my fellows are 1 looking at the playbook and the 33 best practices in the playbook and looking at the ones they can implement without permission. And the one they can implement and sometimes it's either there may be, they can influence both implement and influence, and sometimes they can only want to do one of them. So absolutely there's so many things that you can do that are part of your job and you don't need the approval of anyone. And the second assignment that I give to them is how to pitch them internally, which is what we just talked about. And basically helping them to create a slide deck to pitch this topic internally and see how we can be tied to the strategy of the company. So yeah, it's just so on point.[00:47:45] Gaël DuezSo you can say to your students, no need to use ChatGPT, just listen to the latest Green IO episode and you'll get a good grade. And what about you, Antonia? [00:47:57] Antonia LandiYeah. I mean, I'm all for asking for forgiveness rather than asking for permission. Even in product management, like there's so many of us that wouldn't be able to get to do the things we really want to do. Like, God forbid actually speak to some customers, right? You just need to find different ways of doing it. And like I say, use a Trojan horse if you need to go find it to get your job done without having to ask and don't even let anyone know. Right? Like once you have a strong enough case, that's when you can go talk about it. That's when you can go and say that, by the way, I've actually been doing this for the last month and it hasn't disrupted anything. So why don't we do this for all teams?I'm a huge, huge fan of this approach. [00:48:46] Gaël DuezThanks a lot. I love this approach as well. I'd like to ask you the final closing question, which is, would you share a positive piece of news regarding sustainability, maybe even sustainably in the IT sector? [00:49:01] Antonia LandiActually for me, the positive news is that this podcast is happening, that we're having this conversation, that I now got to know a whole other person who is so passionate about this topic. And I think this is the biggest marker of future success for me, because I think the more people talk about this, the more people with a meaningful voice talk about this and share really specific, actionable, real world strategies, the sooner we can get to a place where this is no longer a niche subject. [00:49:39] Gaël DuezGreat. I'm going to have a mainstream podcast now. I love it. [00:49:45] François BurraI wish. That's all I wish for your podcast. I think my positive news will be in the same sense of what Antonia just said, I think two years back, because I admit by pivot, like not long ago in all respect, you know, like it was two, three years ago. When I talked about it, I was the only one, like, especially if I talk about my context in Montreal, Canada or North America, like when I talk about climate and products, people were just looking at me like, what are you talking about? But now, people reach out to me so that I talk about this topic and it's only my personal story. But I see that the momentum is shifting in North America. Like there's more lights shining on this topic. And before I had to push the topic everywhere so that maybe someone will pick up on it and talk about it, agree to talk to me about it. Now people reach out to me to talk about this topic. And I would hope and I would assume that is the same thing for many other person in that ecosystem, which I think just shows and demonstrate that there there is a momentum happening and that the awareness is slowly picking up and hopefully with the awareness, people will start taking action, which is at the end of the day, what we what we need because t o come back to your introduction, like we need to start acting now and not tomorrow.[00:51:05] Gaël DuezThanks. It's great to see this momentum indeed. It was really cool to have you on the show today. There's very practical insights. I mean, I guess there are a lot of constants to be extracted by anyone working in the product management field. So I would say mission accomplished. Congratulations to both of you and I'm looking forward to continuing our discussion offline.[00:51:30] Antonia LandiIt was lovely being here. Thank you for the invitation. [00:51:33] François BurraThank you, Gaël. And great to meet you, Antonia. I think we have many more discussions to have together. [00:51:38] Antonia LandiAbsolutely.❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.  📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents. 
undefined
Apr 9, 2024 • 54min

#36 - Climate change challenges to data centers: lessons from Singapore with PS Lee

📈 44 cm water level rise under the IPCC business as usual scenario.This number shows that climate change is very real for Singaporeans and for their data centers, close to 10% of the whole of APAC! 🎧In episode 36, Gaël Duez discussed with Professor PS Lee, National University of Singapore Dean's Chair of Mechanical Engineering and one of the top experts worldwide on data center cooling, the challenges in making data centers sustainable.🔍Some key points of their exchange are:🌡️why temperature rise has multiple downside 🌡️energy challenges that affects the sustainability of data centers🌡️liquid cooling technology as an important option🌡️why lessons from DC operating in tropical climate apply almost everywhereAnd much more!❤️ Subscribe, follow, like, ... stay connected the way you want to never miss an episode!📧 Once a month, we deliver carefully curated news on digital sustainability packed with exclusive Green IO contents in your mailbox, subscribe to the Green IO newsletter here. 📣 Green IO next Conference is in Singapore on April 18th (use the voucher GREENIOVIP to get a free ticket) Learn more about our guest and connect: PS's LinkedInGaël's website Green IO website 📧 You can also send us an email at greenio@duez.com to share your feedback and suggest future guests or topics.   Poh Seng's sources and other references mentioned in this episode:Coastal Risk Screening Tool A Heat Dome Hits Virginia: One Data Center's StorySingapore Climate 2023: The Year in NumbersWorld Without End by Christophe Blain and Jean-Marc JancoviciFuture of the human climate nicheTan Tin WeeSUSTAINABLE TROPICAL DATA CENTRE TESTBEDEED TranscriptGaël Duez 00:21Hello everyone. Welcome to Green IO with Gaël Duez. That's me. Green IO is a podcast for responsible technologists building a greener digital world one byte at a time, twice a month on a Tuesday. All guests from across the globe share insights, tools and alternative approaches enabling people within the tech sector and beyond to boost digital sustainability. Because accessible and transparent information is in the DNA of Green IO, all the references mentioned in this episode, as well as the transcript will be in the show notes both on your podcast platform and on our website greenio.tech when it comes to sustainability, I have a sweet spot for Singapore because of its uniqueness. This is one of the top cities in the entire world which has benefited the most from the global extractivist, highly carbonized and financialized economy, and also one of its most at risk of climate change. Hence a blossoming of initiatives there on how to both pivot towards more sustainability and mitigate climate change impacts. And these efforts apply especially to the backbone of its infrastructure data centers. As we are growing aware that the human body has some physiological limits that can be reached during heat waves where the wrong mix of high temperature and humidity is reached. It's death for people staying outside too long and this concerns everyone, not only the populations usually at risk like infants or elderly people. It was really well illustrated by Jancovici Jean-Marc and Blain Christophe in their international bestseller comics World Without End and based on a 2020 mega study published in Environmental Sciences, a third of humankind is now at risk to live in places where temperature could be lethal several weeks per year. Well, everything that I just said about humans applies to data center equipment which also needs to be cooled down 24/7 for many obvious reasons. Starting with latency and sovereignty, we cannot move all data centers in the world where the air will remain cool like the Nordics. Hence a serious challenge for the tech industry, how to run a sustainable data center and more specifically how to build sustainable data center hubs where tropical climate creates hurdles which are getting bigger due to climate change. To discuss this, I'm honored to have Professor P. S. Lee on the show. Based in Singapore, PS is National University of Singapore Dean's Chair of Mechanical Engineering and one of the most cited scientists in mechanical engineering and transport energy rings worldwide. He has specialized in data center engineering for two decades and he is also a field practitioner. The founder of Coolers DC, what a cool name, which advises top DC operators like Equinix or Meta. He will also be our keynote speaker at the Green IO Singapore Conference in two weeks, April 18 to be precise. One reason among many reasons to join the first conference in Asia focusing 100% on green it and Green IO. Listeners can get free tickets using the Voucher GREENIOVIP and how exciting it is to kickstart my discussion with Professor Lee. Welcome, Poh Seng. Thanks a lot for joining Green IO today.PS Lee 03:54Hi guys, thanks for having me. So it's my pleasure to be on your podcast.Gaël Duez 03:59So PS, I played again with the excellent Climate Central's coastal risk screening tool, which enables people to simulate the impact of a sea level rise, among other impacts. And the results were already concerning for a 44 cm water level rise, which is the current IPCC best estimator for the business as usual scenario with most of the iconic gardens by the bay in Singapore under the water. And if we simulate a 2 meters rise, a likely scenario. If some tipping points with ice in Greenland or Antarctica are reached, well, the entire Singapore's harbor is at risk. My point is, climate change is very real for Singaporeans. Could you maybe explain to us why and what is more precisely at stake here?PS Lee 04:53Sure. I would like to refer to the recently published 2023 Singapore Climate Reports, which marks an alarming continuation of global warming trends, with Singapore experiencing its fourth warmest year since unprecedented temperatures, especially during the months of May and October. This underlined the urgency of addressing climate change. The report's projection of up to 326 high heat stress days by 2099 in high emission scenarios starkly highlights the impending challenges, especially for industries like the IT sector.Gaël Duez 05:33Correct me if I'm wrong, but Singapore is a very massive IT hub in Southeast Asia, is that correct?PS Lee 05:42Yes. As of two to three years ago, in Singapore, the total data center capacity is actually close to 1, that is close to the 10% of the whole of APAC. So for city states to be hosting that kind of capacity is actually quite amazing, but it's actually important to manage the power and the associated carbon footprint. So it is. Right now, the data center industry is already consuming 7% of Singapore total electricity consumption. And if we don't manage this in a couple of years, this can actually go up to 12%. So that's why the Singapore government has actually imposed a data center moratorium about two and a half to three years ago, which they finally left the year before. Thereafter, there was a data center call for application, which the industry expected to meet very stringent criteria, including PUE, no more than 1.3, as well as the adoption of innovative and sustainable data center solutions.Gaël Duez 06:59You mean that for two years and a half it was not permitted to build new data center facilities in Singapore.PS Lee 07:07That's right.Gaël Duez 07:08And just to make sure I understood well it's already 7% of electricity consumption and it could triple with the current trend.PS Lee 07:17Yes and especially with the interest in AI in other high power workloads. So if we don't manage this in a sustainable fashion, this percentage is certainly going to increase very sharply over the next couple of years.Gaël Duez 07:35So clearly the Singaporean government doesn't want to go the Irish way with the Irish scenario. Where I recall today the electricity consumed by the data center industry is already above 20% if not 25%. I don't recall the number exactly but it was really mind blowing. And so you were mentioning this very strict criteria to build a data center in Singapore, obviously low carbon electricity and also innovative solutions. Maybe it would be interesting to unpack the challenges that data center operators in Singapore and more largely in tropical climates are facing.PS Lee 08:19Yeah sure. I think the first and foremost operating data centers here in the tropics with a high ambient temperature and humidity is actually going to be a very energy sapping exercise because of the constant cooling needs in order to maintain the IT equipment within acceptable temperature range in order to ensure proper and reliable operations. But associated with the heat stress that we're experiencing during the recent few years due to climate change, this has actually imposed a more challenging condition for data center operators to ensure the resilience as well as efficient operation of data center here in the tropics and specifically here in Singapore. So that's definitely going to be they are constantly looking for more energy efficient cooling solution and this goes beyond operational efficiency, rethinking energy sourcing as well and potentially integrating renewable energy sources to mitigate the carbon footprint. And for city state like Singapore in terms of domestic generation of renewable energy is going to be limited as well. I think the only viable renewable energy source here is actually solar and based on a report published by the Solar Energy Research Institute of Singapore series which is hosted here at NUS. So even if we were to sort of use all our rooftop space for the solar panels, so the percentage of the electric supply that comes from solar is actually still going to be very limited, well less than 10%. So that's why I think Singapore is actually adopting innovative approaches whereby we're actually cutting agreements with our Asian neighbors to look into the imports of green electrons, low carbon electricity. Then I think related to the temperature challenge or the temperature or what we call a thermal management challenge associated with operating data center here, there's also this wear and tear on it equipment because the increased heat stress days will directly translate to accelerated depreciation of critical data center equipment. And this necessitates more effective maintenance and proactive design consideration that can withstand the rigor of a harsher climate, and again, related to a climate. So when it comes to heat rejection, typically we need to operate the cooling towers that actually consume water, then that brings about the water sustainability issues. So given the intensity or the intensification of the extreme weather events. So water cooling, the technologies, while efficient, may actually impose a sustainability challenge. Hence, we are always actually looking at better ways of reusing water recycling along with explorations of air and more efficient liquid cooling alternatives. And these are now becoming imperative. But we are also grabbing with, for example, the challenge of a skilled workforce, especially if you look across the region, Southeast Asia. So the data centers industry is actually booming. So how do we actually adapt and mitigate the impacts of climate change? Data center also means nurturing a workforce which is proficient in engineering and understanding sustainability practices and technology. So there are definitely quite a few challenges that we are confronting today in order to ensure the sustainable growth of the IT industry.Gaël Duez 12:42If I follow you here, in trying to wrap up what you say, I've listed four challenges. You've got the energy challenge, the equipment challenge, the water challenge, and I would say the workforce challenge.PS Lee 12:56Yes. These are the exact four points that.Gaël Duez 12:59I mentioned, and I think it will be worse deep diving on all of them. So if we start with the energy challenge, I've got a first question. So you mentioned solar, and solar not being able to attend more than to contribute to more than 10% of the overall electricity production. But what about wind energy? Is there any ability to put wind from around Singapore or is it just something completely nuts because of all the boat traffic?PS Lee 13:33Yeah. So Singapore is what we call a renewable energy challenge country, right? So when it comes to the wind, unfortunately, our wind speed is actually on the low side. If you were to put in the wind turbines, whatnot, right. In terms of the energy that this can generate will be very limited, right. Because of the low wind speed. Hence comparing the different options, solar is actually the only viable renewable energy option. Although in the recent one, two years, there have been a lot of interest in assessing the potential for geothermal. And there's also talk about going into nuclear, although I think the nuclear question has not been fully addressed because we are in CD states. So certainly the public perceptions, the safety related issues will have to be adequately addressed before Singapore to take a position. So we are very much constrained in terms of renewable energy options.Gaël Duez 14:47I think I'm not 100% sure that regarding this energy innovative approach you mentioned, which is basically outsourcing the production of low carbon electricity, I think you're the one who mentioned it as a geopolitical problem. Could you elaborate a bit on what are the pros and cons of sourcing your low carbon energy elsewhere than on your own territory?PS Lee 15:13So sourcing, renewable, the energy or the low carbon electricity from for example our neighbors. So that actually requires the cutting of bilateral agreements. And if let's suppose the electricity is actually to pass through in more than one country, then that actually requires multilateral agreement. Hence the political issues will then come into the picture. But I think Singapore has been actually very forward looking and over the past year or so, in fact we have signed an agreement totaling more than four gigawatts of clean electricity imports. But obviously the data was in the details. For example, what will be the price that would be paying for such clean electricity? The imports, because our other neighbors also have their net zero commitments. So there are actually a lot of issues. But the one thing is clear, I think once the Singapore government decides on the course of action, they will actually put in all the necessary efforts to realize them. So I feel hopeful that it will be a win-win arrangement. For example, we will collaborate deeply with our asean neighbor so as to sort of increase the deployment of renewable energy projects so that when they have access they can actually look into exporting some of these clean electricity to Singapore. But it may actually require some sort of differential pricing. For example, maybe for their domestic demand the rate will be kept at the lower level, but for those that they are exporting to Singapore, that may actually incur a certain premium.Gaël Duez 17:10What about the infrastructure? I mean, you need a connection cable. Are they already there? Do you plan to construct more?PS Lee 17:16Yes, I think the associated grid infrastructure will be actually critical when we look at linking up the grid that cuts across different countries. For one, I think there may be differences in terms of some of the standards. I think this is something that Singapore will have to work with the neighbors to finance the grid infrastructure and that may actually require quite a big amount of upgrading or the expansion of the existing grid. So I think this will likely be a costly exercise. But I think if this will actually be a win-win arrangement, by adopting a collaborative approach, we can actually increase the renewable generation capacity.Gaël Duez 18:05And if I run a data center facility in Singapore, how does it work concretely for me? Is it more okay, my government is in charge of decarbonizing the electricity, so I consume the electricity that I've got in the grid. That's it, period. Or do they also have to close a deal like purchasing power agreement or some sort of directly with foreign countries?PS Lee 18:32So I think this is the part that is still not quite clear yet. So for example, can you achieve the data center operators, can they cut a direct view, for example, power purchase agreement with renewable generation plants, for example, in Indonesia? So this is the part that is still not quite clear because the agreement that has been cut is still country to country. But I think there's actually a lot of interest from the data center operators to have access to as much green electrons as possible. But the details have not been ironed out yet. But I suppose it will be done in a phase approach. First, country to country, the agreement has to be done then that can actually trigger down to the various sectors. Certainly. I think data centers are one industry that has strong demand for the low carbon electricity. But there are also other sectors, right? So I think the Singapore government will then have to figure out a way. Then when they import such, the green electrons, which sector should it go to? Is it the data center industry or is it the semiconductor industry? Or would they actually allow, for example, the sector to actually cut direct agreement with the renewable, the generation, the plan overseas? So this is not clear yet. So I suppose over the next two to three years, right. There will be actually more clarity. But I think what is clear, Singapore government has always actually had a very consultative approach. So I'm sure they will be actually reaching out to industry, including data centers, to have the dialogue and to sort of come up with a framework, right, such that it will actually benefit, for example, data center operators who have the desire to steeply decarbonize their data center operations.Gaël Duez 20:41And eventually, because I believe the Singaporean electricity grid is fully unified, it's a theoretical question, or at least it's a bit like an analytical accounting question, because an electron is an electron. It's not green or per se, but it's more like between the different business interests, which business interests can claim, between the semiconductor industry, building and heating and or heating, not that much, but cooling. And in the data center industry, which industry will contribute the most to the low carbon electricity sourcing effort? But still. So how many data centers and how many operators are we talking about? Is it a very concentrated market with a handful of firms running larger facilities, or do you still have a lot of smaller or medium sized data centers run by, you know, I don't know, institutions, big companies, etc.PS Lee 21:40I think it's a mixed bag. So for Singapore, the data center industry, it has been reported that there are between 70 to 80 data centers operating here in Singapore. So you look at some of the more recent announcements, in particular, meta -150 mw, obviously it's huge, then, even since they actually consolidated some of the data centers. So they are actually probably in the range of 70 to 80 megawatt. So again, very sizable. And I think there's also the Google, the Microsoft, but certainly there are also the smaller data centers. That's probably in the range of maybe between 10 to 20 megawatt. But I think moving forward, for a mature market like Singapore. We will have to look at the edge data centers, which are smaller in capacity. Maybe we should be actually looking at anywhere between five to ten to 20 maximum in the capacity. And for those data centers, they are handling very high power workloads. For example, if you are doing AI training, I think it makes more sense to actually have this in, for example, Malaysia, in Indonesia, because they have abundance of renewable energy potential, they can then obviously have the green electrons to offset the carbon footprint associated with the very high capacity data centers. So I'm actually very hopeful that the various parties, various countries come together to set up what I call the sustainable data network, whereby we actually work in a very collaborative fashion so as to more effectively manage the carbon footprint of the entire Southeast Asia region.Gaël Duez 23:40Because you're an expert on these topics, whether my electricity is low carbon or not, I think the question of reducing the electricity bill is pivotal here. I would like to ask two questions that are very closely interlinked, as far as I know. How should we build data centers today, or refurbish them, and how should we run them to make sure that we save or we reduce as much as possible, or energy consumption.PS Lee 24:14I think certainly you make an excellent point. There's this constant desire to have access to more green electrons. I think this has to be peltrapped with energy, the efficient technologies specifically addressing the high energy, the consumption associated with operating cooling systems for data center, suddenly we can actually look at innovative cooling technologies, for example, the exploration or the adoption of single and two phase directorship and immersion cooling systems. So these are actually at the forefront of reducing the data center energy use and improving the cooling efficiency. And there's actually another wonderful side benefit that is less reported associated with the use of high efficiency, the liquid cooling that actually allows your hardware to actually operate at optimal efficiency. So, meaning that you'll be able to get the optimal workload accomplished versus, for example, when you are operating your IT equipment under the air cooling mode, which often goes into suggestion like thermal throttling, meaning that you're not able to fully maximize in terms of the performance. So I think adoption of innovative cooling technologies is certainly one of the first things that we want to do in terms of improving the energy efficiency of data center operations.Gaël Duez 25:58Just regarding this water cooling approach, I've got a question, which is how sustainable it is. And my question comes with two faces, and we will talk about the water consumption later. The first, is it mature enough or is it still R & D? Because last time I checked, the moment you put metals in water, bad things happen, corrosion and so on. And that will be my very minimalistic contribution in terms of chemistry. But how much is it like R and D and a bit of hype, or how sustainable it is? And my second question is how big the investment, environmentally speaking. But it also comes with a financial cost, obviously, when you refactor or you refurbish your factory to enable water cooling, because you could have a lot of carbon embodied with the data center facility itself. So hence my two questions, which are two sides of the same coin regarding the sustainability of water cooling technology.PS Lee 26:59So liquid cooling is not new. It has actually been around for a decade. It's just that the data center, being a recent adverse industry, didn't quite adopt it in a big way. But I think we are actually getting to a stage whereby, because of the increase in thermal design power, because of the sustainability imperative, I think we are getting to this stage whereby liquid cooling really needs to be a serious option. And it's certainly well beyond R & D. So to give an example, we actually started a company called CoolDC. We started off with a test bedding at one of the major co-location operators in one of their production data halls. And shortly after the completion of the test vetting, we actually managed to secure a project with a major local bank. And they are actually looking at implementing liquid cooling for 16 racks and this handling production workload. So certainly I think it's actually ready for the big time. But the challenge is actually how do you come up with a system that can scale with demand, when talking about, for example, high power IT equipment, it will not happen overnight, there will be a scaling up over time. So then the challenge is actually how do we configure a cooling system that can scale with demand in a very cost effective fashion, in a fashion that minimizes the disruption to ongoing operations. So I think there are actually ways to do it, because without liquid cooling, in fact there are actually different configurations. It can be air assisted, liquid cooling, which can be pronged into an existing brown fuel air cooled data centers without laying the elaborate piping network. But then obviously, going back to your second question, then what about the impact in terms of the environmental impact, for example, in terms of embodied carbon, so should you, for example, I have to set up new data centers. So again, I think it really depends, right? So we're very innovative in design. In fact, you can retrofit, right? Brownfield Data Centers to be very energy efficient. And that major local bank, the client that I mentioned just now, is actually in fact retrofitting one of their existing data halls. So as to support the equivalent  that can be done. It's just that the engagement, the dialogue with the infrastructure folks and the IT team has to happen at an early stage. So as to minimize the disruption because if after talk, then obviously that will potentially lead to the interruption to the IT operation, but then if, for example, you start the dialogue early, you know, when the IT team is planning to have the next round of hardware refresh, it refresh. If you time the upgrading or the retrofitting of the infrastructure to support liquid cooling in line with your IT refresh cycle. I think that can be done with very minimal disruption. And the other initiative that I have been spearheading, the sustainable tropical data center. In fact, what we did is actually we retrofitted a 40 plus year old power substation into a very efficient data center. That's where, so moving forward, that likely will be the model. Because for one, you want to save on embodied carbon. And the fact that for a mature market like Singapore, we are actually running short of power. We are also running short of space. We can't keep continuing to have new build data centers. But the fact that you see we have a lot of existing building stock, I think there's one perception that liquid cooling may actually lead to increased water consumption. That's not really the case because in fact, when talking about direct to chip or emergent cooling, this can be fully the closed system. For example, instead of rejecting the heat through a conventional cooling tower, which is obviously going to consume water, you can actually perform the heat rejection using a dry cooler. In fact, we have actually demonstrated this at that cold location data center that I mentioned. We actually demonstrated that you can actually simply reject the heat using a dry cooler without consuming the water. So again, really there's a lot of variance in terms of how you configure a liquid cooling system. So I think it really needs to be a very early engagement between the infrastructure team, the IT team, as well as the solution provider, so that you can actually come up with a configuration, a solution that is really fit for purpose because you don't want to over design or under design. Then we need to adopt a holistic approach whereby we actually factor in what's the impact in terms of the embodied carbon, what's the impact in terms of the water consumption. Then obviously, what's the impact in terms of the power or the energy consumption, which is typically measured in terms of power. But moving forward, I think it has to be really a very collaborative approach, and the dialogue needs to actually start as early as possible and not as an afterthought.Gaël Duez 32:49I've got a question for you regarding what you said about water consumption, which was one of the four challenges that you've listed. Where do these big headlines about water consumption of main hyperscalers, at least western ones, come from if water cooling doesn't consume that much water? Because, you know, I mean, all of them, Google, Azure, AWS, they've been blamed for consuming a lot of water. So where is the issue here? Because you tell us that, well, it doesn't, you can run a closed system and it doesn't consume water, or at least marginally so can you enlighten us a bit here?PS Lee 33:32One option, as I mentioned just now, is actually to use, for example, a dry cooler instead of cooling the tower, thereby having a fully closed system without actually consuming water. But then obviously there's a trade off, right? So if you operate a dry cooler versus a cooling tower in terms of the PUE, in terms of efficiency, right? The former, without the consuming water, will be actually less efficient, somewhat less efficient then in terms of the footprint. So the space required for a dry cooler versus cooling tower is actually going to be bigger. So I think it really has to be a holistic assessment. So what makes sense? Do you want to go, for example, the best possible PUE, or do you want to have a more balanced so called matrix involving not only the PUE but also the weight? So I think that's probably what's necessary moving forward. So for data center operators, including hyperscalers, to look at things in a more holistic fashion. But then I think the fact is still hyperscalers actually have very high capacity data centers. So then that obviously will translate to a large carbon footprint. Yeah. So I think that's something actually unavoidable in certain sense, but I think as much as possible we will have to then maybe look into the transition from air cooling to cooling because that will bring about a reduction in terms of the total energy consumption. So I think that's something that we do, but I think the other aspect that can be done is actually the end user. So we also need to be probably more prudent in terms of our consumption of digital services. So maybe you want to limit how much TikTok or YouTube that you watch each day.Gaël Duez 35:27It's a bit like the elephant in the room. Is sobriety like digital sobriety a topic at all in Singapore or in Southeast Asia, as far as you can tell?PS Lee 35:37I think people, obviously so here in the news, in social media, that we need to be more sustainable. Carbon emission is definitely going to be something challenging to manage, but it's just that I think some of the things that don't actually trigger down to the individual, to the personal level. So moving forward, we may really need to consider imposing a personal carbon budget so that you're more conscious in terms of how much digital services that you consume. I think it's all part of being a responsible citizen. Everyone obviously is conscious of climate change. We really need to take it upon ourselves so that we are also very conscious in terms of managing our consumption. I think that certainly will be very complimentary to all the wonderful energy efficient technologies that we are deploying in our data centers. Very complementary to the integration of renewable energy. But if, let's suppose we can in tandem manage the consumption while actually managing in terms of the energy efficiency improvement, integration, and renewable energy, I think all in all, that will make the earth more sustainable.Gaël Duez 37:13The example you gave about making people more aware of the need to refrain from digital consumption, I think it connects pretty well with what you've said several times before that. Dialogue is key. And making sure that all stakeholders start discussing on how to make the data center industry more sustainable. And my question is that a Singapore way of doing things like, it seems to be pretty much in the DNA of Singapore business and governments to talk to each other. I don't know if it includes consumers as well as you've just mentioned right now, and how much of the government is involved and how much more specifically does it use the stick or the carrot.PS Lee 38:07So I can frame it as dialogue and collaboration. So I think Singapore has always been very collaborative. I think one, the key reason being that Singapore is actually a migrant society. We have a very relatively short history. So in our 50 years of nation building, I think Singapore achieved a lot. And I think that happened because we are able to rally people coming from a different background, different races, different religions to come together in a very collaborative fashion, then obviously you see collaboration between government, industry, academia, civil society. So I think that really is very much needed when we want to actually address the multifaceted challenges associated with sustainability, associated with climate change. I think the Sierra Leone government probably will soften, or rather usually will first go with the carrot before they bring the stick. I think really helping the industry, helping the public to see the needs to be more sustainable and the need to be more collaborative, I think that will be actually more effective than, for example, imposing the very draconian measures. I'm still hopeful that Singapore will be able to focus more on inculcating or enculturation. They need to be sustainable, they need to be collaborative so that people come together willingly. Because I think that will be more effective than forcing people to comply with certain regulations. I thought that should be probably the last resort but I suppose it needs to be a balance, because obviously there will still be companies that are non compliant. Certain amounts of regulatory advancement will still be necessary.Gaël Duez 40:20And I'm asking you also the question, because I'm pretty sure you're familiar with what is going on in the rest of the world and in Europe. Very recently, an energy efficiency directive has been launched and it was pretty precise regarding the metrics that a data center should now report above a certain capacity. And it goes way beyond just renewable energy and PUE or we also waste use, etc. Do we have this kind of reporting requirements planned in Singapore, in other areas in Southeast Asia, for instance, or nothing to your knowledge?PS Lee 41:02Not to my knowledge, but at some point we may have to do the same thing. Going back to our earlier discussion that moving forward for Singapore especially, we will need to be very selective in terms of the kind of data center, the kind of workload that we are hosting. Obviously, we are very energy constrained. We have very limited options when it comes to renewable energy sources. So I think at some point we will probably need to mandate that the data center operators will actually need to report various matrixes, for example, maybe pertaining to PUE or WUE, what's the amount of renewable energy, the sources that are integrating into their operations, things like that. But I think it will take a while. I think at least for now, I see Singapore, the government actually engaging industry, encouraging them to go through the green transition. And I think that is something that will take some time. But I'm sure the Sambo government actually provides the necessary support, for example, in terms of grants, in terms of assistance, in terms of technology, so that the existing stock of brownfield data centers, especially those that are quite dated, right, can then go through this green transition.Gaël Duez 42:36Now I'd like to zoom out a bit and talk about resiliency and the lessons for the rest of the world, because tropical, or you meet subtropical climates, are pretty widespread around the globe. And for example, in the state of Virginia, home of AWS, this is a subtract tropical climate. And I don't know Pierce if you read it, but in 2022, the reviewer, security technology published a fictional story about a heat dome descending in the summer of 2025. So it's next year near the town of Ashburn, Virginia. Ashburn is called ‘data center alley’ by its folks in the US, because it's by far the largest concentration of data centers, not just in the United States, but in the entire world. And spoiler alert, it ends with a data center manager having to shut down an entire hyper scaler facility facing the cornelian choice of either losing billions of dollars in equipment or entering its history with the first ever cloud blackout. And everything you said in this interview makes this story some sort of more likely because you describe the worsening climate condition that Singapore faces, all the challenges to cool down data centers, etcetera. And my question is: how realistic it is, and what are the lessons for pretty much everyone running a data center facility in the world from the specific conditions that operators in subtropical areas, the one who are the most at risk of heatwave? Well, what could be these lessons? PS Lee 44:30I think certainly the increasing episode of heatwave is going to pose challenges to the data center industry, especially if they are still using conventional air cooling systems. So I think this is part of that kind of dovetails here very nicely into why I think liquid cooling is actually the way to go because one of the studies that we did, which I thought is pretty interesting. So what we did is actually we increased the supply air temperature as well as the supplied liquid temperature, because we actually have two racks, one which is actually an air cool rack, the other is actually the liquid cool rack, both having the exact same IT configuration. So the only difference is actually the cooling method. So what we have shown is actually for the air cooled rack, the server's performance, the energy consumption, it's actually very strong functions of temperature, because when you're operating your IT equipment using air cooling, all your chip temperature, or what we call junction temperature, are actually much higher, typically in the range of 80 or 90 degrees celsius. And that is actually very close to the temperature threshold, because you want to protect the electronics, [for] the server you usually set the upper temperature limit. So once you cross that limit, then the thermal throttling skips in because you want to protect your electronics. But when you actually operate your liquid cool rack, even with an increase in supply temperature, what we notice actually both the performance as well as the energy consumption actually stays relatively insensitive to temperature. So the implication is actually what if you have a liquid cooling system, or rather if you are using a liquid cooling system, even if, for example, there's actually an increase in your outdoor temperature, ambient temperature, whatnot, that's not going to lead to a big issue in terms of the performance of your IT equipment in terms of the energy consumption by the demand. But if you were to actually use the conventional air cooling, then you'll run into serious problems because of the increase in your ambient temperature that actually leads to your ip, the hardware going into more frequent thermal throttling, which is obviously going to affect the performance. That's also going to mean that your equipment now is going to have a higher power penalty. So I think from the resilience standpoint, I think it makes a lot of sense for the industry to actually look into transitioning from air to liquid cooling. So that's actually my personal experience that I can share when it comes to the data center resiliency from the angle of cooling.Gaël Duez 47:48However, you mentioned earlier that transitioning a full entire facility from air cooling to water cooling comes at a price, and you were advocating more for a modular approach. So if, how much would it cost for a per scaler to migrate entirely air cooling to water cooling? I think it will be ace watering without bad word play.PS Lee 48:15It's true. If you have software looking at retrofitting and air cooled data centers to liquid cool, and suddenly there will be the additional CapEx. The fact that you have a sunken investment on your air cooling infrastructure and for you to replace that with liquid cooling infrastructure, certainly there will be additional CapEx, but it may, you may have no choice, right? If let's suppose your tenants come to you and say, I'm going to bring in one rack of GPU service because I'm going to run generative AI workload. If you don't have the supporting liquid cooling infrastructure, there's no way that your tenants can actually operate his high power servers or his high power rack. Even if they can operate it, they wouldn't be able to get the performance that they paid for. So the multiple factors that you need to consider one is actually, what's the cost, what's the ROI the other is actually what is the business opportunity that you don't do it. Because if you can't support the high power rack, high power equipment, your tenants will take the business and go to your competitors. So I think it really requires a holistic assessment. Certainly ROI is important, but at the same time, do you want to future proof your data centers so that you're able to handle the current, the future workload that's going to come into your data centers? So I think operators probably don't really have too many options. If you don't do it, you are going to fall behind your competitors.Gaël Duez 50:05So when economics incentives meet sustainability incentives, I think that's a nice way to wrap up the entire episode. Professor, before we leave this episode, I've got one final question. Would you like to share a positive piece of news with us? Whether it comes to sustainability in general or more specifically, its sustainability?PS Lee 50:32Sure. Traditionally, the Southeast Asia data center markets are not known to be trailblazers in terms of adoption of the most advanced, most innovative solution. But because of the confluence of various factors, including the increase in thermal design power achieved because of the sustainability imperative. So over the last 12 to 24 months, we have been seeing a lot of interesting developments. For example, data center markets in Malaysia and Indonesia are starting to sort of call for liquid cooling infrastructure. So I think that can propel the region, in fact, to be the leaders when it comes to adoption of sustainable and innovative data center solutions. So I'm actually very hopeful that the region can achieve leadership in terms of embracing the most sustainable and the most innovative solution. But going back to one of the key discussion points, it has to be a highly collaborative approach.Gaël Duez 51:42Yeah, that was the key word here, collaboration. And we've got a lot to learn from a small but very powerful state who has to work well with all its neighbors and to source electricity and to find innovative solutions. Thanks a lot. PS, that was lovely to have you on the show. I'm eager to see you on stage for Green IO Singapore. I hope that many listeners based in Singapore and maybe some of them based in Southeast Asia will come as well. Thanks a lot again and talk to you very soon.PS Lee 52:16Sure. Thanks for having me.Gaël Duez 52:19Thank you for listening to this Green IO episode. If you enjoyed it, share it and give us five stars on Apple or Spotify. We are an independent media relying solely on you to get more listeners. Plus, it will give our little team Jill, Meibel, Tani and I a nice booster. In our next episode, we will talk about product management and more specifically, how to be a climate conscious product manager with Antonia Landi, a leading voice in the european product ops community, and François Burra, co-author of the Climate product management Playbook. Stay tuned. Green IO is a podcast and much more. So visit Greenio Dot Tech to subscribe to our free monthly newsletter, read the latest articles on our blog, and check the conferences we organize across the globe. The next one is in Singapore, but you already know it on April 18, and you can get a free ticket using the Voucher GREENIOVIP and you already knew it. What you might not know is that early bird tickets for London on September 19 are already for sale. And what you might not already know is that we opened the call for speakers for London, so feel free to apply if you've got something interesting to say regarding it, sustainability, whether it's cloud data, operation design, etcetera. I'm looking forward to meeting you there to help you fellow responsible technologists build a greener digital world one bite at a time.❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.  📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents. 

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app