Why Emissions Are Your Next Massive Technical Debt

Hello everyone. I'm very glad to be here. Uh twothirds of the world's disasters are natural disasters. earthquake is landslide storm floods extreme heat. In 2022 here in the UK because of SRM heat waves two major NHS hospitals suffer IT failures and these impacted patient care and natural disaster are getting worse, more frequent, more in intense and they are affecting everything including digital assistance. According to Paris agreement, this will get worse if we don't limit the global warming to 1. 5°C. And to do that, we need to reduce by 45% by 2030 30. Yes. 25 45% by 2030 and we need to reach net zero global emission by 2050 and we have a problem. Do you know why we have a problem? We have a problem because the tech sectors emissions is still increasing. The tech sectors is 6% of the global emissions. When I talk about emissions, I'm talking about greenhouse gas emissions. This surpasses the whole airline industry. And because of AI, emissions are estimated to be to triple by 2030. We have to do something. If not, we won't meet Paris goals and natural disaster will become uncontrollable. I wish I knew all that when I started working uh with green that is four years ago. I was the chief technical officer of Luni, a hardware scaleup in France, the French market leader for uh kids toys in French when I led United Transformation. I wish I learned that four years ago and I'm currently in my mandate as uh the chief technical officer uh at BB France digital. I'm still working on those issues because it's our central role as tech to do something. If not, you won't reach the 5. 5 1. 5 degrees Celsius uh limit. Okay. And today I want to share with you and the seven lessons I learned the hard way to start to get started strong and avoid common traps. Let's go. Before starting one thing I want to make sure I think is to be all aware that we are now building tech in a new operating context. an operating context of a destabilized world and this is our new uh design constraint. Emissions are design constraints. Planet boundaries are design constraints and regulation are also a design constraint. So let's start with the first lesson where to begin right assessment is very challenging. When we started at Luning we have no clue where to begin. We were a hardware company. We make toys connected toys for kids. We published audio books as well and we are operating in Europe and starting to operate in North America. As we don't have a clue, we bring in we brought in um a sustainability expert and you have a lot of concept to learn. It's not overwhelming. It's okay. The

first concept is LCA life cycle assessment. It is the standard method to get started with greeney not only green it with sustainability. So you just um review your whole supply chain from raw materials manufacturing usage and beyond to evaluate your emissions. This is the start you have to learn also about scopes. Are anyone familiar with scopes here? Okay. Scopes. Scopes are framework to categorize emissions. So you have scope one which is emissions from your activities. For instance, you have um a company delivery van. Okay. It's scope one. You control scope one. Scope two is emission from energy you use for your activities. For instance, your office uh electricity and scope three is the rest the whole supply chain you are operating from your logistic partners to your cloud providers. You are also need to have a goal because without a goal you cannot improve anything right and actually we chose to measure carbon our carbon footprint. A trap here is to uh feel overwhelmed with all the discrepancies you can have in measuring the carbon footprint because all the numbers you will get will be wrong. But that's okay. That's okay because uh measuring carbon footprint is very a relative um like I say activity. You can measure some things. Something some something else are declarative and some cannot be measured and that's okay. Feel free with that. Don't don't look for clear measures. Just have a sense of direction to get going. And a trap to avoid is it's not just our carbon because tech end users stress on materials uh resources like water usage uh stress on the power grid but you have to stop start somewhere. So start with carbon. Carbon is good. Carbon is simple. It gets you get going to create momentum. So I started with carbon with my teams and when we did the LCA for the whole company what we learned that is that tech was the third contributors to the global emissions of our company. We choose to look at scope one and scope two. Scope that we can control. And in the tech contribution, you have everything IT related devices, screen, laptops, servers, and you also have infrastructure. So the first thing to do is to get in with infrastructure. Uh let me ask a question quickly. Who how many of you are on the cloud? Okay. On prem. Okay. A mix of both. Yeah. Cool. I was on a platform as a service. The major one. I won't name nobody here. And I used to think that because I'm on a pass, everything is okay. But what I found out is uh the past is the least flexible and the least transparent uh infra cloud infrastructure because the abstraction layer they add to the cloud and you have everything is opaque the scaling is opaque uh you cannot fine-tune your own database okay you cannot look what is working under the hood. So um because everything was difficult to measure, I decided to move to the cloud. Partly to gain control and the lesson here is simple. You have to gain control and push for more transparency. So we move to the cloud and we learned a lot of things. One things you need to understand is

what's called electricity mix. Electricity mix is what is the composition of electricity in your country if you are on prem or in your cloud regions if you are using cloud. Okay. For instance, electricity in France because of nuclear is green and so it's less carbon intensive like electricity in Virginia. By the way when we were on the path we are we were located in Virginia our region was Virginia. So the first thing is to understand the electricity mix and the carbon intensity. Okay. One thing also to understand is uh when you understand the cabon intensity is that you can shape your workloads. You can if you are on the cloud of course you can move your workload to the greener regions or you can let's you can operate your bad jobs for instance during low carbon intensive hours. It's a new competent skill to build to understand your how your demands are um stressing the car intensity your power grid. Okay. One thing that helps and I think the major cloud providers did a great job with that they have all they all have sustainability guidance. Please follow them because it's well done. And if you are on prem, you can look at those sustainability guidances and apply it also to the prem. And if you have the chance to improve the hardware, always looking look for efficient hardware. ARM for instance are very efficient and less costly for the planet. And this make me come to the lesson three. Lesson three is about architecture. Um picture check this we uh in the company we uh we had uh micro service setup. So if I can share some requirements of how we are doing is very simple. We build hardware. So we track logistic. We publish audio books in 13 different countries. So multilingual audio books we uh build ehops e-commerce shops to sell audio books and to sell hardware and those are the big requirements of our architecture. At that time you build everything in house and looking through the green lines and using DDD I make three moves. The first one is simplify. We are overbuilding everything in house. We are scaling and going to North America. We have a lot of tech dep. So we need to make a choice have a focus. So what is our core domain align with the core value of the business at that time it was audio publishing and then what is our supporting domain. So the supporting domain like e-commerce okay to to delegate this part to e-commerce like chify and stuff. So I chase friction because when you're looking at technical depths technical depth signals opportunities green IT opportunities so first simplify so c domain and I made a choice to help the team succeed is to go from microervices which because of the technical that's becoming a distributed monolith back to modular monies to make things simple. That's the first thing. We migrate all of our e shops on Shopify and the third things I refocus the whole team so social technical I focus the whole team onto the core domain bringing simplicity it's correlate with green let's go to lesson four this lesson for is really funny because front end is the most overlooked when it comes to grunit. Fun fact, web pages are energy and emissions.

And this graph from HTTP archive say a lot. In the last 10 years, the weight of our web pages grow by four times. That's that's a lot. And what we don't see is that the heavier web pages swin users devices, make user charge more and recharge more and force the user change its device to compete with the complexity of the web page. So this is something to keep in mind also. And of course when we start with front end we were like uh we need to optimize page speeds because page speeds corate with what uh conversion rate. We were into page speed. We were also into accessibility. Okay. So we were all on all that and we started to look at lighthouse and to improve our core vitals and then I discover a new world of page size weight and also when you throttle the networks you can see how your website is behaving in low network regions and you see that web page weight is not just like is not just bad UX is an equity issue. Okay, so we have a ton of tools to optimize for performance. And when I talk about web pages, I'm talking about HTML, CSS, JS bundles, video, image. So we not we must compress images, we must lazy load, defer the scripts, we must trim them. And there's a lot of resources to do that free on the web. So don't forget web page optimization. And also if you can afford it, go for static rendering and edge catching to be close to the your user and to limit the data transfer. Data transfer is very expensive when it comes to emissions and use efficient protocol. You have HTTP2. Now you can have you are going to have HTB3. If you can afford it afford it, go for it. Okay. And keep measuring your real front end impact. This one is also funny because when I talk about uh to uh operation teams and devops teams, the thing that's come up even with C CFO is PHOPS correlates with green. Yes, until it's done. It doesn't. Why? Don't assume cloud is always greener. And I'll tell you why. Let's take a simple example of PHOPS. Okay, let's go for cheaper regions. are often carbon intensive regions. That's why they are cheap. Okay. When you move your workloads to a cheaper regions, you are worsening your drivers. Okay. Let's I want to to tell about speak about for instance spot who uses instance spot in the room who know about spot it's um you can ask for instance spot is um service that you can ask so if you are doing green it or phops you prefer downsizing you prefer use what you need right spot are there are here for that. So you can just add for an instance when you need it. This appear very green. But if you look under the hood instance spot used to be old virtual machines and you think that you are doing phups you think that you're doing green but under the hood you are making things worse. So the thing is you need to rethink your decision and you need to be curious to see what is working under the hood even if you are using a cloud and for your fabs to work you need to include in your PHPS impact metrics and to measure them and you need to measure the electricity you need to use your

cloud calculators it's a whole new subject I want tackle here and you need to build what we that's what we did at uni. You need to build great ways in your procurement and the lesson says as Q folks we all know that it's social technical okay it's not because you did you do LCA life cycle assessment it's not because you choose a goal like to measure your carbon footprint that it will work out of the You need to enable change. I prefer the term enabling change than managing change. Okay. And how can you do that? What we did and what some of the companies do is to build external accountability. It's like if you want to like decide to lose weight or to go to gym, you tell your friends or I decide to go to gym two times a week. You build external accountability to help you when things become hard not to give up. So this works and even at BE France we have we build external accountability as well and BE France the accountability of BE front is we are the climate bank that is one of our core missions is to accelerate uh the green transition of the French economy. So you start with building accountability when you are talking about it's very new someone is telling me this morning it's niche it's very new so you have to train the people you have to educate them they need to learn planet boundaries system thinking things new reflex to to learn some reflex performance optimization reflex to learn about sufficiency, not just efficiency. And when you create this space of learning and integration, you want to build in feedback loops. It's not a three years goal like I want to build in green and full stop. You have to put milestone, metrics, celebration, and a space for the team to reflect on the challenge. And sustainability cannot be a side quest. I made that mistake. [snorts] And what I learned is you need to build sustainability in your delivering backlog. It's another just another user stories like your whole user story in the backlog and you need to have the companies okay translate sustainability. This will be easier to do that to have sustainability in the tech uh in the tech team and don't centralize mandates. Build a distributed leadership team. I put an enable team of champions in the whole in all the teams. So tech teams, product teams, design teams, marketing teams and let the champions build the ch together and that's what I did uh when I was at you have champions. It's very difficult to build the chance to enable the change. So distributed distribute the leadership And the fellow lessons is not a story of the past. It's my current work. As a member of the green IT working group at BB France, we are looking at emerging AI patterns when it comes to sustainability. I think that it's not just I think nobody is talking about sustainability. at the governance level right now. Let's me let's take an example even the AI uh um the even if you have in Europe right the AI act even the AI act very touched touch touches the question of sustainability in the same times AI is accelerating everywhere we are talking about agitating I think AI so embedded AI everywhere. We are watching a lot of models coming

to life and dying right away with heavy training cost and for the first time it's not today it's like many months ago inference is the new energy thing that is inference surpasses training cost in terms of emission and not to mention the hardware generates Do you know that GPU chips their lifespan is about two three two three two three years if you are lucky five years. So nobody is telling that nobody is telling the users is telling us what it cost today to have AI everywhere and there's a lot of bands in different countries against the building of new data centers because the data centers is are stressing the local power grids also and we need to do something about that. Okay, because everything is free, but we know that it's not free. No pricing means no restaurant and our product and tech teams are building AI tools without ever thinking about do I really need use AI here? How can I built in telemetry, environmental telemetry and the AI usage? How can I inform the users to make the right choices? Nobody is doing that. I'm trying to do that uh right now. And one thing we are doing at uh BE France is to promote air literacy. and I work on prompt sage GPT which is a custom GPT that you can try to learn how to prompt better with uh the tool and the tool will help you prompt better and build sustainable prompts and as well and what I'm trying to do to to say here is there are many tools to achieve telemetry to achieve the control of our usage of tokens because token usage is emission and you can refer to tools like ecologists. Ecologates maybe some of you know knows about the tools. ecolog and you have lances to monitor your usage of AI. Okay. And if you have when I began this journey already when I wish I knew all that. that we are operating in a new operating system of a destabilized planet. I wish I was aware that tech is playing a central role as software engineers, tech leaders, we are playing a central role in bringing the global warming to 1. 5 degrees CC. If you have one thing to give from this talk, it is you have to learn a lot. Yes. But you have to unlearn a lot also and build a new set of question. Why? Why is the most important one? Why am I building this? about what I'm what where how and where am I running it for whom am I building this tool to build an intuition for sufficiency that's very important and I will let you with three pieces of advice the first one is resources are abundant everywhere Well, it's not a matter of tooling. It's a matter of space and reflection. Okay? And first be curious. Be curious to go out of your comfort zone because green requests to go outside out of one's comfort zone. And the second one is honesty to assess our impact and to own our negativeities

that we offload for years for years. And three, we need to be collectively have this courage to push our providers, cloud providers, SAS, AI for more transparency. So happy journey. — Does anyone have any questions for Ludy? Um, I'm curious if you could elaborate on why data transfer is such a major cost in terms of emissions. — Sorry, I don't have the question. — Could you elaborate a bit on why data transfer is such a large component of emissions? — Yes. Yes, sure. Okay. So, yeah. Uh, it's not the largest, it's the third largest. Okay. As a this is typical as a hardware company the production was the first contributor because we are having chips from uh countries and producing the toys for the kids and the second one was the distribution that's logistics that's normal and what's this is very current to have this kind of uh setup right and the tag and that come in second or Third part uh third row and um when we measure the valuation from scope one what you electricity office electricity and scope two what we learn that is more the IT side that you know that's uh make the whole uh contribution goes up and IT side is per employee so you have the laptop the mobile phone you have the servers, you have the networks and this is the crit part of uh our carbon emission. — Okay. — Yes. — Hey, thanks for the great talk. Um, one of the slides mentioned imperfect data. — Yes. — So, I wanted to hear about your experience of measurement also scaling that measurement and handling imperfect data and making people believe it. That's a good question and I tell you why. I think that you have to be comfortable not uh having first data have relative data because you need to get going and I started we started uh like three years ago with a tool called uh cabu carbo it's a French tool right and cabbo you cannot measure everything to measure everything because for instance you have okay you have your application running on the cloud to have your um the emission to calculate the emissions of your application you have to rely on the cloud calculators. Okay. For instance, uh you have uh Google cloud has uh a very uh what to call the local based uh calculator that is the your actual uh emission according to your region the region you are using but some cloud AWS for instance has market base so it's uh it's not it's unclear so you cannot have the finest number. So you have to be okay with that and the number varies uh the numbers vary between those also I told you about cowboy it's an example let's take a bigger company because I don't say it at France we have more than 5,000 okay and so you have uh I have uh 20 applications and some parts you have maybe four 400 application to measure that is very difficult So you have cloud, you have onrem. How do you do that? You need to ask your provider. So the data center provider to give you the calculations. You have to measure. How can you measure like um your web pages? So you have to go with relative numbers. But it's okay because you need you are looking for a sense of direction and you need and you need what is more important is to have a method to do that and we still thinking about the

good method at baby front right now and it's an ongoing work and that's okay it's a green it's not uh simple and you we need to learn and in three years I can see shifts In the tools there are many tools and I will complement my talk with a document about the tools that were available. So things are getting better but don't look for the numbers. It's not the real numbers. Look for the relative the sense of your numbers to get going. — Uh yeah. So, um, how much buyin does it like how much convincing does it take from the business side? Uh, because I'm guessing you can't just be like this is for saving the world. They might want to be more convinced on the money side of things. Like how much of that is just Yeah, — I think you get it right. It's not just about saving the world. I think it's for resiliency. And don't forget the initial u message. I think it's very important to acknowledge that we are working in a destabilized world and that natural disasters are causing supply chain disruptions are threatening lives. Okay. And when it come to your companies is to let them know this is it's not easy. You have to build a business case around that and uh looking for external accountabilities even if it's for not green rashing but it's for communication it helps and look looking for to find how the your heavier web pages are slowing the business are slowing your scale and looking for business disruption uh signals and tech depth signals efficiency. You can do green it through tech as well. So if you have friction somewhere just look at those the friction to base uh green IT cases and have champions with you like talk to your your early adopters. Yeah, there's someone in the back. — Okay. — Yeah. Sorry. Uh last question. Um how do you help dev team to take time for queen it subject? Uh do you have like dedicated Jira tickets in the spring backlog or weekly meeting to explore metrics or stuff like that? Yeah, for me it's uh um so it's different. Uh at I tried okay do green IT so come up with the champions help the team come up with uh user stories. So green IT user stories and I try to let them uh do the stories when they have time. And this is a huge mistakes. So I decided that they will have 20% of their time for the grunge. So it's it's an option. But if you don't have this option, just make sure that every team embark one at least one green it user story in their sprints if they have sprints. So you need is I think that's the trap here is to me make it a side quest. You want to build a normal a new normal behavior. A new normal behavior is please uh be alert be aware of greenity optimization and build that into your backlog. Don't wait for team gathering and let's celebrate also let's celebrate reaching a green goal also and my what I hope is um in few years we will forget about the green side of the equation is just normal work but we we're not there we are not there yet. So building on your question right which is around enticing leadership on this journey. So really what I want to know is what advice do you have around bringing developers on this journey and enticing them and also is a way to wrap in um regulation on top of that. [snorts] — Uh okay so that's two questions.

I think that um I would say something maybe uh the chances I have it's a chance is to have uh young developers want to change the world. I don't say that older developers don't but young developers are ready for that are ready for the change. And something that's very interesting at UN it's not just so I talk I talked about granite but is a part of a whole sustainability movement. So grate but when you think of circular economy it's part of the sustainability movement and the younger generation is ready for that they need methods approaches and helps make them time to do that for people not ready for that culturally not ready it's a question of education you have a sustainability leader at Looney and her goal was to train everyone on sustainability on Paris agreement of how to measure and to make everybody easier with the fact that we don't know we think we estimate but need to improve to get improving and we don't know the real answer okay and to get cool with that so I think that education is a part that and maybe meeting with other teams that are doing sustainability well BBL so things like that and regulation it's funny because it's interesting because I can build up um on leadership thing u regulations are here I think the most constraining regulation I'm in fintech right now you have to disclose your climate risks it's a lot of reporting but it's you can use it as a lever to start green journey the climate risk you have ESG um you have uh you have those reporting and many of them I don't remember the whole one so use the regulations so this kind of regulation like that is constraining us to report our carbon emission is concerning sustainability, the global sustainability, equal equality and use that to foster a sense of a movement uh a sense of transformation where regulation is not helping right now for me is uh in terms of AI and sustainability and we need to build uh that kind of governance. Yes. —