 All right, guys, before we start, quick reminder, you can post your question Q&A on the app and then we'll read it from here or you can ask me for the mic that I'll bring the mic to you. Yep, that's it. You guys can start. Okay. Well, welcome everyone. Thank you for joining us today, and this is a little talk that we've put together not so much in the idea of back slapping. It's more like the idea of saying we're setting a base camp. We're saying there's a great big mountain in front of us, there's something we want to actually achieve together, there's a summit that we want to climb towards. So it's not about saying this is all done work, this is about actually the beginning of the work. Let's put it that way. So I'd like to give a little bit of a background about myself and then Jerome and Antonella are going to do the same. I've been working in Drupal in the kind of open source scene since around 2010. So that's like quite a long time now. My first experience of Drupal was actually going down to an open source conference. I live in Germany and I went to an open source conference in Bonn. I sat down next to a spiky head guy who turned to me and said, hey, who are you? And I said, Andrew. And he said, yeah, I'm Dries. So it was a good way of getting to know Drupal. First person I ever met was Dries himself. And yeah, I used to work for Acquia for a few years. Actually just before Nestle started working with Acquia, they managed to avoid me at that point, but they couldn't avoid me later. And I'd like to now pass on to Jerome. Yeah, sorry. I look after Platformer's H's relationship with Nestle globally. That's kind of my role. So, Jerome. Hello. Yes. I'm Jerome. I also work for Platformer's H and I do strategy works, which means a lot of different things from environmental impact to a lot of other stuff. I'm actually a life science engineer at Art. And my most profound relation with Drupal was that I don't like PHP and I've been doing GoLang most of the time, but what I liked about it was actually trying to reduce it most of the time, like reduce the number of servers that I use for that. And I've been doing that in many occasions, like that's been my hobby for a long time. Hi. I'm Antonela Severo. I'm the Product Manager of WebCMS, which is part of the Marta Group at Nestle Global IT Hub in Barcelona. And we take care of a suite of solutions that power Nestle's brand websites. So it includes Drupal as a CMS, our past platforms, governance models, processes, and lately DevOps. And so before that, I was actually on the agency side. I worked at Third & Grove, which mainly is a Drupal agency. And I've also been around a bunch of years. I mean, I was here before websites. I built my first websites in HTML. And also, just to show you how long I've been around, I was an early adopter of LinkedIn. So I actually was able to get LinkedIn slash Antonela, believe it or not. So first name. Yeah. So let me tell you a little bit about Nestle. Just to give you a bit of context, in 2015, there were about 2,100 brand websites. And of course, it was completely decentralized. Hundreds of CMSs, many different hosting platforms, many individual local agencies not connected. And as you can imagine, this is inefficient. The total cost of ownership is very high. And we just couldn't manage the security risk. So we undertook a journey as all solutions in IT. And we went through a digital transformation. So what were our goals? So the first thing we wanted to improve digital experiences for our users, we wanted to provide the businesses and markets freedom in a box. So instead of building all our products, we were going to adopt, adapt, assemble as much as we could. And we also wanted to make sure that the businesses and markets could use their marketing budget more for the custom features that are relevant for them. And we would take care as much of the technology as possible, like 80% or more. So how did we work towards achieving this? So one of the things we did is follow the IT strategy, which is to simplify, standardize, and share. And we relied on best in class solutions. And we consolidated the tools and processes. We masterized as much as possible. And we even decommissioned sites that didn't have as much value. And the results today is we have 1,300 brand websites, 1,200 are in Drupal. And to understand the scale, that's like 300 million page views per month globally. So just to give you an idea of how we scaled up what our model is, we are undergoing the master approach, which maybe is like 60% of our sites. And there's about 30 masters. So it's a unified code base. And then we have templates, and we can have many templates against each master. And then we have child sites, each master could have like 60, 70 child sites. So yeah, and we built that on Drupal Core, but that wasn't enough. We also built a Drupal distribution, which is called Lightness, to even kind of standardize and share more. And then united with our past platforms. And then on the next level up is web application, which the business owns with their technical agencies. And so I will pass it on to you. Great, okay, yeah, so thanks, Antonella. So that's a bit more about the way that Nestle is doing things. Platform as H, hopefully there are people in this room that have used Platform as H who are familiar with Platform as H, can anyone? Okay, so maybe it's just short of half of people actually using it or maybe familiar. Hopefully you know that Platform as H has been around the Drupal space for a while, we actually come out of that sort of background. Drupal commerce was something that was built by the same people as built Platform as H. Our CTO was the guy who had the most commits on that beautiful old beast Drupal 7. So we're kind of Drupal through and through from where we began. And this little video or this little image gives you an idea of the concept behind Platform as H. You as a developer simply insert your code and Platform as H does the rest. So the idea is that you can then actually have the whole way in which you develop your project is essentially completely code driven from infrastructure as code. So you actually define the servers, the services, all of that kind of stuff is then done simply through Git. You manage all of your data, you've got your scaling, your operations, compliance, access control, all of that is run via Platform as H. So we like to think that that means that we also fit well with companies like Nestle doing large scale Drupal, doing large scale operations where you need to manage lots of stuff at once. And that's kind of like the sweet spot for Platform as H, that kind of fleet management where you've got hundreds of projects, where you need to manage those efficiently, which kind of fits nicely with what the Drupal people at Nestle are also doing. And then the next thing that we're going to talk about is how Drupal actually got us together. Right, so I think in 2017 at DrupalCon Vienna, not far from here, some members of my team went to DrupalCon and of course met the Platform as H team. And subsequently when we started to do an RFP to expand our panel of past suppliers, we invited them to attend and they were chosen as one of the three. And yeah, that's how Drupal, the Drupal community really helps to share and make connections. Exactly, so we will just talk a bit more about Drupal as the history of that side of things. Yes, so yeah, as we talked about earlier, but Drupal is actually important. There are large deployments of Drupal and for us we lost a lot of the website, but it's set to represent like 2% of the internet, it's a huge thing. And so for us, we found millions of users, millions of projects actually. I don't know if you will say anything more about this. Well, yeah, great, okay, next slide then, yeah. So fast forward to last year when our product group manager reached out to our vendors to find out what they were doing in digital sustainability. And why did we do this? Because Nestle was on a path, we're working hard to reduce the carbon footprint across the whole company according to the Paris agreements to achieve net zero emissions by 2050 at a minimum. And so there's a lot of actions underway throughout the whole company. And so in our department, or our area of IT, we started to look at what we could do with brand websites. So we started on this journey. So we learned that platform as H was already actively investigating the impacts and possible ways to reduce this. And also there were other efforts going on at Nestle. So there was an innovation challenge on sustainability. And another team, one with their project, which is to look at brand websites, they're associated with us. So of course they contacted us to work with them when they won. And we also did a roadmap study for our Drupal distribution. And one of the features recommended was sustainability. I don't think we knew what, but we knew that it was there, so. So we started to have discussions with all of these different groups. And at some point we're like, okay, let's get everybody together. And see, just started discovering and share where we are and what we're doing. And as a result, we created this kind of green action committee. And so that's how we started on this road together. Yes, one thing we started from this, the thing is that we performed our first carbon audit last year on the 2020 data. That is something that is important because that's the start of trying to measure things. A carbon audit is just basically paying someone to look into your data and give you numbers about what you are actually emitting. And for us, it's about what the company does, what the infrastructure we run for our customers does, and what the transit, the network aspect, for instance, impacts all. So yes, measure all the things. Why is there an interrogation mark there? Is that because we are a past provider, we could just ask Google, Amazon, and Azure, and OVH, and all the one we were working with. What is our current carbon emission? Just use their word and they just believe what they say. But you may know that they all claim to be carbon neutral. And what our carbon auditor says usually is that, okay, carbon neutrality is just actually buying carbon credit or actually buying renewable energy and sending it back to the grid. So that it compensates where it offsets the emissions. But the infrastructure are always growing. So they are always emitting. Taking into consideration the fact that they are buying or selling renewable energy is called market-based accounting. And the most important regulation that are being passed are insisting on using the location-based approach. So what is actually the data centers, the green data centers are actually plugged in the same plug, electricity plug than every other system. In France, for instance, there is no greener data center in terms of energy, one to another. Some of them are a bit more efficient, but that has to be measured. So this is an important thing. So if we are going to that, measuring is something we recommend for everyone and that's something we've done and that we've been trying to share. And the way we measure things is about taking actually the resources we use and try to see what the electricity consumption can be. So some provider can report this. And then getting the carbon emissions from that is just applying the carbon intensity that is local to the country we use. So carbon intensity is the amount of carbon produced by a kilowatt or megawatt hour of electricity. So in France it's pretty low because we have nuclear energy. It's not really carbon emitting, but it's nuclear. In Sweden it might be hydro or in some other place might be wind farms, so different aspects. So it can vary from one to 20, say, 15 in France and maybe a thousand in Australia. So it varies a lot in the world. And we'll get back to its measuring is actually something important due to just a different appreciation. If you compare, for instance, Google and Amazon there are very different numbers for the same infrastructure. So that's important. And going from there, I think carbon is interesting. But for Nestle, for instance, getting to know what project might be more impacting or what system could be requires us to go a bit higher and to get into the metrics of how much CPU is used by this project, how much RAM, how much storage network, and what is the most contributing, can we improve on things? And how can we model this information to eventually break down our general carbon audit per customer? So this is something that we'll be doing this year for all of our customers on the data from 2021. And this is an ongoing process because the more we do it, the more we learn, the more we identify where the fine details or where the fine prints from our providers are written to. So let's get to the rest. So speaking about how we measure things with the initiative project that we're doing, we decided to start small and do actions that we can achieve. So we have three different goals. The first one is to set the baseline. We want to measure where we are now. And we want to find out what is the carbon footprint of each site. So of course we were like, OK, let's just get a whatever calculator that's on the web and throw a badge up on the site. Of course, when we dug deeper, we know, yeah, but what are those calculators really measuring? We need to understand better and find the right calculator. Then they helped us to understand that. So we need to figure out what are the meaningful indicators, because we don't want to greenwash it. We want to really understand what's happening and display it. So the second one is to bring awareness. So to visualize the data, we're going to create dashboards. And to kind of start having our markets and our vendors understanding that and helping us in that area. And the third is to encourage, change, and track the process, the progress. So we're going to provide best practices. And we're going to do the tweaks along this road and keep finding out new ways to improve the system. And then so one of the things that we also discovered is we also have to optimize our application level. So to make the application resources more efficient, as Jerome had just mentioned, we don't want to throw metal at it. We need to understand what's going on. We don't want to just upsize a container when it's getting higher. We need to dig in a little bit and kind of make sure that we're just we're finalized. Analyze what's actually happening inside of that particular application, how that's actually putting resource strain on the particular resources and so on. So we're going to provide best practices for our websites. But we also have to look at lightness, which is the distribution, because that's the basis of many sites. And we understand how to make it more performant. And actually, we're going to be doing some research projects with them to kind of understand that. And the results are still to come on that. Yeah. Great. Thank you. Yeah, so the application performance, as we said, we are measuring things. The general model of the carbon emission impact is just trying to get the most efficient applications, the resource efficient. It shouldn't require a lot to deploy it as densely as possible, and that's what we're going to discuss, and try to densely as in using fewer resources, but use those fewer resources more. And maybe once we have this very efficient system, try to locate it where it's actually emitting the less carbon on the best data center in the most move to the side, Jerome. Yes. Yeah, there's only one picture, but it's an interesting one. That's what we do at Platform. We pack application very tightly, very densely, so that we add the carbon in at a third party independent, try to look into it, and try to get information from the market to tell us, OK, how dense are we when we deploy all this application, those Drupal applications that usually are deployed on dedicated resources? And as you may know, servers, they usually idle, because we over-project vision a lot as a community. I mean, that was the machine is cheaper than the workforce. So by hour, it's easier to throw metal at problems. And if you have problems, you just push on a server, or you upsize it, and then you forget it, and you don't pay engineers or persons to work on it and optimize them, because sometimes it's easier. So yeah, most of the time, a lot of servers are too big, and they are just wasting energy. They're wasting energy, because yes, they had to be produced at some point, and that supply chain, everything. It's not without impact. And moreover, a server actually consumes a lot of energy, just doing nothing. The idle, the consumption per operation, it's not neglectable. And the more we use a processor, the more you use the CPU, which is the most energy consuming part of the server that we use. We don't do you CPU or things. Storage is actually kind of cheap. So the more we pack application together, the most efficiently. So using fewer resources, but more like loading them at 80% say. So we are, in general, our auditor says like 12 times better than the most naive deployment on AWS, for instance. The others are not really sharing, unfortunately, other providers of container service, for instance. It's very unclear how the server less function from GCP or AWS, we don't have real numbers about their density, so their overcome it level, for instance, it's difficult. Our global mantra is about right sizing when we do dedicated deployment. So make it as small as possible and allow it to grow when needed. And in our container approach, then try to pack it. And this allows for great gains in general. So density, and then yes, once we have this very efficient system, let's say we have used observability, optimize everything, the Drupal that we have is the most cacheable, it's great, it's densely deployed with all of the desapplication from different persons. Now we need to deploy it somewhere. Maybe we could advance through there. So we will try to deploy it where it's greener. What does it mean, greener? As we said, carbon intensity is a thing that varies a lot and it's a function of the energy grid. So the energy grid is in France, for instance, it's a mix of mostly nuclear energy and then wind, I think is the second energy we use, but it's very different, for instance, in Canada. So you see here where it's green, it's actually greener, it's the lowest carbon intensity possible. So the lowest, actually, I think it's in Quebec, it's maybe below 50, something like that, 50 grams per megawatt hour or something. So this is very different, it can vary a lot and depending on the location of your customer users, you may want to be able to locate your application where it means it could have the lesser impact. For most of our customers, for instance, for the Drupal project we asked, we often rely on a CDN to handle almost of the static assets kind of network impact. So locating a website in Germany or France won't have a big impact on the performance of the website considering everything, but then why should you deploy in Germany? Maybe for data locality, maybe for compliance, maybe for security reason, maybe because we have a good partnership with the local oyster, but it's difficult to make a case from an environmental impact standpoint if there is a 5 to 10 X ratio in carbon intensity. So we are considering locating our project, our region, public region, where the intensity is actually lower. There are no solution for every place on earth. For instance, in Southeast Asia, it's difficult. There is nothing quite good. Currently, Australia has a very ambitious program, but you can see that it's probably one of the worst place on earth to place a data center. And in the US, it's not much better, but we are Canada for Australia. It's always the answer in North America. Okay, great. Thanks, Jerome. I'm gonna speed it up a little bit because we don't have too much time, but I think to take a step back, what we talked about basically is three different areas where we can improve things. And those are the areas we're working on, not just with Nestle, but in general. But specifically, we're trying to look at app efficiency. How do we make Drupal, for example, work better in the cloud? Secondly, the density and right sizing. How can we make sure that we're packing as many projects into the most efficient regions? And then thirdly, those greener regions. So what we've got there is it's like a, you know, it's a Venn diagram. Where we have that overlap is where we can actually get to much higher levels of improvement in terms of what a particular project can actually do in terms of its carbon footprint. So as you can see, it's not quite exponential maybe, but the more that overlap happens, we're gonna move from a good improvement to a better improvement to a best improvement where all three of those come together. And I think that's the exciting bit is where, you know, we can say, well, you know, we've got things that our customers can do together with us as partners. We've got stuff that we as PlatformSH can do in terms of the density, in terms of the right sizing. And then we can also look at finding the best places to move these projects. And, you know, in the most humble way possible, not us standing there waving a finger, but trying to educate ourselves. That's really about what we want to do here, you know. As I said, it's about having a base camp, not about having a back slap, you know. So what we're trying to do is find a way to learn together, to advance together, and you know, to grow in terms of what we're doing here. And I think that sort of like comes to be, like I think the final message, that's what we're really here for. And I'd like to hand over to Antonina for a conclusion there on that point. Right, I mean, maybe I'm not the most technical, you know, person on this, but I'm learning a lot through this journey. And it's actually getting very exciting to see the impacts that we could have, especially at Nestle where we have 1200 websites. You know, on our team, you know, we're beginning to understand, you can't just look up something in Google and find a proven methodology because it doesn't really exist yet. And you know, we can only challenge our preconceptions and fears and ask questions to probe deeper. So you know, again, the goal of this session was not to give you results, it's just to invite you early on, you know, give you a glimpse early on in our journey. And you know, I'm kind of curious how many have been thinking about sustainability in their projects with a show of hands. Okay, great, great. So I'm glad you attended this session and you know, joined us so we could share, you know, what these small steps are that we're taking, the conversations we're having. And perhaps we could jumpstart the sustainability session, you know, here at DrupalCon. I mean, what can we do in that sense? How can we make Drupal as a whole more efficient? We know that the Drupal Association website is committed to it, but you know, so that's a small step which is great, but what about all of Drupal sites? You know, how can our community contribute? What standards will we use going forward? Should we do a working group? You know, what are the next steps? So feel free to contact any of us, you know, with any questions. Absolutely. Yeah, whatever. Yeah. Yeah. And any questions? I think we got some maybe, there was a Q&A. Yeah. We have one from the app, let me read it to you guys. Lukash, I think it's Lukash. Which hosting platforms does Nestle use? How does Nestle choose the right one for the project? We have three hosting providers, Acchia, Pantheon and PlatformSH. So it really depends on the type of project that we're talking about, but we give the options to our markets and our businesses and brands, and they are free to choose in the end which one they want to be working with. Right now we're focusing for Masters with Acchia and Pantheon and single sites with Pantheon and PlatformSH, but this is going to be changing soon and we want to work with our three providers for all kind of sites. I hope that. Anyone else got questions? Yes, I'm running to you. So just I want to understand, if you don't share results with us, it's still in progress or to have already the impact? And did you already choose at least the tools that you will use for the measurement or not yet? So regarding results, I'm into the microphone. So regarding the communication of the results, I will let Nestle answer about the communication, communicating this number, but we are sharing every month about between us and these numbers and regarding the tools now. We have many tools already in place because we are collecting every day, like we have data lakes, or we are collecting all the metrics for the usage of the system, the application, the reservoirs. So we are collecting them. We've been collecting it for a long time. What we were, what we are not was a model to compute and to have actually electricity consumption from the resource usage. So now that we have this in place in our systems and we are checking these numbers with a carbon auditor third party that comes and check if our model works fine. So we started this work last year with the first carbon audit. We did some POCs for first version of that. Now there were problems we've been learning. We've been using different models for network computation, all this is progressing. So now we are about to finish our second carbon audit for 2021 and to make a breakdown of this number for platform usage as a role for all our customer per project. So that will be the deliverable for the Q4. I can't really tell when it's going to be. So in terms of tools, it's just things that we already have. Observability tools, honestly. Yeah. Thanks for the session. It was very, very interesting. I just have a question on, so you spoke a lot about measuring the emissions from the hosting, from the network operations, but are you already or do you have plans of considering the resource operational emissions? Like say people working on these applications, the emissions that gets generated, let's say in the logistics part of this chain, or any related logistics of these data centers, transport, et cetera. Actually, to be frank, no, because there are many things that we like to come to.