Show notes are at https://stevelitchfield.com/sshow/chat.html
…
continue reading
A tartalmat a Gaël DUEZ biztosítja. Az összes podcast-tartalmat, beleértve az epizódokat, grafikákat és podcast-leírásokat, közvetlenül a Gaël DUEZ vagy a podcast platform partnere tölti fel és biztosítja. Ha úgy gondolja, hogy valaki az Ön engedélye nélkül használja fel a szerzői joggal védett művét, kövesse az itt leírt folyamatot https://hu.player.fm/legal.
Player FM - Podcast alkalmazás
Lépjen offline állapotba az Player FM alkalmazással!
Lépjen offline állapotba az Player FM alkalmazással!
#33 - Is open source and sustainability a perfect match? with Oliver Cronk and Katie Davis
MP3•Epizód kép
Manage episode 403307176 series 3346125
A tartalmat a Gaël DUEZ biztosítja. Az összes podcast-tartalmat, beleértve az epizódokat, grafikákat és podcast-leírásokat, közvetlenül a Gaël DUEZ vagy a podcast platform partnere tölti fel és biztosítja. Ha úgy gondolja, hogy valaki az Ön engedélye nélkül használja fel a szerzői joggal védett művét, kövesse az itt leírt folyamatot https://hu.player.fm/legal.
🔍How open source software slashes energy consumption, extends hardware lifespan, and drives sustainability?
🎙️In this episode, Gael Duez unravels the intricate relationship between open source software and IT sustainability together with Oliver Cronk, the host of Architect Tomorrow podcast and Sustainable Technology practice lead at Scott Logic (UK), alongside his software engineer colleague Katie Davis.
💡Discover how open source software (OSS) and hardware (OSH) are not just benefiting companies, but also driving global sustainability efforts. With insights from a recent HBR study revealing OSS's monumental impact on reducing software development costs, the stage is set for a game-changing conversation. ⚙️
💻Katie breaks down the key to OSS success: transparency. Learn how freely available code empowers developers to make efficient, sustainable choices, while Oliver highlights the importance of common hardware standards and circularity to combat waste and inefficiency.
📊Dive into the battle for open-source data transparency with Katie, and explore Oliver's vision for the tech community's future direction. From reducing emissions with initiatives like the Tech Carbon Standard to innovative projects like the Open Compute Project, the episode is packed with actionable insights and inspiring solutions.
❤️ Subscribe, follow, like, ... stay connected the way you want to never miss our episode, twice a month, on 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 next Conference is in Singapore on April 18th (use the voucher GREENIOVIP to get a free ticket)
📣 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:
📧 You can also send us an email at greenio@duez.com to share your feedback and suggest future guests or topics.
Oliver and Katie's sources and other references mentioned in this episode:
- Scott Logic’s Technology Carbon Standard
- Harvard Business School Strategy Unit Working Paper No. 24-038 on Open Source Software
- Open Compute UK
- Tristan Nitot's talks about eroom
- GSF Carbon Hack
- Boavista Cloud Scanner
- Cloud Carbon Footprint
- EU Energy Efficiency Directive
- Climate Q&A
- Built and planned electricity Interconnect across Europe
- Viking Link: Interconnector from Denmark to Great Britain
- Reporting requirements on the energy performance and sustainability of data centres for the Energy Efficiency Directive
Transcript
Gael 00:00
Hello everyone. Welcome to Green IO, the podcast for responsible technologists building a greener digital world, one byte at a time. Every two Tuesdays, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the tech sector and beyond to boost digital sustainability. So today we're going to talk about a relationship, a deep, pivotal, and yet troubled relationship. Open source and IT sustainability every person I talk in our sustainability field takes it as a no-brainer. Open source is good for sustainability. But which open source are we talking about? And is this link? Is this correlation so strong? For instance, is open source software always more frugal or sustainable in the way it is designed?
We don't know. Let's investigate it. But there is at least one certain thing. Open source software is big. A recent post made by Sasha Lucioni, one of my absolute favorite thought leaders in AI, caught my eye. A recent HBS study found that companies would have to spend 3.5 times more on their software development if open source were not around. And we're talking about an $8.8 billion estimated market. So open source is big. Let's see how it connects with sustainability.
Today I am delighted to have two guests to talk about sustainability. Who has done open sources from the trenches? Oliver Cronk is a fellow YouTuber. Now I can say because, yes, Green IO launched its YouTube channel a week ago to offer a wider choice for accessing its content than traditional podcast platforms like Apple or Spotify. Anyway, Oliver is the Architect Tomorrow host, a channel for enterprise and platform architects with a soft spot for sustainability, among other topics. He's also a tech strategist at Scott Logic, a software and data consultancy based in the UK, and has a significant track record in both IT and energy. As an example, he built carbon calculators, air quality databases, and industrial emissions reporting systems, which made him experience firsthand the issue of accessing open, transparent, and reliable data. Katie is a software engineer with a Math’s Degree, so not easy to mess with her when it comes to data and statistics, and she is a driving force in this sustainable technology practice at Scott Logic, especially involved in the recent release of the technology carbon standard at open UK several weeks ago. A standard under the Creative Commons license, of course. By the way, they will present it at Green IO London on September 19 this year, but that's a different story. Welcome, Oliver. Thanks a lot for joining Green IO today. So, are you on your bike or more comfortably sitting at your desk?
Oliver 03:02
No, today I'm stationary. I'm at my desk.
Katie 03:04
Thanks for having us.
Gael 03:06
You're welcome. That's nice to have you on the show. And welcome, Katie. Thanks a lot for joining Green IO today as well. That's vegetable time, not food time. I guess because you love to garden and to grow your own food. That's a very nice angle for sustainability. Katie, maybe the first question will be for you, with your academic track record. When we talk about open source, it can actually cover a lot of different angles. Could you give us your definition or your way of approaching what is open source?
Katie 03:40
Of course. So when I think about open source, I always think about open source software. So this is software whose source code is freely available to modify, to distribute, deploy, and even extend it into their own projects. So I think one of the key parts of open source is that it's transparent. You can view the methodology behind the code. There are also types of open source, which I think maybe Oliver could expand on.
Oliver 04:08
Yeah, no, happy to. So I think you've touched on open data, which is an important one. But then open hardware is really interesting as well when it comes to sustainability. So at the open UK conference, you mentioned, Gail, there was a great presentation about open compute, the open hardware sort of initiatives to ensure that there are sort of standards for hardware. And there are companies now taking the used hardware out of the big hyperscaler data centers and reusing it because it still has a life. Right? The big hyperscalers no longer want it, but you can easily get additional life out of it by hosting it in another data center and getting other people to run workloads on it. So there's a whole host of different strands of the open source kind of ecosystem, really. Software, hardware, data, significant ones.
Gael 04:48
I think that would be a good way to define these pillars of data, software, and hardware. So, Oliver, you started with hardware, not maybe the most obvious when we are in the IT world. What are the progress today that you've seen in open source hardware?
Oliver 05:05
Yeah, I'd love to throw to Katie to talk about the sort of data that's available, hopefully in the future, in more of an open form about hardware. But in terms of hardware itself, I think there are some really interesting developments. Like I talked about, the open compute project, they have a whole sustainability division that's looking at the circularity of the hardware. I think having open standards when it comes to hardware is important because otherwise, it's the compatibility challenge. Right. If everyone is inventing their hardware to kind of be unique and proprietary, it just means we have a whole lot of potentially wasted components that won't work with other systems and we're producing lots of different types of components that only work with one particular platform. So I'm a really big proponent of the open compute project also because they're looking at things like can we increase the ambient temperature of data centers, for example, to reduce their cooling requirements and things like that. So they're doing some really clever stuff in that area. But I think also open hardware is interesting when you look at the sort of hardware hacker sort of culture that's created as well. Right? What can you run on an embedded device like a Raspberry Pi? How can we extract every last efficiency out of a couple of watts that a Raspberry Pi runs on? So I think there's some really interesting sort of things that come out, the constraints that you impose when you look at sort of consumer grade hardware and the open source movements around that.
Gael 06:20
Yeah. And just to not close, but to comment on the hardware topic, it really resonates with the new battle of my good friend Tristanito, who's one of the founders actually, of Modzia Europe. So he's been in the game for quite a while and he's got his amazing talk now that he's given several conferences already about Moore's law is dead now it's eroom. So that's kind of reversing the Moore name. The idea is every year you don't double the size of your computing power, every year you divide by two the size of your code. So every year you can run your code on lighter and lighter machines, which means older and older machines, which means saving potentially millions, avoiding building new computers that actually don't need. But anyway, Katie, you wanted to elaborate a bit more on the data side. I think this is one of your main battles, isn't it?
Katie 07:18
Yeah. So especially with the sort of manufacturers of hardware, the data they provide themselves, often the figures they're given, these, it's not clear on how they actually calculate them. So for example, if you wanted to find out the typical energy consumption, they might provide this figure, but it's hard to find out where they're getting this from. How long are they running the device to arrive at this figure? Additionally, there's another sort of documentation that these manufacturers provide, like product carbon footprint, which gives the emissions from each use stage of the lifecycle. So manufacture transportation usage, end of life. But again, it's just not clear how they get these figures. So the benefit of open source is having that methodology transparent to everyone to see how these figures are arrived at.
Gael 08:06
Yeah, the big issue if you've got quite diverse hardware estate is that we might add apple and orange because the way one manufacturer would calculate things is not the same as the others. And eventually, these numbers don't make sense at all. What do you see as a potential path to overcome this very big difficulty in data sustainability?
Katie 08:34
I think open source is key, just having a sort of level playing field for the methodology that we use to calculate these things. So I mean, I think we'll probably discuss it later, but like tools like Cloud Carbon Footprint, sort of having that methodology that you can use across cloud providers, so it's more comparable, easier to track your improvements across the estate. And I mean, that's just one part of the standard.
Oliver 08:59
Katie touched on a really important point there, which is at the moment because the manufacturers can essentially almost makeup, I'm exaggerating a little bit to make a point, but because they can almost make up their methodology, their numbers, it almost becomes a marketing exercise to make it look like their product consumes less energy. And it's a bit of a game, right? So if you come up with a really creative way of measuring the power consumption of your device, I mean, we've seen servers, for example, with different sort of load ratings and energy consumption, different load ratings, and each manufacturer seems to sort of measure at different loads. So some are at 20%, some are 80, some are 100. It's like, okay, well, you can make yourself look cleaner by just presenting the characteristic that your model of hardware tends to perform best at. And of course, then you don't really know whether you're making an improvement or you've just bought better marketing when you change suppliers. So I think that's why what Kate is saying is really important. This apples-for-apples comparison is super important because we already have the same with the cloud. Like you run AWS's cloud carbon tooling versus Azure's versus Google's, they all have different methodologies, and so they're not comparable.
Gael 10:04
And that's a point with carbon footprint. And I'm very happy you raised both of you, actually, you raised the point. Is that true? It's a very powerful tool. Cameron was on the show last year. I'm a big advocate of CCF, among other open source tools, but at the end of the day, they have to recalculate things that should be provided by hyperscalers. And I think here we've got two very serious issues. The first one is the methodology. Tools like open source tools like Boa Vista cloud scanner or CCF, obviously the methodology is transparent, so at least we know how they manipulate the data. But then there is the issue of the data itself. I mean, do we have the granularity to calculate things in a wise enough way? And I'm not always sure, because for cloud. But Katie, I will let you elaborate on it. For cloud, it's always a question of transforming financial data almost all the time, financial data, into sustainability data. And some data would be more, I would say, accurate, like starting with CPU usage or GPU usage or whatever. But what's your thought on it, Katie?
Katie 11:17
So with regard to cloud, one of the key issues is that each cloud provider has their own tooling, but the methodology is not the same. So I think one of the main challenges is to do with sort of greenhouse gas scopes. So I think AWS, especially, doesn't include scope three. I think that one of the benefits of the cloud carbon footprint tool is that it provides a way to at least try and estimate that proportion allocated to scope three and sort of compare on the same level with those from Google as you are. I think Google does provide data on scope three emissions. So I think it's sort of just picking the right methodology for your estate. So, for example, if you just had Google Cloud, maybe you just go with their tool, since it's probably more up-to-date and includes scope three. Whereas, if you had multi-cloud or AWS, you might want to look at a tool like Cloud Carbon Footprint.
Oliver 12:14
But I think the point, Gael, around using financial numbers to come up with sustainability metrics is an interesting one. Clearly, it'd be great if we could trust the native tool, like Katie said, to use the Google tool, the AWS tool, because they have access to the underlying data that allows them to more accurately calculate the emissions. The trouble is, there's a trust thing there, isn't there? I think it's how transparent are they being and how much we trust them to kind of calculate this fairly, rather than just putting forward a polished view of their emissions, a managed kind of view of their emissions rather than reality. And so this is where I think the open approaches really would help because if they were open and transparent about what went into providing the service that you pay for, you would be able to make a more informed decision. And I think the calculations would be far more robust than basing them just on billing data. Ideally, you find situations where you save money, and you save carbon, but those two aren't necessarily going to go hand in hand in all cases.
Gael 13:12
Yeah, absolutely. I had an issue with a client very recently when we made the calculation that the bill would go up by 20%. And then you have also the question of how do you calculate a clean energy region. Because I don't know if you've seen this beautiful post from Mark Butcher migrating from Scotland to Ireland, where almost a time six difference in the carbon intensity of the electricity grid. If you do very basic math, you will still double your carbon footprint. Still, I would like to go back to it. Okay, so we claim that open source and open data because we are 100% on this open data subpart of open source here should be there, but the question is why not? One of the pushbacks I've heard from Isposkella representatives is like the business secrets and that they don't want to share because it's sensitive data. But what's your thought on it? And do you believe there are other obstacles, I would say on the road toward mods transparency, at least from the data perspective?
Katie 14:18
I think that sort of garden data for a commercial reason, especially with regards to sustainability, is just not really the way forward. I think to sort of make sustainable software key and forefront, we really need to be transparent, so people can build on methodology. It's not really any good sort of reinventing the wheel when there are already so many good established, like data point methodologies out there. I think as well, open source projects, have such a large user base, especially compared to some enterprise software, especially internal sort of company software, not always, but sometimes. So I think any efficiency gains that we make to open source software can have a real downstream impact.
Oliver 15:05
Yeah, Casey's kind of moved on to one of the other key kind of conversation points we have around open source. Right. But before we go there, my take on what you're saying, I think the commercial pushback is an interesting one, and I'm not convinced it's as simple as if we just reveal some more data, and all of a sudden the cat is out of the bag and everyone will know how we're running our data center. I think the reality is running a modern data center, certainly for a hyperscaler, is super, super complex. Right? And they're even using AI to optimize. Google and Meta, for example, have used AI to optimize their cooling and stuff like that. So I think to think that if you just release some data about your energy consumption and the high-level waste footprint and water footprint and all that sort of stuff, that that's going to be commercially dismantitious if you release that information, maybe to a point, because it will reveal perhaps to your competitors how efficiently you're running things. But yeah, I think it's just an excuse that they're hiding behind, quite frankly. And I think what will happen if they continue to drag their feet is the EU will regulate. And I think the EU is already regulating. Right. If you look at the data center regulation that's coming out, if you're running a data center over a certain power consumption now or in the near future, you will have to start reporting on more data. So I think that the reality is if companies continue to drag their hills on being transparent, regulation will follow. And it's probably better to kind of get ahead of that.
Gael 16:27
Yeah, you're mentioning the energy efficiency directive, and actually I will put the link in the show notes as everything that we've discussed so far. But is it the only way forward? Is it okay to regulate? I'm a big advocate of better norms and more transparency, but do you believe this is the only trigger that will force big actors to become more transparent? Is there any other way?
Oliver 16:51
I think Google has shown that this can be a competitive edge. Right. I am pretty sure that Google is still the only ones that give you near real-time carbon intensity information of the different regions that they operate in. Why is Google doing that? In my view, they're doing that because they are probably still number three in terms of enterprise cloud adoption. Like it's AWS and Azure, depending on which stats you look at, are the leading two. Right? Google is still trying to compete. So in order to compete, I think they are offering more transparency. They're offering more options around sustainability than the other two are, arguably. And so I think this can be a competitive edge. I do wonder whether if Google went a bit further in the near future, the EU companies might all of a sudden go, do you know what? We're just going to adopt Google Cloud because they just give us the data we need for regulatory compliance, for example. So I think the regulatory lever can be really powerful. But you're right, it probably can be too big a stick at times when actually industries can get ahead of that by just saying, do you know what? We'll just be a bit more transparent, a bit more open, a bit more sensible and pragmatic about how we operate, and then that way the regulators won't have to force this out of us.
Gael 17:58
I think it's pretty obvious that the European Union has a... I don't remember who coined the word, but I really love it. The long arm of EU regulations that you also see in California, etc. That is when the EU regulation actually doesn't stop at the European border, and it goes pretty much everywhere in the world, as we've seen with the privacy laws. First, the European Union is not as powerful economically speaking as it is today, even still a very, very big market, obviously. But what about the UK? Because you're both based in the UK. The UK is not a European Union member anymore. But on this specific aspect, do you think that whenever the European Union regulates, somehow it will lead the UK to adopt a similar pattern? Or is it a different way of doing things in the UK now?
Oliver 18:48
So I think the reality is, as much as the UK thinks it's left the European Union, the reality is we still live in Europe and Europe is one of our biggest trading partners. And so, therefore, whatever Europe does, we'll almost certainly have to follow in some regard. So I think we haven't abandoned GDPR, for example. We still have that in our law, a very similar law to that. And I think it's a matter of how much we want to sort of stay in harmony with our biggest trading partner or not. And I think we've got political changes probably happening in the next few months for the UK, and it will be interesting to see quite what happens there.
Gael 19:23
Regarding the link that you made between kind of transparency, a competition I would say on one hand, and regulation on the other, I think it's a very interesting point that one will lead the other. Do you believe that, except maybe for Google, there are other cases where transparency could become a competitive advantage?
Oliver 19:42
I definitely think hardware, like the end user hardware. I honestly think if there are organizations that are far more transparent about the supply chain and the kind of full lifecycle analysis of their products, and they do that in a way that you don't have to fight free PDFs and extract information from data sheets, but actually, they perhaps provide an API or an open standard or something where you can just get that information. I think that fairly soon will be a competitive advantage. And so I think the first organizations that do that will win. I also look at companies like Fairphone that are more modular and repairable with the right to Repair Act that the EU has also just put out. I honestly think that there is scope for innovation in repairability, and it will be really interesting to see, for example, how Apple responds to that regulation. Do they do it to the minimum or do they go beyond and be more transparent about the supply chain and so on at the same time, I think that would be amazing if someone chooses to innovate on that angle? And I think they would find a lot of success in that space, in honesty.
Gael 20:45
We'll see. I think it's some kind of pattern. Usually with Apple, it's like they pour tens of millions of dollars in lobbying against, then they drag their feet a bit, and then when everyone starts to follow, they put tens and millions of dollars on moving forward and then claiming they're the best. And I simply don't understand why they don't skip the two first parts of the dance. But that's a different story. So we talked a lot about hardware, data transparency, methodology, and transparency. But Katie, you mentioned before the correlation, not that obvious correlation, between open source software and sustainability. And please could you elaborate a bit on it?
Katie 21:25
Yeah, of course. So as I was saying, open source projects can often be adopted widely by many organizations and can also be extended or used in their own projects. So I think due to this scale, sort of any efficiency gains or sustainability improvements that are made to the code can have a real impact downstream. So like a ripple effect. What might seem like quite a small optimization, like in the core code base, can have massive impacts downstream on the community. So sort of reduce the environmental impact across the ecosystem. So sort of reduce the environmental impact across the ecosystem, baking in these sustainable practices into the actual open source code. In the same way, sorting other nonfunctional requirements like security have already been, would be hugely beneficial, something that we're missing at the moment.
Gael 22:19
Do you have some examples of maybe communities, open source communities on a dedicated software starting to pay attention to that? Or is it way too early stage? At the moment, I think it's pretty early.
Oliver 22:32
Right. But the comparison I would make to this is a bit like when you buy your laptop, what power profile does it ship with? Right? Does it ship with the high-performance profile enabled by default or the power-saving profile enabled by default? And I would encourage all open source software contributors and maintainers to think about that same sort of thing. When someone downloads your software or includes their library in their software, what mode is it running in by default? And does someone have to specifically go and tune it for either efficiency performance or security? Because for security now it's pretty frowned upon, right? If you ship your open source software in a way that's unsecure by default, I would say that Katie was encouraging us to think about our open source projects and making them run efficiently by default. And maybe if you really need to crank the performance out of it, then, yeah, you go and tune, and you go and make optimizations. But that is something I think that could have a massive impact. And Katie's right, the scale at which open source libraries and software are deployed means that any savings that we make there are going to massively outweigh perhaps any optimizations we might make on our enterprise code. So we do a lot of work with big government and big financial services customers, but even still internal applications for them are hundreds of users, maybe thousands of users. They're not the millions of users that open source projects have. So where do you focus your optimization efforts as an open source developer? Do you optimize your own code that's running on your company systems, or do you optimize open source? So I would say as far as coding efficiency goes, think about what code you're optimizing.
Gael 24:08
And especially knowing that we have only to target like 5% of open source developers because it's like 5% contribute to almost 90 to 99% of the code base. So we can target very active developers. My question would be, is it a bit different as an end user, obviously a developer, but downloading some kind of libraries, whatever, et cetera, do I have today all the information? And honestly, the answer is no. So I'm kind of self-answering my question, but what would it require being able to understand, okay, this is kind of a frugal design or low carbon open source code. There is nothing in I'm pressure like 99.99% of the MD in GitHub or any other repository that doesn't mention the carbon footprint at all or any kind of environmental approach. There is this project, I'm kind of thinking out loud at the moment, but is the impact framework provided by the Green Software Foundation the right approach? Or any other tools that could be used or leveraged to access this information. What do you think about this?
Oliver 25:20
I think the impact framework shows potential to have a standardized way of calculating things. It's very early and we're excited. We're going to be part of carbon hack this year and that's all about the impact engine framework. And we're going to be, for example, probably looking at how you measure S3, the impact of S3, the Amazon S3 service using impact framework. I think what we need though is more of a standard, and you've perhaps given me a bit of an idea here of where the carbon standard that we launched recently could go, right? I mean, we've primarily designed it as a way of navigating the technology space and working out where your emissions live in the upstream operational and downstream sort of categories that we have sort of touched on throughout this episode. But actually I do wonder whether there is almost a metric that you should be looking for in each of those. And I think we're signposting open source tools and standards. But I do wonder if ultimately that might be where we lead. Is giving people an easy, almost like an eco label for software or a bit like the calorie counting you get on food, right? We need something as simple as that for someone to go red, amber, green on a library or a piece of software. We're not there yet, but I hope we get there.
Gael 26:29
Yeah, I think that was the idea behind the API green score, main french companies. It's an open source approach as well, but it's more like a best practices checklist. Like, okay, is my API management clean? Do I provide information on carbon, et cetera? And then you've got a score. I kind of like this approach. It's very straightforward still, it requires pretty transparent methodology and data to be sure that we compare Apple with Apple. Once again, I'm really about fruit salad. Katie, as a developer, because you are, what would you like to have when you use open source code to make sure that you're using a very efficient one?
Katie 27:12
I think the idea that you discussed, Oliver, was really quite a good one about sort of the comparisons to calorie counter. I think having that sort of metric, just so when you're looking and researching what you actually want to use in your code, you don't really have to dig in too much. That would be really useful if you could just see high level. What sort of sustainability level has this library got, for example? And for a developer, you want to be looking at how you can code. So I think learning from how upstream dependencies, open sourced libraries have been coded efficiently is a really good way to learn.
Oliver 27:49
I think that exemplar approach is really interesting. It would be nice to see those sort of standards sort of applied. I think in most cases you could argue they are right, because lots of open source libraries have been optimized for performance, which means they should be fairly efficient. But that's necessarily, that's a big assumption. Right? And so I think what we're talking about here is something that validates that and ensures that there isn't excessive memory usage or excessive CPU consumption, perhaps polling anti patterns. I think maybe we aren't that far away from maybe being able to take some of the Green Software foundation principles and standards and patterns and maybe running some tests because that's the other part of this. We talk a lot about coding, but the testing is really important as well. And so we're equally looking at other software development lifecycle roles, like testing to say, should there be a test suite for efficiency, for energy consumption of code? So part of your pipeline, your builds break if energy consumption is over a certain level or efficiency is not hitting the bar. So this is the sort of bigger picture software development lifecycle thinking we're also doing beyond just sort of classifying emissions into their different buckets. It's also thinking about how does each role play its part when it comes to building more sustainable software.
Gael 29:07
I'm very enthusiastic about what you describe, and actually I know that there are a few projects in CI CD pipeline trying to automate it. So, full disclosure, I've launched a project with my good friend Benoît Petit at the Boavizta association, but that's not public yet. But we've released the version zero of our repository of green it tools because we want to increase the transparency in the landscape. So the idea is not to assess if the tool is good or bad, but just to assess how transparent is the methodology, the data used, and all the information that will help people choose the right tool. And that's something that I will at Green IO Singapore and in forthcoming podcasts. It's really a transparency battle, it's not an assessment quality assessment battle, and a good practical exercise to close the podcast. How would you launch this open source initiative if you launch it?
Oliver 29:59
Well, we'd certainly come back on the podcast if you have us. Katie, what are your thoughts? You're more in the development world than I am these days, I think in.
Katie 30:09
The sort of same way as projects like Green Software Foundation, I think sort of co-pilot tool. It's the same sort of thing. So the way that Green Software foundation have sort of promoted their principles and their green coding patterns sort of promote the tool in the same way.
Gael 30:28
But when you launch an open source, mean you have an idea you want to launch. Obviously you will create a GitHub repository and the license, blah blah blah. But hey, you're still two or three of you. I don't know how many people actually, how many people will join from Scott logic for that kind of project? Will it be only the two of you, or are you?
Oliver 30:53
We have a team of about ten people working in this space at the moment, but people have rotated in and rotated out as we're a consultancy, right? So when we have client projects, people will go on to client work. So it very much depends on what client projects are going on at any one time as to how many people are involved in our R & D initiatives. But ultimately we're rolling out sustainable software thinking across the whole of the company and that's 500 people in total, about 450 or so consultants. Not all of those are developers, of course, but this mindset and these best practices are being pushed out across the organization because we don't want this to be a bolt on, we want this to be how we just work by default back to the sensible defaults thing. So I think yeah, for launching it we'd almost certainly launch it internally, we'd alpha test it ourselves and then I think we would look to kind of work with friendly clients and then kind of use open source ecosystems. We're part of like Linux foundation and Finos. So shout out to Finos that we do a lot of work with. They're a great organization for financial services, open source adoption. So I think yeah, it's kind of use the community, right? I think would be the straightforward answer to that one because that's the beauty of the open source effort, right, is it's always community driven. So kind of embrace that community approach.
Gael 32:06
And does it respond well, the community, to that kind of project?
Oliver 32:10
Well, it's interesting, right? My presentation at OpenUK, I would have hoped for a slightly larger audience, but I was competing with AI. There's a lot of AI talks and everyone, I think if you put an AI talk on everyone wants to go to that talk right now because it's just so hot. It's kind of ironic that everyone wants to go and listen to the really power hungry, hardware hungry topic of the day rather than talk about how we can tread more lightly. I do fear there is a culture war, Gael, happening like some of the tech meetups I've been going to recently there's almost two camps, there's people that are almost falling for the techno optimist manifesto from Marc Andreessen Horowitz and just think, growth and energy consumption and increasing energy consumption isn't a problem. Why is it problem? And there are others who are more aware of the issues that we have around. We don't have endless power sources that are renewable.
Gael 32:58
Sadly though, the trick, and I will do it in London again, is that my keynote speaker for closing the day was Théo Alves da Costa and he's the co-founder of data for good, and he's the one who trained the ChatGPT for climate. I think it's climate Q&A, but they didn't use the ChatGPT tool, the model they use. Another one, I think it was. I don't remember which one it is. S hungry energy hungry model. But anyway. And the idea, know if you train a chatbot on IPCC, report only usually the answer to the question is pretty good. So that's a good usage. And so I managed to get some people in there because it was artificial intelligence and sustainability. So we will do it in London again, don't worry. But anyway, Katie, some thought on what we've just discussed.
Katie 33:54
I think what you've just been discussing there about the sort of war between AI consuming so much energy and sustainability. It's really interesting because I think they're both so forefront at the minute, not so invested as I am in sustainability. For example, I think that my sort of passion for sustainability has really grown last few months, especially just even coming to terms with the terminology and the language of it. It's a whole world really, that I'd never really thought about. It's really shifted my thinking, especially not just within development, but everyday life.
Oliver 34:30
I'm living a Jekyll and Hyde personality day by day at the moment. Right, Gail? Because I am spending almost equal amounts of time on AI R&D and sustainability R&D. So at some point the two streams will cross, and it's interesting to hear about Climate Q&A. That's great. I didn't know that existed. So thank you. And almost, yeah, what we'd love to do is almost to create a climate Q&A for technologists. So that the bit like we were just talking about earlier, the know, that's the sort of ultimate goals we have at AI for good. But yeah, I do struggle with this. And in fact, I've got a blog in drafts that I'm not sure I'll publish. But it's sort of talking about this conflict I have between technological progress using AI and on the other hand, all the power consumption and the sustainability impacts that that will have. So I do have this sort of schisms in should we be embracing this thing or should we be being more cautious? It's a really fascinating time, I think, to be a technologist right now, like the different challenges and opportunities that we have in front of us, it's incredible. Yeah.
Gael 35:32
And I fully agree with you. And I think actually it's also a good thing not to be made in one piece of wood, I would say. And because our world is complex, it requires subtility. And even if it's not very comfortable coming up with subtle approaches to the big questions that we have, and not having one single answer for every question. I think it's a good approach, whether it's AI or. I don't know how to say it, but yeah, kill the tech or whatever approach, or go back to the stone edge, I don't know. I don't want to use it as a caricature. But you see my point. We're thinking complex, and it's very uncomfortable because I think we are more and more aware of how complex is the world we live in. And that's just us adapting to the reality of our world. So quite a lot has been covered here. To close the podcast, as usual, I would love both of you to share one piece of good news. Sustainably related, even green it related, but sustainably at large is cool. Something that makes you happy about our path towards a more sustainable world, I would say.
Katie 36:44
I think for me because I'm quite new to the tech space, especially the technology sustainability space. I think just seeing the community growing, it's really positive. Seeing all the initiatives that are out there, whether they're open source. I think just like seeing the work that our team's been doing as well, the tech carbon standard, I think it's all really positive. It's going in the right direction.
Oliver 37:08
I'm going to choose something that's not tech, but it's energy. So I'm probably as much an energy nerd as I am a technology nerd. And the thing that really excited me the other day was the announcement of quite a big electricity interconnect between Denmark and the UK. Right? And I hadn't heard about this, but a huge investment, like we're talking billions of pounds or euros investment to create this undersea link between Denmark and the UK. And it makes a lot of sense because both Denmark and the UK have a lot of wind power, and because of the time difference between Denmark and the UK, our peak electricity demands are at a different time. So there's a lot of sense in this interconnect, because when it's windy in Denmark and there's lots of demand in the UK, they can send their wind power to the UK and vice versa. And so I didn't know this thing was even being built. And it's gone live literally in the last couple of weeks. And it's significant, I think it's like at least a gigawatt of interconnect. So a serious DC power interconnect between the two countries. And that just made my day, because I knew we've been rolling back on various sorts of environmental policies, but this has just gone live and I didn't even know they were building it.
Gael 38:09
And that's really, that's actually, that's so fun because I stumbled on a map of all these new connects being built across Europe, both for sustainability reasons, also for security reasons, after the Russian aggression against Ukraine and all the energy issue that it raised. And that was mind-blowing that there is already a lot of being built and even more being planned. And as you say, that's just perfect. It's sharing energy and low-carbon energy as much as you can. So I really love it. I think I will try to find the map and put it in. That'd be great. Thanks for sharing it. That made my day as well. Super cool. So it was great. I will put all the references in the show notes as usual. And what is also super cool, is that there is like a very good deal of chance that we will meet in London in September. So thanks for joining. Talk to you soon. Keep up this amazing work with a very open source technology, carbon standard but sustainable pilot idea. I love it. So let's stay in touch. Thanks a lot for joining and have a very nice and sustainable day.
Katie 39:17
Thank you.
Oliver 39:18
Thanks for having us.
❤️ 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.
47 epizódok
MP3•Epizód kép
Manage episode 403307176 series 3346125
A tartalmat a Gaël DUEZ biztosítja. Az összes podcast-tartalmat, beleértve az epizódokat, grafikákat és podcast-leírásokat, közvetlenül a Gaël DUEZ vagy a podcast platform partnere tölti fel és biztosítja. Ha úgy gondolja, hogy valaki az Ön engedélye nélkül használja fel a szerzői joggal védett művét, kövesse az itt leírt folyamatot https://hu.player.fm/legal.
🔍How open source software slashes energy consumption, extends hardware lifespan, and drives sustainability?
🎙️In this episode, Gael Duez unravels the intricate relationship between open source software and IT sustainability together with Oliver Cronk, the host of Architect Tomorrow podcast and Sustainable Technology practice lead at Scott Logic (UK), alongside his software engineer colleague Katie Davis.
💡Discover how open source software (OSS) and hardware (OSH) are not just benefiting companies, but also driving global sustainability efforts. With insights from a recent HBR study revealing OSS's monumental impact on reducing software development costs, the stage is set for a game-changing conversation. ⚙️
💻Katie breaks down the key to OSS success: transparency. Learn how freely available code empowers developers to make efficient, sustainable choices, while Oliver highlights the importance of common hardware standards and circularity to combat waste and inefficiency.
📊Dive into the battle for open-source data transparency with Katie, and explore Oliver's vision for the tech community's future direction. From reducing emissions with initiatives like the Tech Carbon Standard to innovative projects like the Open Compute Project, the episode is packed with actionable insights and inspiring solutions.
❤️ Subscribe, follow, like, ... stay connected the way you want to never miss our episode, twice a month, on 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 next Conference is in Singapore on April 18th (use the voucher GREENIOVIP to get a free ticket)
📣 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:
📧 You can also send us an email at greenio@duez.com to share your feedback and suggest future guests or topics.
Oliver and Katie's sources and other references mentioned in this episode:
- Scott Logic’s Technology Carbon Standard
- Harvard Business School Strategy Unit Working Paper No. 24-038 on Open Source Software
- Open Compute UK
- Tristan Nitot's talks about eroom
- GSF Carbon Hack
- Boavista Cloud Scanner
- Cloud Carbon Footprint
- EU Energy Efficiency Directive
- Climate Q&A
- Built and planned electricity Interconnect across Europe
- Viking Link: Interconnector from Denmark to Great Britain
- Reporting requirements on the energy performance and sustainability of data centres for the Energy Efficiency Directive
Transcript
Gael 00:00
Hello everyone. Welcome to Green IO, the podcast for responsible technologists building a greener digital world, one byte at a time. Every two Tuesdays, our guests from across the globe share insights, tools, and alternative approaches, enabling people within the tech sector and beyond to boost digital sustainability. So today we're going to talk about a relationship, a deep, pivotal, and yet troubled relationship. Open source and IT sustainability every person I talk in our sustainability field takes it as a no-brainer. Open source is good for sustainability. But which open source are we talking about? And is this link? Is this correlation so strong? For instance, is open source software always more frugal or sustainable in the way it is designed?
We don't know. Let's investigate it. But there is at least one certain thing. Open source software is big. A recent post made by Sasha Lucioni, one of my absolute favorite thought leaders in AI, caught my eye. A recent HBS study found that companies would have to spend 3.5 times more on their software development if open source were not around. And we're talking about an $8.8 billion estimated market. So open source is big. Let's see how it connects with sustainability.
Today I am delighted to have two guests to talk about sustainability. Who has done open sources from the trenches? Oliver Cronk is a fellow YouTuber. Now I can say because, yes, Green IO launched its YouTube channel a week ago to offer a wider choice for accessing its content than traditional podcast platforms like Apple or Spotify. Anyway, Oliver is the Architect Tomorrow host, a channel for enterprise and platform architects with a soft spot for sustainability, among other topics. He's also a tech strategist at Scott Logic, a software and data consultancy based in the UK, and has a significant track record in both IT and energy. As an example, he built carbon calculators, air quality databases, and industrial emissions reporting systems, which made him experience firsthand the issue of accessing open, transparent, and reliable data. Katie is a software engineer with a Math’s Degree, so not easy to mess with her when it comes to data and statistics, and she is a driving force in this sustainable technology practice at Scott Logic, especially involved in the recent release of the technology carbon standard at open UK several weeks ago. A standard under the Creative Commons license, of course. By the way, they will present it at Green IO London on September 19 this year, but that's a different story. Welcome, Oliver. Thanks a lot for joining Green IO today. So, are you on your bike or more comfortably sitting at your desk?
Oliver 03:02
No, today I'm stationary. I'm at my desk.
Katie 03:04
Thanks for having us.
Gael 03:06
You're welcome. That's nice to have you on the show. And welcome, Katie. Thanks a lot for joining Green IO today as well. That's vegetable time, not food time. I guess because you love to garden and to grow your own food. That's a very nice angle for sustainability. Katie, maybe the first question will be for you, with your academic track record. When we talk about open source, it can actually cover a lot of different angles. Could you give us your definition or your way of approaching what is open source?
Katie 03:40
Of course. So when I think about open source, I always think about open source software. So this is software whose source code is freely available to modify, to distribute, deploy, and even extend it into their own projects. So I think one of the key parts of open source is that it's transparent. You can view the methodology behind the code. There are also types of open source, which I think maybe Oliver could expand on.
Oliver 04:08
Yeah, no, happy to. So I think you've touched on open data, which is an important one. But then open hardware is really interesting as well when it comes to sustainability. So at the open UK conference, you mentioned, Gail, there was a great presentation about open compute, the open hardware sort of initiatives to ensure that there are sort of standards for hardware. And there are companies now taking the used hardware out of the big hyperscaler data centers and reusing it because it still has a life. Right? The big hyperscalers no longer want it, but you can easily get additional life out of it by hosting it in another data center and getting other people to run workloads on it. So there's a whole host of different strands of the open source kind of ecosystem, really. Software, hardware, data, significant ones.
Gael 04:48
I think that would be a good way to define these pillars of data, software, and hardware. So, Oliver, you started with hardware, not maybe the most obvious when we are in the IT world. What are the progress today that you've seen in open source hardware?
Oliver 05:05
Yeah, I'd love to throw to Katie to talk about the sort of data that's available, hopefully in the future, in more of an open form about hardware. But in terms of hardware itself, I think there are some really interesting developments. Like I talked about, the open compute project, they have a whole sustainability division that's looking at the circularity of the hardware. I think having open standards when it comes to hardware is important because otherwise, it's the compatibility challenge. Right. If everyone is inventing their hardware to kind of be unique and proprietary, it just means we have a whole lot of potentially wasted components that won't work with other systems and we're producing lots of different types of components that only work with one particular platform. So I'm a really big proponent of the open compute project also because they're looking at things like can we increase the ambient temperature of data centers, for example, to reduce their cooling requirements and things like that. So they're doing some really clever stuff in that area. But I think also open hardware is interesting when you look at the sort of hardware hacker sort of culture that's created as well. Right? What can you run on an embedded device like a Raspberry Pi? How can we extract every last efficiency out of a couple of watts that a Raspberry Pi runs on? So I think there's some really interesting sort of things that come out, the constraints that you impose when you look at sort of consumer grade hardware and the open source movements around that.
Gael 06:20
Yeah. And just to not close, but to comment on the hardware topic, it really resonates with the new battle of my good friend Tristanito, who's one of the founders actually, of Modzia Europe. So he's been in the game for quite a while and he's got his amazing talk now that he's given several conferences already about Moore's law is dead now it's eroom. So that's kind of reversing the Moore name. The idea is every year you don't double the size of your computing power, every year you divide by two the size of your code. So every year you can run your code on lighter and lighter machines, which means older and older machines, which means saving potentially millions, avoiding building new computers that actually don't need. But anyway, Katie, you wanted to elaborate a bit more on the data side. I think this is one of your main battles, isn't it?
Katie 07:18
Yeah. So especially with the sort of manufacturers of hardware, the data they provide themselves, often the figures they're given, these, it's not clear on how they actually calculate them. So for example, if you wanted to find out the typical energy consumption, they might provide this figure, but it's hard to find out where they're getting this from. How long are they running the device to arrive at this figure? Additionally, there's another sort of documentation that these manufacturers provide, like product carbon footprint, which gives the emissions from each use stage of the lifecycle. So manufacture transportation usage, end of life. But again, it's just not clear how they get these figures. So the benefit of open source is having that methodology transparent to everyone to see how these figures are arrived at.
Gael 08:06
Yeah, the big issue if you've got quite diverse hardware estate is that we might add apple and orange because the way one manufacturer would calculate things is not the same as the others. And eventually, these numbers don't make sense at all. What do you see as a potential path to overcome this very big difficulty in data sustainability?
Katie 08:34
I think open source is key, just having a sort of level playing field for the methodology that we use to calculate these things. So I mean, I think we'll probably discuss it later, but like tools like Cloud Carbon Footprint, sort of having that methodology that you can use across cloud providers, so it's more comparable, easier to track your improvements across the estate. And I mean, that's just one part of the standard.
Oliver 08:59
Katie touched on a really important point there, which is at the moment because the manufacturers can essentially almost makeup, I'm exaggerating a little bit to make a point, but because they can almost make up their methodology, their numbers, it almost becomes a marketing exercise to make it look like their product consumes less energy. And it's a bit of a game, right? So if you come up with a really creative way of measuring the power consumption of your device, I mean, we've seen servers, for example, with different sort of load ratings and energy consumption, different load ratings, and each manufacturer seems to sort of measure at different loads. So some are at 20%, some are 80, some are 100. It's like, okay, well, you can make yourself look cleaner by just presenting the characteristic that your model of hardware tends to perform best at. And of course, then you don't really know whether you're making an improvement or you've just bought better marketing when you change suppliers. So I think that's why what Kate is saying is really important. This apples-for-apples comparison is super important because we already have the same with the cloud. Like you run AWS's cloud carbon tooling versus Azure's versus Google's, they all have different methodologies, and so they're not comparable.
Gael 10:04
And that's a point with carbon footprint. And I'm very happy you raised both of you, actually, you raised the point. Is that true? It's a very powerful tool. Cameron was on the show last year. I'm a big advocate of CCF, among other open source tools, but at the end of the day, they have to recalculate things that should be provided by hyperscalers. And I think here we've got two very serious issues. The first one is the methodology. Tools like open source tools like Boa Vista cloud scanner or CCF, obviously the methodology is transparent, so at least we know how they manipulate the data. But then there is the issue of the data itself. I mean, do we have the granularity to calculate things in a wise enough way? And I'm not always sure, because for cloud. But Katie, I will let you elaborate on it. For cloud, it's always a question of transforming financial data almost all the time, financial data, into sustainability data. And some data would be more, I would say, accurate, like starting with CPU usage or GPU usage or whatever. But what's your thought on it, Katie?
Katie 11:17
So with regard to cloud, one of the key issues is that each cloud provider has their own tooling, but the methodology is not the same. So I think one of the main challenges is to do with sort of greenhouse gas scopes. So I think AWS, especially, doesn't include scope three. I think that one of the benefits of the cloud carbon footprint tool is that it provides a way to at least try and estimate that proportion allocated to scope three and sort of compare on the same level with those from Google as you are. I think Google does provide data on scope three emissions. So I think it's sort of just picking the right methodology for your estate. So, for example, if you just had Google Cloud, maybe you just go with their tool, since it's probably more up-to-date and includes scope three. Whereas, if you had multi-cloud or AWS, you might want to look at a tool like Cloud Carbon Footprint.
Oliver 12:14
But I think the point, Gael, around using financial numbers to come up with sustainability metrics is an interesting one. Clearly, it'd be great if we could trust the native tool, like Katie said, to use the Google tool, the AWS tool, because they have access to the underlying data that allows them to more accurately calculate the emissions. The trouble is, there's a trust thing there, isn't there? I think it's how transparent are they being and how much we trust them to kind of calculate this fairly, rather than just putting forward a polished view of their emissions, a managed kind of view of their emissions rather than reality. And so this is where I think the open approaches really would help because if they were open and transparent about what went into providing the service that you pay for, you would be able to make a more informed decision. And I think the calculations would be far more robust than basing them just on billing data. Ideally, you find situations where you save money, and you save carbon, but those two aren't necessarily going to go hand in hand in all cases.
Gael 13:12
Yeah, absolutely. I had an issue with a client very recently when we made the calculation that the bill would go up by 20%. And then you have also the question of how do you calculate a clean energy region. Because I don't know if you've seen this beautiful post from Mark Butcher migrating from Scotland to Ireland, where almost a time six difference in the carbon intensity of the electricity grid. If you do very basic math, you will still double your carbon footprint. Still, I would like to go back to it. Okay, so we claim that open source and open data because we are 100% on this open data subpart of open source here should be there, but the question is why not? One of the pushbacks I've heard from Isposkella representatives is like the business secrets and that they don't want to share because it's sensitive data. But what's your thought on it? And do you believe there are other obstacles, I would say on the road toward mods transparency, at least from the data perspective?
Katie 14:18
I think that sort of garden data for a commercial reason, especially with regards to sustainability, is just not really the way forward. I think to sort of make sustainable software key and forefront, we really need to be transparent, so people can build on methodology. It's not really any good sort of reinventing the wheel when there are already so many good established, like data point methodologies out there. I think as well, open source projects, have such a large user base, especially compared to some enterprise software, especially internal sort of company software, not always, but sometimes. So I think any efficiency gains that we make to open source software can have a real downstream impact.
Oliver 15:05
Yeah, Casey's kind of moved on to one of the other key kind of conversation points we have around open source. Right. But before we go there, my take on what you're saying, I think the commercial pushback is an interesting one, and I'm not convinced it's as simple as if we just reveal some more data, and all of a sudden the cat is out of the bag and everyone will know how we're running our data center. I think the reality is running a modern data center, certainly for a hyperscaler, is super, super complex. Right? And they're even using AI to optimize. Google and Meta, for example, have used AI to optimize their cooling and stuff like that. So I think to think that if you just release some data about your energy consumption and the high-level waste footprint and water footprint and all that sort of stuff, that that's going to be commercially dismantitious if you release that information, maybe to a point, because it will reveal perhaps to your competitors how efficiently you're running things. But yeah, I think it's just an excuse that they're hiding behind, quite frankly. And I think what will happen if they continue to drag their feet is the EU will regulate. And I think the EU is already regulating. Right. If you look at the data center regulation that's coming out, if you're running a data center over a certain power consumption now or in the near future, you will have to start reporting on more data. So I think that the reality is if companies continue to drag their hills on being transparent, regulation will follow. And it's probably better to kind of get ahead of that.
Gael 16:27
Yeah, you're mentioning the energy efficiency directive, and actually I will put the link in the show notes as everything that we've discussed so far. But is it the only way forward? Is it okay to regulate? I'm a big advocate of better norms and more transparency, but do you believe this is the only trigger that will force big actors to become more transparent? Is there any other way?
Oliver 16:51
I think Google has shown that this can be a competitive edge. Right. I am pretty sure that Google is still the only ones that give you near real-time carbon intensity information of the different regions that they operate in. Why is Google doing that? In my view, they're doing that because they are probably still number three in terms of enterprise cloud adoption. Like it's AWS and Azure, depending on which stats you look at, are the leading two. Right? Google is still trying to compete. So in order to compete, I think they are offering more transparency. They're offering more options around sustainability than the other two are, arguably. And so I think this can be a competitive edge. I do wonder whether if Google went a bit further in the near future, the EU companies might all of a sudden go, do you know what? We're just going to adopt Google Cloud because they just give us the data we need for regulatory compliance, for example. So I think the regulatory lever can be really powerful. But you're right, it probably can be too big a stick at times when actually industries can get ahead of that by just saying, do you know what? We'll just be a bit more transparent, a bit more open, a bit more sensible and pragmatic about how we operate, and then that way the regulators won't have to force this out of us.
Gael 17:58
I think it's pretty obvious that the European Union has a... I don't remember who coined the word, but I really love it. The long arm of EU regulations that you also see in California, etc. That is when the EU regulation actually doesn't stop at the European border, and it goes pretty much everywhere in the world, as we've seen with the privacy laws. First, the European Union is not as powerful economically speaking as it is today, even still a very, very big market, obviously. But what about the UK? Because you're both based in the UK. The UK is not a European Union member anymore. But on this specific aspect, do you think that whenever the European Union regulates, somehow it will lead the UK to adopt a similar pattern? Or is it a different way of doing things in the UK now?
Oliver 18:48
So I think the reality is, as much as the UK thinks it's left the European Union, the reality is we still live in Europe and Europe is one of our biggest trading partners. And so, therefore, whatever Europe does, we'll almost certainly have to follow in some regard. So I think we haven't abandoned GDPR, for example. We still have that in our law, a very similar law to that. And I think it's a matter of how much we want to sort of stay in harmony with our biggest trading partner or not. And I think we've got political changes probably happening in the next few months for the UK, and it will be interesting to see quite what happens there.
Gael 19:23
Regarding the link that you made between kind of transparency, a competition I would say on one hand, and regulation on the other, I think it's a very interesting point that one will lead the other. Do you believe that, except maybe for Google, there are other cases where transparency could become a competitive advantage?
Oliver 19:42
I definitely think hardware, like the end user hardware. I honestly think if there are organizations that are far more transparent about the supply chain and the kind of full lifecycle analysis of their products, and they do that in a way that you don't have to fight free PDFs and extract information from data sheets, but actually, they perhaps provide an API or an open standard or something where you can just get that information. I think that fairly soon will be a competitive advantage. And so I think the first organizations that do that will win. I also look at companies like Fairphone that are more modular and repairable with the right to Repair Act that the EU has also just put out. I honestly think that there is scope for innovation in repairability, and it will be really interesting to see, for example, how Apple responds to that regulation. Do they do it to the minimum or do they go beyond and be more transparent about the supply chain and so on at the same time, I think that would be amazing if someone chooses to innovate on that angle? And I think they would find a lot of success in that space, in honesty.
Gael 20:45
We'll see. I think it's some kind of pattern. Usually with Apple, it's like they pour tens of millions of dollars in lobbying against, then they drag their feet a bit, and then when everyone starts to follow, they put tens and millions of dollars on moving forward and then claiming they're the best. And I simply don't understand why they don't skip the two first parts of the dance. But that's a different story. So we talked a lot about hardware, data transparency, methodology, and transparency. But Katie, you mentioned before the correlation, not that obvious correlation, between open source software and sustainability. And please could you elaborate a bit on it?
Katie 21:25
Yeah, of course. So as I was saying, open source projects can often be adopted widely by many organizations and can also be extended or used in their own projects. So I think due to this scale, sort of any efficiency gains or sustainability improvements that are made to the code can have a real impact downstream. So like a ripple effect. What might seem like quite a small optimization, like in the core code base, can have massive impacts downstream on the community. So sort of reduce the environmental impact across the ecosystem. So sort of reduce the environmental impact across the ecosystem, baking in these sustainable practices into the actual open source code. In the same way, sorting other nonfunctional requirements like security have already been, would be hugely beneficial, something that we're missing at the moment.
Gael 22:19
Do you have some examples of maybe communities, open source communities on a dedicated software starting to pay attention to that? Or is it way too early stage? At the moment, I think it's pretty early.
Oliver 22:32
Right. But the comparison I would make to this is a bit like when you buy your laptop, what power profile does it ship with? Right? Does it ship with the high-performance profile enabled by default or the power-saving profile enabled by default? And I would encourage all open source software contributors and maintainers to think about that same sort of thing. When someone downloads your software or includes their library in their software, what mode is it running in by default? And does someone have to specifically go and tune it for either efficiency performance or security? Because for security now it's pretty frowned upon, right? If you ship your open source software in a way that's unsecure by default, I would say that Katie was encouraging us to think about our open source projects and making them run efficiently by default. And maybe if you really need to crank the performance out of it, then, yeah, you go and tune, and you go and make optimizations. But that is something I think that could have a massive impact. And Katie's right, the scale at which open source libraries and software are deployed means that any savings that we make there are going to massively outweigh perhaps any optimizations we might make on our enterprise code. So we do a lot of work with big government and big financial services customers, but even still internal applications for them are hundreds of users, maybe thousands of users. They're not the millions of users that open source projects have. So where do you focus your optimization efforts as an open source developer? Do you optimize your own code that's running on your company systems, or do you optimize open source? So I would say as far as coding efficiency goes, think about what code you're optimizing.
Gael 24:08
And especially knowing that we have only to target like 5% of open source developers because it's like 5% contribute to almost 90 to 99% of the code base. So we can target very active developers. My question would be, is it a bit different as an end user, obviously a developer, but downloading some kind of libraries, whatever, et cetera, do I have today all the information? And honestly, the answer is no. So I'm kind of self-answering my question, but what would it require being able to understand, okay, this is kind of a frugal design or low carbon open source code. There is nothing in I'm pressure like 99.99% of the MD in GitHub or any other repository that doesn't mention the carbon footprint at all or any kind of environmental approach. There is this project, I'm kind of thinking out loud at the moment, but is the impact framework provided by the Green Software Foundation the right approach? Or any other tools that could be used or leveraged to access this information. What do you think about this?
Oliver 25:20
I think the impact framework shows potential to have a standardized way of calculating things. It's very early and we're excited. We're going to be part of carbon hack this year and that's all about the impact engine framework. And we're going to be, for example, probably looking at how you measure S3, the impact of S3, the Amazon S3 service using impact framework. I think what we need though is more of a standard, and you've perhaps given me a bit of an idea here of where the carbon standard that we launched recently could go, right? I mean, we've primarily designed it as a way of navigating the technology space and working out where your emissions live in the upstream operational and downstream sort of categories that we have sort of touched on throughout this episode. But actually I do wonder whether there is almost a metric that you should be looking for in each of those. And I think we're signposting open source tools and standards. But I do wonder if ultimately that might be where we lead. Is giving people an easy, almost like an eco label for software or a bit like the calorie counting you get on food, right? We need something as simple as that for someone to go red, amber, green on a library or a piece of software. We're not there yet, but I hope we get there.
Gael 26:29
Yeah, I think that was the idea behind the API green score, main french companies. It's an open source approach as well, but it's more like a best practices checklist. Like, okay, is my API management clean? Do I provide information on carbon, et cetera? And then you've got a score. I kind of like this approach. It's very straightforward still, it requires pretty transparent methodology and data to be sure that we compare Apple with Apple. Once again, I'm really about fruit salad. Katie, as a developer, because you are, what would you like to have when you use open source code to make sure that you're using a very efficient one?
Katie 27:12
I think the idea that you discussed, Oliver, was really quite a good one about sort of the comparisons to calorie counter. I think having that sort of metric, just so when you're looking and researching what you actually want to use in your code, you don't really have to dig in too much. That would be really useful if you could just see high level. What sort of sustainability level has this library got, for example? And for a developer, you want to be looking at how you can code. So I think learning from how upstream dependencies, open sourced libraries have been coded efficiently is a really good way to learn.
Oliver 27:49
I think that exemplar approach is really interesting. It would be nice to see those sort of standards sort of applied. I think in most cases you could argue they are right, because lots of open source libraries have been optimized for performance, which means they should be fairly efficient. But that's necessarily, that's a big assumption. Right? And so I think what we're talking about here is something that validates that and ensures that there isn't excessive memory usage or excessive CPU consumption, perhaps polling anti patterns. I think maybe we aren't that far away from maybe being able to take some of the Green Software foundation principles and standards and patterns and maybe running some tests because that's the other part of this. We talk a lot about coding, but the testing is really important as well. And so we're equally looking at other software development lifecycle roles, like testing to say, should there be a test suite for efficiency, for energy consumption of code? So part of your pipeline, your builds break if energy consumption is over a certain level or efficiency is not hitting the bar. So this is the sort of bigger picture software development lifecycle thinking we're also doing beyond just sort of classifying emissions into their different buckets. It's also thinking about how does each role play its part when it comes to building more sustainable software.
Gael 29:07
I'm very enthusiastic about what you describe, and actually I know that there are a few projects in CI CD pipeline trying to automate it. So, full disclosure, I've launched a project with my good friend Benoît Petit at the Boavizta association, but that's not public yet. But we've released the version zero of our repository of green it tools because we want to increase the transparency in the landscape. So the idea is not to assess if the tool is good or bad, but just to assess how transparent is the methodology, the data used, and all the information that will help people choose the right tool. And that's something that I will at Green IO Singapore and in forthcoming podcasts. It's really a transparency battle, it's not an assessment quality assessment battle, and a good practical exercise to close the podcast. How would you launch this open source initiative if you launch it?
Oliver 29:59
Well, we'd certainly come back on the podcast if you have us. Katie, what are your thoughts? You're more in the development world than I am these days, I think in.
Katie 30:09
The sort of same way as projects like Green Software Foundation, I think sort of co-pilot tool. It's the same sort of thing. So the way that Green Software foundation have sort of promoted their principles and their green coding patterns sort of promote the tool in the same way.
Gael 30:28
But when you launch an open source, mean you have an idea you want to launch. Obviously you will create a GitHub repository and the license, blah blah blah. But hey, you're still two or three of you. I don't know how many people actually, how many people will join from Scott logic for that kind of project? Will it be only the two of you, or are you?
Oliver 30:53
We have a team of about ten people working in this space at the moment, but people have rotated in and rotated out as we're a consultancy, right? So when we have client projects, people will go on to client work. So it very much depends on what client projects are going on at any one time as to how many people are involved in our R & D initiatives. But ultimately we're rolling out sustainable software thinking across the whole of the company and that's 500 people in total, about 450 or so consultants. Not all of those are developers, of course, but this mindset and these best practices are being pushed out across the organization because we don't want this to be a bolt on, we want this to be how we just work by default back to the sensible defaults thing. So I think yeah, for launching it we'd almost certainly launch it internally, we'd alpha test it ourselves and then I think we would look to kind of work with friendly clients and then kind of use open source ecosystems. We're part of like Linux foundation and Finos. So shout out to Finos that we do a lot of work with. They're a great organization for financial services, open source adoption. So I think yeah, it's kind of use the community, right? I think would be the straightforward answer to that one because that's the beauty of the open source effort, right, is it's always community driven. So kind of embrace that community approach.
Gael 32:06
And does it respond well, the community, to that kind of project?
Oliver 32:10
Well, it's interesting, right? My presentation at OpenUK, I would have hoped for a slightly larger audience, but I was competing with AI. There's a lot of AI talks and everyone, I think if you put an AI talk on everyone wants to go to that talk right now because it's just so hot. It's kind of ironic that everyone wants to go and listen to the really power hungry, hardware hungry topic of the day rather than talk about how we can tread more lightly. I do fear there is a culture war, Gael, happening like some of the tech meetups I've been going to recently there's almost two camps, there's people that are almost falling for the techno optimist manifesto from Marc Andreessen Horowitz and just think, growth and energy consumption and increasing energy consumption isn't a problem. Why is it problem? And there are others who are more aware of the issues that we have around. We don't have endless power sources that are renewable.
Gael 32:58
Sadly though, the trick, and I will do it in London again, is that my keynote speaker for closing the day was Théo Alves da Costa and he's the co-founder of data for good, and he's the one who trained the ChatGPT for climate. I think it's climate Q&A, but they didn't use the ChatGPT tool, the model they use. Another one, I think it was. I don't remember which one it is. S hungry energy hungry model. But anyway. And the idea, know if you train a chatbot on IPCC, report only usually the answer to the question is pretty good. So that's a good usage. And so I managed to get some people in there because it was artificial intelligence and sustainability. So we will do it in London again, don't worry. But anyway, Katie, some thought on what we've just discussed.
Katie 33:54
I think what you've just been discussing there about the sort of war between AI consuming so much energy and sustainability. It's really interesting because I think they're both so forefront at the minute, not so invested as I am in sustainability. For example, I think that my sort of passion for sustainability has really grown last few months, especially just even coming to terms with the terminology and the language of it. It's a whole world really, that I'd never really thought about. It's really shifted my thinking, especially not just within development, but everyday life.
Oliver 34:30
I'm living a Jekyll and Hyde personality day by day at the moment. Right, Gail? Because I am spending almost equal amounts of time on AI R&D and sustainability R&D. So at some point the two streams will cross, and it's interesting to hear about Climate Q&A. That's great. I didn't know that existed. So thank you. And almost, yeah, what we'd love to do is almost to create a climate Q&A for technologists. So that the bit like we were just talking about earlier, the know, that's the sort of ultimate goals we have at AI for good. But yeah, I do struggle with this. And in fact, I've got a blog in drafts that I'm not sure I'll publish. But it's sort of talking about this conflict I have between technological progress using AI and on the other hand, all the power consumption and the sustainability impacts that that will have. So I do have this sort of schisms in should we be embracing this thing or should we be being more cautious? It's a really fascinating time, I think, to be a technologist right now, like the different challenges and opportunities that we have in front of us, it's incredible. Yeah.
Gael 35:32
And I fully agree with you. And I think actually it's also a good thing not to be made in one piece of wood, I would say. And because our world is complex, it requires subtility. And even if it's not very comfortable coming up with subtle approaches to the big questions that we have, and not having one single answer for every question. I think it's a good approach, whether it's AI or. I don't know how to say it, but yeah, kill the tech or whatever approach, or go back to the stone edge, I don't know. I don't want to use it as a caricature. But you see my point. We're thinking complex, and it's very uncomfortable because I think we are more and more aware of how complex is the world we live in. And that's just us adapting to the reality of our world. So quite a lot has been covered here. To close the podcast, as usual, I would love both of you to share one piece of good news. Sustainably related, even green it related, but sustainably at large is cool. Something that makes you happy about our path towards a more sustainable world, I would say.
Katie 36:44
I think for me because I'm quite new to the tech space, especially the technology sustainability space. I think just seeing the community growing, it's really positive. Seeing all the initiatives that are out there, whether they're open source. I think just like seeing the work that our team's been doing as well, the tech carbon standard, I think it's all really positive. It's going in the right direction.
Oliver 37:08
I'm going to choose something that's not tech, but it's energy. So I'm probably as much an energy nerd as I am a technology nerd. And the thing that really excited me the other day was the announcement of quite a big electricity interconnect between Denmark and the UK. Right? And I hadn't heard about this, but a huge investment, like we're talking billions of pounds or euros investment to create this undersea link between Denmark and the UK. And it makes a lot of sense because both Denmark and the UK have a lot of wind power, and because of the time difference between Denmark and the UK, our peak electricity demands are at a different time. So there's a lot of sense in this interconnect, because when it's windy in Denmark and there's lots of demand in the UK, they can send their wind power to the UK and vice versa. And so I didn't know this thing was even being built. And it's gone live literally in the last couple of weeks. And it's significant, I think it's like at least a gigawatt of interconnect. So a serious DC power interconnect between the two countries. And that just made my day, because I knew we've been rolling back on various sorts of environmental policies, but this has just gone live and I didn't even know they were building it.
Gael 38:09
And that's really, that's actually, that's so fun because I stumbled on a map of all these new connects being built across Europe, both for sustainability reasons, also for security reasons, after the Russian aggression against Ukraine and all the energy issue that it raised. And that was mind-blowing that there is already a lot of being built and even more being planned. And as you say, that's just perfect. It's sharing energy and low-carbon energy as much as you can. So I really love it. I think I will try to find the map and put it in. That'd be great. Thanks for sharing it. That made my day as well. Super cool. So it was great. I will put all the references in the show notes as usual. And what is also super cool, is that there is like a very good deal of chance that we will meet in London in September. So thanks for joining. Talk to you soon. Keep up this amazing work with a very open source technology, carbon standard but sustainable pilot idea. I love it. So let's stay in touch. Thanks a lot for joining and have a very nice and sustainable day.
Katie 39:17
Thank you.
Oliver 39:18
Thanks for having us.
❤️ 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.
47 epizódok
Tất cả các tập
×Üdvözlünk a Player FM-nél!
A Player FM lejátszó az internetet böngészi a kiváló minőségű podcastok után, hogy ön élvezhesse azokat. Ez a legjobb podcast-alkalmazás, Androidon, iPhone-on és a weben is működik. Jelentkezzen be az feliratkozások szinkronizálásához az eszközök között.