 Hello. Hope you had a good break. This is going to be the most important talk of your life, so please listen. My name is Jakko Alajöki. I'm from a company called Evermade. I'm founder of the company and also CTO there. And today we are going to talk about environmentally friendly WordPress development. Let's start by checking few numbers. It has been estimated that in six years data center power consumption is 20% of all electricity produced on this planet. 20%. Today the yearly number is over 400 terawatt hours per each year. For me, this number doesn't say that much. But if we think about what it takes to produce that, it would require over 50 nuclear power plants. One we have at all about the one here in Finland, it's 100 megawatts. Or it would take 200 coal power plants. For example the one here in Helsinki, Salamisari, you might know it, 200 of those. Or the biggest windmill in Finland for megawatts, 12,000 of those. To power all the data sensors in the world. And in a way, that's our fault. We have done the services taking all this energy. And when we talk about climate change and environment, we talk about flying. We talk about meat, food production. We talk about traffic. All of these very important things. I think we are missing one here. And I found one source saying that the data sensors actually emit more carbon. The carbon emission is bigger than flying. So it's quite a big thing. Of course there's a lot of discussion on how the energy is produced for the data sensors. But there's not that much discussion on how to reduce the amount of energy required. And especially like what I can do to reduce the amount of energy. What services I create. I have one case example. We have been working on one event site. And last fall, few days before the event, we realized that the site was quite slow. And we started to investigate the issue. And we found out that it was the varnish, the cache. It wasn't caching properly. And that's why the site was slow. So we had pretty much two options. We could fix the issue. Or we could scale up. Fixing the issue would take roughly four hours of work. That's worth of 400 euros roughly. Scaling would take a few minutes. Maybe additional server costs 30 euros. So you can guess what we did. From four cores to 24 cores. Because it was so cheap. It was so easy. A few minutes from four to 24. And I think that right now we have unlimited resources in our hands. If we need to scale, we just scale. More CPUs, more memory. There's no financial penalty for scaling. There's some, but there's more. I started to think about what I can do for this. Is there something like I'm just a single developer. Can I do something for this issue? So I decided to build my own data center. And run some experiments. Here it is. I configured Raspberry Pi server. To be as a host of workers site. And then I ordered this beautiful power consumption meter from Amazon. And I created a set of experiments to see how various scenarios affect the power consumption of the server. I have here four of them. First one is pretty obvious. Caching. Enable cache. That was super cache. It's 70% less energy consumed. I did this by requesting the front page with the pass benchmark. I guess this was 5000 times or something like that. Enable cache is pretty obvious. But to be honest, I have done maybe hundreds of sites in my life. If the site is very small, I know it's a low traffic site. It's on a fast server. I have deployed many sites without cache. Even though it's obvious that the user is always cache. Second example, unused plugins. Please never do that. But I went to WordPress plugin directory. Just install 20 from the most popular list. Like 20 most relevant plugins. And suddenly and enabled all of them. And suddenly double as much energy consumed. I know also that this happens all the time. We forget to remove plugins we don't use. And all those consumes energy. The next one. Nested query. Imagine an example where you have a list of say apartments. And on each apartment you have a contact person. So what you could do is just like load the list of apartments. Have a loop and inside loop load that's related contact person. So you do multiple queries. It's quite easy to code. But what if you write a single join database query. It's a little bit takes like a little bit more time. A little bit more complex. But significant change on the power consumption. The last example I have is the very classic PHP optimization. If you Google for PHP optimization you will always find this. Do not use regular expression if you don't need to. Use string replacements. I did an experiment with JSON loading. With like doing some replacements for multiple items. I was surprised how fast that actually affected it. But yeah 20% I can take it. So based on these experiments I can personally say that. From now on all the services I create. They will consume at least like save at least 20% of energy. I can say it based on these experiences. That itself is not significant. But I'm going to create multiple sites hopefully in my life. And if you all do that as well. That all adds up. And now imagine if I do that. If everyone in our company do that. If like all developers here globally do that. 20% off. For example 40 coal power plants less. I'm happy to take that. Of course it's obvious that I'm standing here and saying that please consume less CPU. That saves energy. But I think it is the mindset that needs to change. We optimize our sites for various SEO. We want very Google rankings. We optimize our sites for our users. People want fast sites they don't want to wait. Sometimes even though the scaling is easy. We optimize follower costs. But what I would like us to do is to also optimize for polar bears. That was my message today. I'm going to write a blog post to evamates blog about this. Like all the calculations and details information about how I did the experiments. So if you want to see that you can follow me on Twitter. You can follow evamate on Twitter. Or you can just uninstall Twitter. Go out into nature as long as this lasts. Thank you.