Links, quotes & articles (Week 26, 2024)

I was in Australia for the last week and a bit with family, so didn't get around to reading or listening to too many things. This week's update is very much me catching up on some really great material!

A deeply insightful episode with Adrian Cockcroft, one of the leaders in the Cloud sustainability space and a driver behind the Green Software Foundation's Real-time Cloud project. Adrian is leading the charge to develop a standardised system for Cloud providers to report on the carbon intensity of their regions, both in real-time and also in their reporting.

Sticking with data centers and the Cloud, this episode of Catalyst takes a look the energy demand of data centers, how we might meet that demand, how it could impact the broader energy sector, and the role that data centers could play in dynamic grids of the future. The episode spends a bit of time talking about AI (of course), one of the key drivers behind the increased demands on the grid from data centers.

So the one thing that always comes up in this context is, well, data centers should just go off grid. Why not just have all of your own power, make your own power, you don't need the utility. But the problem with that is, there's no such thing truly as going off grid. Because what you mean by that, at least if we're talking about this decade, is I'm going to build a natural gas power plant, which, surprise, is actually connected to a grid. So you're just saying, I'm going to go from connecting to the power grid to connecting to the gas grid, which, and of itself also has constraints on it. And you have issues with the firmness of the gas supply. So there is no magic button to say, well, I just got to completely get rid of the utility, certainly not at the scale we're talking about. You're not going to put solar and storage to supply a 500 megawatt to gigawatt scale data center. I mean, it's just not feasible. Nuclear comes up a lot, but again, we're trying to be realists about what we're going to do this decade. I'm very bullish on nuclear. You and I have talked about nuclear a lot, very excited about what's happening in that industry. But the focus that these companies have right now is getting data centers online in the next three, four years. And nuclear is not going to solve that problem. So taking the data center completely off-grid is not the solution.

I think probably the perception today inside of those companies is we need power now and we need a path to get lots of it. And we also want to have line of sight to where that zero-carbon energy is going to come from at some point. And so I think that's sort of the trade-off that's probably happening at this point, is there is such a land grab or a power grab to get as much capacity as possible to plug in as many GPUs As possible. Because this is really a game about scale and building the biggest machines. And there's a lot of discussion right now about efficiency and what if Nvidia comes out with a more efficient chip, which they have and they continue to innovate and come out with more Efficient chips. But the example that I used with utilities is that if, let's say Meta, for instance, finds a piece of land or a point on the grid where they can pull three gigawatts of power in one location, And then the next day, Nvidia comes out with a chip that's twice as efficient. Is Meta going to build a 1 .5 gigawatt data center? No. They're going to build a three gigawatt data center that is twice as powerful as the one they thought they were going to build. That's what this AI game is all about, is building the most powerful models that you can. And so that's why efficiency isn't really going to save us in this particular context.

A good listen if for anyone interested in where digital sustainability might overlap with other more established parts of IT or business operations. Anne and Navveen touch on a few topics, but the one that interested me the most was their discussion around using operational cost as a proxy when estimating carbon emissions. Anne makes a very good point about the freebies a lot of Cloud providers dish out, and how that is likely to make cost a not-so-useful metric. She also makes an interesting point about when and where efficient coding practices are useful, and where time might be better spent on other endeavours.

I think, at the moment are very aligned with FinOps and cutting your costs. And some things are not aligned yet, but are almost certainly going to become aligned in the future. So, for example, you talked about operational efficiency and automation, which is interesting. Obviously, operational efficiency, if you use fewer machines and you use less electricity, your bill goes down. So that's all good. So in that respect, your FinOps and GreenOps are really well aligned. Fewer machines, less stuff, less carbon goes into the atmosphere. That's all fantastic you get. And in that respect, I would say that your hosting bill is a really good proxy metric for your carbon emission. But of course, it's almost stupid, it's so obvious to say it, but it can't just use your cloud bill totally blindly as a guide. I used to do startups in my youth, quite often Azure or AWS would give you loads of free credits. Then of course, but that doesn't mean that's carbon free. There are times when you just need to use your head, don't you?

I always say there are loads of different types of developers in the world, but two types of back-end developers. So front-end developers is just almost a separate thing. But back-end developers, you've got people who are just working in an enterprise and the code that they're producing is not going to ever be deployed. It's the code that they are writing. It's never going to be run by billions of people in their own data centers. And then you've got people who are writing platforms. And the whole purpose of the platform is to try and get some billions of people or it's millions of people to write this code, to run this code. Those people, they absolutely need to write efficient code. I think everybody needs to get used to getting out their performance profiler and just making sure there are no egregious performance problems with their code. Because performance problems are your code's slower and you're burning a load of carbon and it's total waste. So all, again, very aligned with the business. You want your systems to run fast, your customers want your systems to run fast. So having a decently performance system is good. But beyond that, you probably don't want to be writing code which is massively efficient because that takes a long time. But you do want the platforms you're running on to make that investment.

In this episode of Syntax, Wes and Scott answered a listener question about website sustainability. It was great to hear Green Web Foundation and Website Carbon mentioned by them - though they'd not heard about either 😅. I ended up with a smile on my face as I listened to them getting nerd sniped by results they were seeing for their websites when checking on Website Carbon. Hopefully this discussion on such a popular show will bring some more attention to digital sustainability.