 Welcome back everyone to theCUBE's coverage here in Las Vegas. We are on location for AWS re-inventor. 11th year, I'm John Furrier, I'm your host with Dave Vellante. We're here extracting the signals. We've got a great guest, CUBE alumni, former AWS, former, a lot of things, Adrian Cockcroft took some time to step away from AWS, traveled the world, give a bunch of talks, do some digging in, tech advisor, OrientX, and also CUBE collective friend of theCUBE coming on, dropping knowledge, always having a master course. Adrian, great to see you. How are things? Got the great jacket, Thursday jacket? It's the Thursday re-invent. I did have a nice shirt on earlier in the week, but this is a party tonight, and this goes in. 2015, all the AWS heroes got these golden jackets, so bring it out one night. Last year, when we were last year on theCUBE, I asked what's your plans for the year, you don't go to a bunch of events, give some talks, just hang out, and that's the version of being retired in tech, I guess, and let's do whatever you want. That's a cool thing to do. Couple things, what did you learn this year? I said, this show here, what's your impression? And then we'll get into some of the things you're working on now with sustainability, you're digging in hard, but let's first find out, what did you do this year? What's the big discovery? And then, what do you think about the show? Yeah, about a year and a half ago, I basically retired from AWS, and I'm independent. This little, Orion X is all consulting vehicle to just do consulting, work and advisory work and analyst sort of things. So that's been, then just trying to manage my workload so I didn't make myself too busy, right? So, and you know, some people end up getting way busy with when they, quote, retire. I saw about half the time traveling around, the other half the time, I'm helping out a number of different companies with interesting problems, people that I generally knew before I left AWS. So that's been fun. I've also joined the Green Software Foundation, so we can get into that little bit and talk a bit more about sustainability in that scenario. The last job I had at AWS was at Amazon actually, because there's AWS to do open source and a few other things, but then I switched to Amazon for the last couple of years and I was in the sustainability team. Basically helping sort out the AWS sustainability messaging. You notice that reinvent there's a sustainability track. We put that together the first time and they've continued around to the third year. We've had a sustainability track and yeah, they're still, I recognize a bunch of the slides they use. So that was what we did. But fundamentally on the sustainability side, AWS has been slow to release information and it's still not transparent enough and we can talk a bit more about that. But I've been given that I'm now sort of outside AWS. I've spent more time working with some of the other players figuring out what's going on, what's available, and just getting a much better handle on what's going on in cloud sustainability overall. So back in 2008, 2019 timeframe, there was this push toward green IT, I don't know if you recall. And it didn't really stick. Why does this matter so much today? Why was it all of a sudden getting so much traction? I think, well, one of the things is that the cost of renewable energy over the last decade got started to get cheaper than just buying it, right? Like the cheapest energy you can buy now is solar and wind. And maybe a bit of battery thrown in. So the economics of it just drove all the large energy users to just start to being renewable, right? And some of the power providers for the city and state level are gradually getting more renewable energy because they've got all these old power plants and to time out. AWS goes and the other cloud providers have been buying renewable energy to the point where they are some of the biggest buyers in the world what's called a power purchase agreement. So there's a wind farm or a solar farm that's dedicated to AWS. They paid for it to exist and they take all the energy from it and they count that against their total usage of energy. And all the cloud providers are doing that. AWS is doing, they don't separate out AWS from Amazon in the sustainability report. It's all one big thing, right? So, but there's, you know, tens of gigawatts of generating capacity that is purely there because AWS paid for it to be there and that they have a bigger investment overall than Microsoft and Google. But everybody is in that sort of tens of gigawatts kind of level. But I'm inferring you think it makes sense for them to split out AWS. It's cheaper, right? That's the cheapest way to buy energy, right? And also you get to decarbonize and everyone's got commitments to decarbonize and you know, the general industry commitment is sort of net zero, 2050, right? Most of the cloud providers have a 2030 goal, more aggressive and AWS says though, I think they still under 2025 is when they say they'll get to 100%. Well, what's the, give us the analyst framework. Like, how should we think about this topic? You know, what metrics matter? Can you help us kind of paint a picture of what we should be looking at here? I think, well, the way it matters a bit, on the one side economically it's cheaper, right? But on the other side, there's regulation coming in. We've got regulation now on the books in Europe and in California, when California does, if you want to do business in Europe or California, it's pretty hard to not be involved in those two parts of the world. You're going to have to disclose carbon and conform to various reporting requirements. And this is, so this is sort of coming down board level through financial, CFO, board level. Like, we have audit requirements that mean you have to figure out what your carbon footprint is to just do business if you're above mineral size, right? And then, so that's been happening for a while, but then bottom up, there's like employee enthusiasm. Like, there are a lot of people where when your kids come home from school and say, hey, I hear the world's going to end. And what are you doing about that? He's like, you don't want to be working for, oh, I've been selling oil for the last, whatever. That's not the right answer. So we're getting, and it's amazing how strong that effect is. So there's enormous amounts of employee enthusiasm to tap into this. And then there's that sort of bottom up, top down. And then the end-to-end is the supply chain. Like, you need to figure out the carbon in everything you buy and in everything you sell. So it's an end-to-end going right through the industry. So it's becoming a pervasive need to do that. And more and more people are doing it voluntarily in advance of the regulations, which are just pushing it. So they're going to be getting the muscle down. So you've got to do it anyway. And then people have decarbonizing commitments. So you've got to have a plan to meet this commitment. And that's a country level, and it goes down to businesses and all the way down. Yeah, Dave and I always talk about theCUBE and how we're always trying to get those puzzle pieces when we're putting the puzzles together. Cube and tech landscape. And every time we get these little puzzle pieces, it's like, okay, the picture's clear. I'm just so blown away by how complicated sustainability is. I mean, you're into it. I'm looking at it like, okay, what do the corners look like in this puzzle? So how do you put this puzzle together? Do you have a systematic approach? How do you look at it? Do you just dive in? Do you zoom out? I mean, what's your approach when you, because there's a lot of things that just, some places know data. Is the data truthful? Are they hiding the ball? Are people really volunteering? What data should be tracked? Yeah, what is your approach? What's your strategy? Yeah, so I spent the last year, the last few years just thinking, okay, well, I'll go figure this out. And it gets more and more complicated the more I know about it. It's a rabbit hole for sure. So it is a complete rabbit hole and you see people saying, oh, I see, no, that's definitely wrong. No, there's all these nuances and things. So I'm still trying to assemble something that makes sense. But so what I've been doing to try and gather this thing together is I joined the Green Software Foundation, which is one of the Linux Foundation, like CNCF, the Linux Foundation is all these sub foundations they run. Which was, and I have a project called Real Time Cloud. So you can, it's all done in the public, you can go to GitHub and find the Real Time Cloud project on Green Software Foundation and see what we're doing. And I've got to gather together a bunch of people and these are people that individually know a huge amount about carbon and the supply and there's people there from Google and Microsoft and I know some of the AWS people so I can kind of give that perspective. My chance to gather to get what data is available, it's all different. What models is it based on? They're all different models. These numbers, it's all apples and oranges everywhere. You can't compare any of the numbers, okay. And if you get... That's a cloud market share to me, Dave. Yeah, yeah, yeah, yeah. Time's 10. You get two people in a room and you'll find you've got 10 opinions on how to do it as well because if you're trying to do it, why are you doing this? That you have different models and different things depending on what you're trying to do. So, and there's this whole idea about, I wrote a blog post recently about consequentialism. It's like, do you want to just account for the carbon you're producing? It's one thing. You do as one methodology. If you want to have a change the world, you have to use a completely different methodology. Is that what? This is crazy stuff, all right, fine, whatever. But you account for the numbers because anyway, I literally wrote a blog post as I was trying to understand this and I checked it with some of the people that really understand this and kept going around and around and, oh no, I got there. So that was a recent blog post. So I'm trying to sort of document that. One of the things we do in the Green Software Foundation is they built a map. There's a mirror board and we just put everything we know on that board. Okay, this thing, this piece of software here gets its data from there, which gets its data from there. So I've got end-to-end flow of all of the carbon, power, data center, PUE, all of that information, power usage efficiency. All of that data, where do you get it from? Where does it flow to? Where does it get to? And we've just been mapping it all out. So we're kind of in that stage of just laying everything out. So you're laying the puzzle out now. So we've got the puzzle laid out and if there's projects, what is this, there's a bunch of projects with weird names. Like CNCF has a project called Kepler, which is the Kubernetes-efficient power exporter or something. But basically it will tell you the power consumption of your workload on Kubernetes, which is awesome. Thank you for building that and save me from like arguing why somebody should build it, right? And that, when you run it in the cloud, it can't measure the power of the cloud instances for security reasons because it's all blocked for multi-tenant, right? Unless you're on a bare metal instance, it can't read the power meter data. And so like, ah crap. So they have a little AI model in the bottom of it that's trained on machines you can get to and then they just guess for the cloud. So there's all this data, when you get to the cloud, the models, there's so much guessing going on because you're not getting real data out of the cloud. It's a very opaque. So nobody's really got this figured out, right? But is there a one-eyed person in the land of the blind? Does anybody stand out that's doing a good job? In different ways, Google and Microsoft are doing much more transparency. I think if you actually, I mean, if we had all the data about everybody, I think you'd find that AWS is actually doing, in many ways, a better job of actually buying more power in difficult to do areas. They're actually, they're all doing a good job of actually decarbonizing their energy supply. So that's not the question, right? And there's not hiding the ball and there's no smoking. There's an economic imperative for this. It's economic, but they are actually all doing really well. And when you move to cloud, you will mostly be running on low, low carbon energy. And so that's good. But then the transparency of like, I need to take my workload, I need to allocate all the bits of the workload to customers and I need to allocate and attribute all the carbon data. And that's what Kepler does. So you've got, you've done to pod levels, it'll figure out how much energy a pod is using out of a node, out of a machine and all that stuff. So you have to do that. And then you know what a workload's doing. And then, okay, now I know my energy. You know, well, what's the carbon intensity and the energy and then everything goes down. Right? You can go so many different places to get numbers and there's so many different ways to do it. And this is where it's getting complicated. So what we're doing in this real time cloud project is just trying to say, well, which ways make sense? And can we get, can we define what those are? And can we get at least a spec for what we think all the cloud providers we'd like to do? And then we can get, there's a bunch of big customers in this green software foundation. We get the customers to go beat up their cloud providers to go, can you just conform to this spec thing so that we can all have the same data? And then you'll be able to build like real time tooling. Things like, like right now you get the data, the carbon data is like, you get it a month or two later. Yeah, I do a benchmark. Okay, I got to wait two months to find out what the carbon number is. That's not real time. Yeah, when I was talking about this in the keynote, he was crossed as a proxy. But cost is directionally correct, but not proportionally correct. Like the carbon of a dollar's worth of storage or a dollar's worth of CPU have radically different amounts of carbon in them. Yeah, right. So, okay, so if you had a magic wand to wave over the industry, where would you start? Just having everybody standardize on their reporting or what advice would you give to the industry generally and AWS specifically? While I was still at AWS, we were looking at joining the Green Software Foundation. I even had a blind item and a budget to go do it and somehow I left and it didn't happen. So, it's not a big deal. Don't need this anymore. And it's gone. Just join the foundation, please. So that would be a simple thing for them to do. It could do any time. It's not expensive. Just so then have somebody turn up to the meetings to give their viewpoint on what's hard for them to do, what's easy, what data can they do. Just start aligning the data sources on what's going on. So that's a simple thing. And then over the next few months, we're starting to define effectively a schema or like a, you know, which metrics, what are the units of the metric? What are we going to call it? What does this thing mean? And we're going to figure out what that looks like and plug in the existing data that is basically Google or Microsoft mostly are providing. And then see if we can get AWS to do more of that data. But at least we'll have a spec for what makes sense. Then you'll be able to tools on it. All the cost optimization vendors are using the same questionable data that's made up. I know it's made up data, right? They're all using the same data sources and that you get a number. I know that number's wrong, right? But say all models are wrong. Some models are useful. It's like Phenops. You got to have green ops. This is all Phenops. The Phenops people are all trying to do like carbon on the side. And it's when you get to the cloud. It has apples and oranges right there too. You guys might as well just to green ops right of the gate. Yeah. But you can't compare the cloud providers and they're ignoring a whole bunch of data. And it's just like that. All right, so how do you feel about this? Do you feel like you can go forward? Do you think that you're going to come out of the rabbit hole and just say blockchain? You know, like, what happened to Adrian? He had his sanity with him and then he just like lost it. I mean, your question. How do you feel about it? Do you think you can get this done or? I've done a number of standards bodies in the past. I have a 40-year-long career. I did it with Son. I did it with a few different people. I know how standards work. You feel good then. And there's a pace that you run at and it's slow. Because it's like they used to say, if you want to go fast, go alone. If you want to go far, go together. Right, it's that African proverb thing that they use. And it's a place to stand. You have to gather everybody, get everyone comfortable and gradually inch them forward that everyone can do that. So this is going to take a while, but nobody else was trying to do it. And the PSA decided to stop. But you've got great results so far. You've got great results. The data you're getting is, it's like you just got to keep iterating on it. We're finding, actually, making some sense. I'm discovering a bunch of data sources that I didn't realize Google was providing another data source than Microsoft. Somebody from Microsoft said they didn't have some data and then I actually hunted her and somebody from Google told me where it was in Microsoft. In Microsoft. They have really detailed data on every region they have. They have PUE and the carbon-free energy sources and that data. There it is, yeah. We're pulling up all this data and that's really, you know, I'm just, you know, it doesn't take a huge amount of time because if I try to go too fast it'll get pushed back. Super exciting. Adrian, it's super exciting for you and we're excited to have you back on theCUBE and looking forward to doing more. Now that you've got the groove swing, let's get an update in Palo Alto, come into the studio, do some remotes. Anytime we want to unpack some stuff, let us know we're ready to go. And again, super important area. And again, it's about getting the muscle, right? It's like the muscle's got to be built, you know? And it's a system. So it's not just one muscle. It's multiple muscles. Somebody's got the stomach to take on a hard problem. So thanks for doing that. And they generally right now, the time when most of the sustainability data updates is in the summer because it takes about six months to do the sustainability analysis for a big company. They, AWS publishes in June, July time, so does Google. So you tend to get new sustainability stuff drops somewhere in July. So maybe next summer, once that data's dropped we can say, okay, because what that will be the 2023 data will be. And it's just, it's like years later. I will tell you, in a year's time what happened this year, it's like, we need something that we can. It sounds like an analyst report. What the? That's called balls and strikes for the game today, not last year. So anyway, I'm going to be blogging on it. You can go to the GitHub for Green Software Foundation. If this is interesting, you can join the GSF, turn up at the meetings. There's all kinds of standards and things they're working on. Azure and great leadership. They're taking on a big problem. Got congratulations at the Linux Foundation having a group to do it. Heard the cats, whatever they say. You got to get people together. Good work. That's the cue. We're bringing you more action. The energy here is high and we have a low carbon footprint here on the cue. But a lot of intellectual footprint. We've got all the best guests here on the engine. Kukroff, who's doing great work sustainability. We'll be tracking this hard. Back to the Palo Alto studio. We'll be coming back here for more day two action here at the cue.