Keynote: From Orbit to the Grid: Automating a Greener... Faseela K, C. Holmes, M. Reichenbach (ASL)

This talk explores the intersection of cloud native technologies and sustainability, emphasizing the role of Kubernetes and cloud architectures in enhancing resource efficiency and addressing environmental challenges. The speaker, Fasiha, sets the stage by discussing the three layers of cloud sustainability: sustainability of the cloud, in the cloud, and through the cloud. Chris Holmes from Planet Labs shares insights on how their cloud native systems enable real-time Earth observation and environmental monitoring, including tracking illegal deforestation in the Amazon using satellite imagery and AI. Michael Reifenbach from 1,5 discusses the transformation of the energy grid through cloud native solutions, highlighting the operation of virtual power plants that decentralize energy production and consumption. The session concludes with a call for the tech community to actively contribute to sustainability initiatives.

Good morning, everyone. And welcome to this interactive discussion on cloud native sustainability. So, in Europe, sustainability is no longer a future ambition, which means Kubernetes and the cloud native systems we are designing are part of our planet's energy story now. So, when we design cloud architectures, it's very important to think in three layers. So, are we designing cloud native systems that are maximizing resource utilization? Are we

making use of these systems to solve environmental challenges? And are these systems powered by renewable energy? Before we get into the details, I'm Fasiha, [snorts] a cloud native developer at Ericsson, a member of the CNCF Technical Oversight Committee, a CNCF Ambassador, and one of your KubeCon CloudNativeCon co-chairs. So, today I'm going to drive this that basically spans three layers of sustainability, sustainability of the cloud, in the

cloud, and through the cloud. So, let's get started with sustainability through the cloud first. Please welcome Chris Holmes, VP at Planet Labs. Chris is from Amsterdam, and he's going to share with us Planet Labs story of cloud native adoption for planet Earth. So, welcome, Chris. >> [applause] [music] >> So, Chris, when we look at the planet from the orbit, what does cloud native make possible today which

wasn't even possible before? Thanks, Fasiha. The answer is visibility at a scale that was unimaginable a few years ago. At Planet, we operate the world's largest fleet of Earth observation satellites, imaging the entire Earth's landmass every single day with over 200 on orbit. The satellites generate more than 7 petabytes of data annually, all processed through Kubernetes to deliver real-time planetary insights that make environmental change visible and

actionable. Our mission is to use space to help life on Earth. Let me show you what that looks like in practice. In Brazil, we're working with the federal police to combat illegal deforestation in the Amazon. Using our daily satellite imagery and AI-powered change detection, authorities have collected nearly 3 billion euros in fines for illegal logging and mining. Over 100,000 users across 500 institutions now monitor 8.6 million

square kilometers in near real time, identifying illegal operations and not just assessing the impact, but giving authorities the timely insight to stop the deforestation before it finishes. We've also built the world's first global-scale forest monitoring system at 3-m resolution. Updated quarterly, it provides precise measurements of carbon stocks, plus canopy height and cover for every hectare of forest worldwide. This is critical infrastructure for voluntary carbon markets and

regulatory compliance like the EU Deforestation Regulation, enabling governments and companies to actually measure and manage their carbon impact. And a year and a half ago, we launched Tanager-1, our hyperspectral satellite built with Carbon Mapper and NASA JPL. It detects and quantifies methane and CO2 emissions from individual facilities with unprecedented precision. In its first year, it detected over 5,000 methane plumes across nearly 3,000 sources globally, giving operators

the data they need to find and fix leaks fast. So, this is really amazing, Chris. So, how does cloud native fit in your story? Well, with so much data coming down every day, it's a massive challenge to process it all and make it available in near real time to our users. The only answer we found that to scale is Kubernetes and the CNCF ecosystem. We've built a

fully cloud native platform running on Kubernetes with Prometheus and Grafana for observability across our entire pipeline, Argo and Keda for event-driven workflow orchestration that scales dynamically with satellite downlinks, Helm and Crossplane for reproducible infrastructure-as-code management from dev all the way to petabyte-scale production, and Backstage powering our internal developer platform for hundreds of engineers. This infrastructure enables us to process 30 terabytes daily through ML and AI pipelines,

deliver near real-time insights to thousands of customers through our Planet Insights platform of web UIs and APIs, and scale seamlessly as we launch new constellations. This isn't just supporting our business, it's infrastructure for planetary-scale From predicting protecting rainforest to detecting methane leaks, cloud native architecture enable us to deliver near real-time actionable data that governments and companies need to make better decisions for our planet. Thank you. Thank

you so much for Chris for sharing this amazing story, and uh so, this story actually highlights >> So, this story actually highlights one thing, like we are no longer observing the planet after the fact, which is definitely great, but that brings the next question. What what how it looks when such massive compute is being run, how do we operate responsibly at scale? So, that that's the next

biggest question that comes, and that is exactly what the CNCF Sustainability Ecosystem is tackling. So, let's come back down from the orbit and now discuss about sustainability in the cloud. So, this map here shows the global footprint of data centers. So, for many organizations, moving from on-premises to cloud was the first step towards, you know, efficiency through better resource utilization and shared infrastructure. But then, the next

question is, how do we measure the impact? Organizations are now measuring scope one, two, and three emissions. Governments are including sustainability goals in public bids, which means carbon reporting is slowly becoming mandatory, and all major cloud providers are now, you know, offering carbon footprint tools as well. Within CNCF, this work becomes practical. It starts with measure. So, we make energy and carbon as visible as cost and

latency, so that teams can see impact in their day-to-day operations. Then comes observe. Signals become insights, where waste is happening, what is improving, and where we should focus next. From there, based on the insights, we decide we get to decide scaling, placement, and scheduling policies. And then, we act. We execute, right-size workloads, reduce idle capacity, and make efficiency the default behavior. And finally, we close the loop,

report, repeat, and improve. Small improvements add up across the entire system, and that's how sustainability becomes an operational property. And this is exactly what CNCF TAG Operational Resilience does. As a CNCF TOC member, I'm so proud to see how the TAG helped the community run some of these guidelines into repeatable practices, and it's also exciting to see other projects like Sylva under the Linux Foundation Europe making

use of these guidelines. For example, using Kepler in Telco CNF deployments. So, how do we get involved in the sustainability mission? So, beyond CNCF TAG Operational Resilience and across the Linux Foundation ecosystem, we have got initiatives that powerful stack GreenOps. So, if you want to be part of any of these, scan the QR codes, join the discussions, and contribute your expertise. So, so far, we have seen

how cloud native choices shape sustainability within our platforms. So, let's go and check how these systems are also enabling sustainability within real energy infrastructure. So, now, let's move from clusters to homes. Please welcome Michael Reifenbach, senior platform engineer at 1,5, and he is going to show us how virtual power plants are run using cloud native technologies. So, welcome, Michael. >> [music] >> So Michael, what happens when

actually this energy systems move from static infrastructure to cloud-native software? Thank you, Fercilla. So, we've all been through a revolution. We took monolithic applications and broke them up into microservices. We went from one server to 1,000 containers. It was hard. It was messy. But, we've built the tools to make it work. And right now, the exact same revolution is happening to the energy grid. For a century,

massive power plants have been powering the entire country. Burning coal, burning gas. Simple, centralized, and destroying our But, what if every home could become its own power plant? What if we could live on wind and sunlight forever, for free? That's exactly what's happening right now. Solar panels on roofs, batteries in basements, EV chargers in garages. Every home becoming a tiny power plant. But, here's the problem. Just

like a million microservices without Kubernetes is chaos, a million power plants without orchestration is 1,5 we're solving this exact problem with the same tools and technologies whose creators and maintainers are sitting right here in the room with us. With cloud-native technology. We call it Heartbeat AI, one of Europe's largest virtual power plants. With over 50,000 connected systems and 4.8 million optimization decisions every single day, 1,5 is

driving the new energy revolution, decentralizing the power grid. And in the last years, we've already saved over 2 million tons of CO2. That's 90 million trees, a forest stretching all the way from Amsterdam down to Frankfurt. That's really amazing to know. So, >> So, how does that actually look like for someone who is completely new to virtual power plants? Yeah, great question, Fercilla. Come, join me, everyone,

on this brief journey into the future of new energy and the architecture behind Heartbeat AI. Everyone, take out your phones and scan this QR code. Join me in our very own KubeCon virtual power plant. And while you're connecting and stress testing the Wi-Fi, let me show you the architecture behind the power grid of the future. This is a home. There's solar panels, a battery, a heat pump,

an EV charger. It stores energy. It produces energy. It consumes energy. It is a tiny power And it's connected to the cloud through an IoT gateway. But, when you go from 12 power plants behind fences to 12 million power plants in people's basements, your attack surface explodes. That's why use Falco, detecting threats and intrusions across our entire fleet. Well, but how do you observe 12 million Well,

the same way you would observe 12 million containers, using open telemetry. OTel is the very reason this decentralization is possible at all. It's free, it's open, and it scales to infinity. And all of this flows through Kubernetes, the orchestrator that connects every home to the cloud, to Heartbeat AI, making consumption follow production. And on our dev stack, we're using OTel, Prometheus, and Grafana for observability, Helm for

deployments, and Backstage as our developer portal. This is really amazing to know about the whole architecture. So, maybe you could show us this idea live, probably how cloud-native architecture is really orchestrating real energy grid. Yeah, of course, Fercilla. >> Let's try. Let's get to the fun part here. So, for the next minute, you're all part of our KubeCon virtual power plant. Your phone is a node, and

you are simulating Is everyone connected? Good. Then, let's have a look at the day in the grid of the future. Today, we need to use a little bit of imagination. Let's just imagine it's a sunny day in Amsterdam. The wind is blowing hard, but there is so much clean energy that the grid can't handle it. In the old world, we would shut down wind turbines, disconnect solar

parks, waste that free and clean energy, just to keep the grid stable. we saw this coming. You saw this coming. Our batteries have the space, and we were waiting exactly for this moment. On your phones, tap charge. Store that clean energy. Relieve the grid of its stress. And as you do this, you can see that Heartbeat AI is sending out thousands of individual signals to the nodes

in our virtual power plant, filling our battery with clean energy. But now, the day is coming to an end. The wind is dying down. The sun is setting. People are coming home. Demand is surging. In the world of old energy, we would burn fossil fuels to meet demand, blowing CO2 atmosphere, polluting our But, our batteries are full. We have Our cars are charged. We have the energy.

Let's give it back. On your phones, tap discharge. Give that energy back to the grid. Help me stabilize it again. What you just did is what Heartbeat AI does 4.8 million times a day, and we're just getting started, orchestrating the power grid of the future, always providing the cleanest and cheapest That is the power of cloud-native technology used for good. The same tools we use to scale

our applications, we use to scale clean energy. Living on wind and sunlight forever, for free. Thank you. Thank you so much, Michael. >> with that, we have seen sustainability over the three layers of cloud. Sustainability of the cloud, in the cloud, and through the cloud. And together, let's make sure sustainability is a default property of cloud-native systems. And with AI being everywhere and being part of every

stack, let's not forget to ask the question, how green is your prompt? And before we wrap up, we would like to thank everyone who helped us shape this keynote, CNCF contributors and maintainers, the events team, and our friends and colleagues who supported with the visuals and reviews. So, thank you once again, and enjoy the rest of KubeCon. Thank you. Yeah.