devopsy.kinsta.cloud

Tag: sustainability

Daddy, Where Do Software Engineers Come From?
There’s one question every parent dreads hearing from their child: “Daddy, where do babies come from?”

We’ve all heard the responses over the years—storks, God, mommy’s belly, “ask your mother,” you name it. But beneath the awkwardness lies a fundamental question: how do we make more of us?

Oddly enough, the tech world is now grappling with its own version of this question: Where do software engineers come from?

And lately, the answers we’re getting aren’t much better than “the stork.”

The Vanishing Junior Developer

Over the last year or so, we’ve seen a troubling trend take root. Across LinkedIn, Reddit, public layoff announcements and even private chats, junior software developers—bright-eyed, bootcamp-graduated, degree-holding or not—are becoming a bit of an endangered species:
- Offers rescinded.
- Layoffs hitting only the least experienced.
- Job postings asking for “junior engineers” with 5 years of experience and Kubernetes mastery.
- Companies saying they have put a hold on hiring junior developers
Meanwhile, hiring managers say: “We’re only looking for top-tier, senior-level talent who can hit the ground running… preferably with AI superpowers.”

I get it. In today’s hyper-competitive, AI-infused economy, the temptation to lean on experienced developers—and increasingly, AI tools—is strong. Everyone wants 10x output yesterday. But this mindset has a blind spot.

The Software Talent Pipeline Is Drying Up

We all love to talk about pipelines in tech. CI/CD pipelines. Security pipelines. Data pipelines. But there’s another pipeline we rarely talk about: the human pipeline.

If you don’t hire junior devs today, you won’t have senior devs tomorrow.

There is no Hogwarts School of Senior Software Engineering. Developers don’t just emerge from a secret lab fully formed with a GitHub repo of 10,000 commits. They start small. They mess up staging. They crash prod. They learn.

And yes, some of that early-stage work—debugging, boilerplate, test writing—is now within AI’s grasp. But that’s no excuse to stop growing new talent. The danger is clear: we’re optimizing ourselves out of a future.

Are We AI-ing Ourselves Into Mediocrity?

Let’s play this out. If junior roles disappear, then over time, fewer devs learn the fundamentals. That means fewer senior engineers down the line. Will AI replace them all?

Maybe. But do we really want to bet the future of software—and by extension, the economy—on that?

Even AI needs human oversight. Someone has to prompt it well. Review the outputs. Understand the systems it’s building. And as AI gets better, the bar for good human oversight gets higher, not lower.

We risk a future where the few remaining engineers are glorified AI babysitters with just enough understanding to catch hallucinated nonsense—but not enough to innovate, secure, or scale. That’s not a future I want to live in.

So What Do We Do About It?

Glad you asked. Here are 4 steps we can start taking right now to make sure tomorrow’s developers exist—and thrive—in a world where AI plays a bigger role:

1. Mentorship-First Engineering Cultures

Don’t just assign junior devs Jira tickets. Pair them with mentors. Give them structured feedback. Celebrate learning, not just delivery. If your senior engineers don’t have time to mentor, you don’t have a team—you have a support group for burnout.

2. AI-Human Apprenticeships

Why not flip the script? Let junior devs learn by prompting. Have them use GitHub Copilot, CodeWhisperer, or open-source models as learning tools—not replacements. Teach them to critically evaluate AI output. Think of it as co-piloting their way into expertise.

3. Hire for Potential, Not Perfection

Companies love to say “we hire smart people and train them”—until it comes to engineering. Create entry-level pathways. Host internal bootcamps. Offer returnships. Today’s “junior” can be tomorrow’s technical leader if you give them the chance.

4. Establish an Industry Pledge

Let’s take a page from the DEI playbook and build a Tech Talent Sustainability Pledge. Imagine a coalition of DevOps orgs, vendors, and communities committing to hire and develop a minimum percentage of junior engineers. Call it ESG for talent.

Don’t Let the Stork Handle Your Hiring

At the end of the day, we have a choice.

We can tell ourselves comforting myths—that senior devs will always be around, that AI will save us, that someone else is training the next generation. Or we can act like grown-ups and take responsibility for nurturing the future.

Because if we keep cutting the bottom out of the software talent ladder, don’t be surprised when no one’s left to climb it.

Final Thought

Software is the backbone of modern civilization. But without the human minds to build, challenge, and evolve it—even with all the AI in the world—we’re just one stork story away from collapse.
July 28, 2025
How Far Are We From Truly Sustainable DevOps

Sustainable DevOps has become a top priority for environmentally conscious engineers and business leaders alike. Customers, regulators and investors now expect release pipelines to move fast and tread lightly on the planet. Yet a nagging question lingers: How close are we to DevOps tools and processes that genuinely avoid environmental harm?

In this article, I’ll examine the hidden carbon cost of everyday DevOps activities and ask whether the raft of tools, practices and policies being introduced are enough to bend the emissions curve towards sustainability. Along the way, I’ll do my best to cut through the haze of green-washed marketing to find the real answers on how far we are from truly sustainable DevOps.

The Environmental Footprint of Modern IT and DevOps

The broader information and communication technology (ICT) sector already emits roughly the same amount of greenhouse gas as commercial aviation and, on its current trajectory, could swallow 14% of the global carbon budget by 2040. Within that total, data centres are the fastest-growing contributor. Reports warn that global data-centre electricity demand could soar beyond 3,000 TWh by 2030, more than ten percent of worldwide consumption and comparable to the present-day electricity output of Germany.

The energy hunger, however, isn’t confined to physical servers. Software architecture dictates how often the hardware runs flat-out or sits idle wastefully. The Cast.AI’s 2025 Kubernetes Cost Benchmark Report, which analysed 2,100 production clusters across AWS, Azure and Google Cloud, found that 61% of requested CPU time never executed real user work. All that energy was consumed for no reason at all.

Generative AI compounds the problem. A study published in April 2025 calculated that training a single cutting-edge large-language model can emit as much CO₂ as flying a passenger jet more than a million kilometres.

The demand for software supported by all of this tech is likely to continue and even escalate. The pursuit of faster releases risks becoming a vicious cycle of escalating emissions.

Current Strategies and Best Practices for Sustainable DevOps

Despite those bleak numbers, tangible progress is being made. ‘Green DevOps’ principles are becoming popular, where carbon costs are treated as a production metric to be optimised alongside latency and cost. Case studies have shown that simple development hygiene principles, like aggressive dependency caching, shutting down idle agents, and running tests in parallel, can cut build-time energy on a project by 28% without slowing release cadences.

Cloud Infrastructure Powered by Renewables

Hyperscale providers are racing to green their grids. Microsoft’s 2025 Environmental Sustainability Report reveals that the company has contracted more than 34 GW of renewable power, an eighteen-fold increase since 2020, keeping it on track to match 100% of Azure’s electricity use with clean energy by the end of this year.

Water consumption is also a major concern for data centers, and Microsoft has pledged to be water positive (i.e., they produce more clean water than they consume) by 2030.

Amazon Web Services has also been active in developing partnerships and deals with sustainable energy producers to meet its Climate Pledge. Their stated goal is to be carbon neutral by 2040. They’re apparently on track to do so, as they reported having matched energy consumption with renewable sources seven years ahead of schedule.

Google leads in renewable procurement, matching 100% of its global electricity use with wind and solar since 2017 and now pushing for 24/7 carbon-free energy in its data centers by 2030.

However, these claims and pledges shouldn’t be taken at face value, as all the hyperscalers have been accused of greenwashing (see below).

Smarter Orchestration and Auto-Scaling

Inside the cluster, the conversation is shifting from cost optimisation to explicit carbon optimisation. Studies on energy-aware elastic scaling for microservices showed that feeding real-time grid-carbon data into Kubernetes’ Horizontal Pod Autoscaler reduced energy use by up to 23 percent without breaching latency objectives.

Observability is improving too: The Kepler project uses eBPF (extended Berkeley Packet Filter) to expose live power metrics for every container and pipes them into Prometheus dashboards, giving platform teams the feedback loop they need to tune idle thresholds and right-size nodes.

Leaner Pipelines and Change-Aware Testing

Build systems are an obvious lower-hanging fruit. Change-aware, or selective, testing runs only the test cases affected by a commit and has been shown to slash CI energy usage by up to 60%.

Practitioners can apply the same philosophy to infrastructure automation. Powering down idle CI agents outside office hours and caching container layers can reduce a pipeline’s annual electricity bill by almost a third.

Green Coding and Minimal Architectures

At the code layer, the debate has moved beyond language wars to architectural minimalism. Stripping unused features and favouring a modular monolith over a sprawl of micro-services reduces runtime energy by 20–35% before any hardware tuning.

Emerging runtimes such as WebAssembly System Interface (WASI) and GraalVM native images promise near-instant cold starts with memory footprints a fraction of equivalent container images, making them compelling building blocks for low-carbon services.

Carbon-Aware AI and ML

Machine-learning workloads remain energy gluttons, yet DevOps can still help. Simply scheduling model-training jobs during low-carbon grid windows can deliver a 30% emissions reduction on unchanged hardware. Teams can combine that scheduling with model-compression techniques, like pruning, quantisation and knowledge distillation, to avoid the spiral of throwing ever-bigger GPU clusters at marginal accuracy gains.

Automations can also be adapted to take carbon optimization into account. While automated cloud cost monitoring might be the norm, these same processes and technologies could be applied to automatically monitor and optimize energy consumption.

Measurement and Accountability

You can’t improve what you can’t observe and measure. The open-source Cloud Carbon Footprint project added multi-cloud billing and real-time grid data support in February 2025, enabling unified cost-and-carbon dashboards.

Meanwhile, the Software Carbon Intensity specification is evolving. The Green Software Foundation convened a workshop to extend SCI to AI workloads, with an ISO draft expected before year-end.

These initiatives mirror the rise of practices like CarbonOps and GreenOps are emerging disciplines that borrow accountability practices from FinOps and apply them to energy and emissions, ensuring every engineering squad sees and owns the carbon cost of its code.

Challenges on The Path to Truly Sustainable DevOps

Despite momentum, sustainable DevOps is still in its adolescence. Tooling is patchy: most cloud calculators omit embodied emissions (the carbon released during server manufacturing), so teams risk optimising only a slice of their footprint. There’s also a massive lack in standardisation: here can be as much as a 70% variance between vendors’ estimates for an identical workload, underscoring the need for standardised measurement.

Trade-offs also persist. Developers accustomed to abundant compute often resist refactoring for efficiency, prioritising delivery velocity. The average cluster utilisation actually fell year-on-year as organisations over-provisioned to protect against unpredictable AI spikes.

Regulatory pressure is only beginning. The United Kingdom will mandate carbon disclosures for large IT estates from April 2026, but metrics, verification methods and enforcement mechanisms are still taking shape.

In the interim, green-washing flourishes. A U.S. congressional hearing exposed how some ‘100% renewable’ claims rely on certificates that do not match hourly consumption, especially when GPU clusters ramp suddenly.

Examining Greenwashing

Despite the claims of big-name hyperscalers in pursuing sustainability, they’ve also received extensive accusations of greenwashing.

While Amazon has pledged to make AWS carbon neutral, AWS remains part of Amazon’s vast logistics empire, which means you can’t take this wing of the company’s emissions in isolation. Amazon is particularly infamous for greenwashing to obfuscate its actual carbon footprint.

Microsoft’s lofty goals of going beyond neutrality to positivity should also be called into question, as they continue to invest heavily in fossil fuels.

And while Google pursues 24/7 carbon-free energy in its data centers by 2030, between 2019 and 2024, its emissions increased by almost 50%, driven largely by AI and data center consumption. Furthermore, Google continues to profit from oil and fossil fuel companies and climate change denial through avenues like advertising revenue.

What Could Move the Needle Next?

Edge intelligence is finally living up to the hype. Moving analytics to on-premises gateways or 5G base stations avoids multi-hop network traffic, reduces latency, and allows waste heat from processors to be captured locally, turning a problem into a resource.

Clean microgrids are another promising advancement. Technical trials of small-modular nuclear, geothermal, and long-duration storage promise round-the-clock green power unconstrained by weather.

Finally, carbon-aware tooling is getting smarter. Research prototypes already adjust Kubernetes node pools, CI run-queues and ML-training windows against live grid-carbon signals. Once such feedback loops become mainstream, choosing a lower-carbon timeslot could be as routine as selecting a larger pod size.

Conclusion

So, how far are we from truly sustainable DevOps? The answer is a qualified closer than ever, but not yet there. In 2025, we’re already seeing double-digit energy cuts in CI pipelines, container clusters that scale with carbon-aware signals, and multi-gigawatt renewable power deals that begin to decouple cloud growth from fossil fuel demand. Yet the industry still lacks universal, trusted metrics and organisational incentives continue to reward speed over thrift.

The good news is that the tools, standards, and economic tailwinds are finally aligning. DevOps teams that embrace green practices not only shrink emissions; they also cut cloud bills, stay ahead of regulation and strengthen their employer brand. With disciplined measurement, carbon-aware automation and a willingness to prune excess complexity, digital innovators can ensure their next release updates the planet as well as the product.

June 27, 2025
MLOps for Green AI: Building Sustainable Machine Learning in the Cloud
The tech world is buzzing with two transformative trends — artificial intelligence (AI) and sustainability. As machine learning (ML) powers everything from predictive analytics to autonomous systems, its environmental footprint is growing. Cloud data centers powering ML workloads consume vast amounts of energy, contributing to significant carbon emissions. Enter MLOps for Green AI: A fusion of machine learning operations (MLOps) and sustainable DevOps practices that promise to make AI not just smarter, but greener.

With over 18 years in IT and a deep focus on DevOps, I have seen firsthand how cloud infrastructure can be a double-edged sword — unlocking innovation while straining resources. My recent work on a cloud sustainability software-as-a-service (SaaS) platform taught me that MLOps can be a game-changer for eco-friendly AI. Here is how we can harness it to build sustainable ML workflows, drawing from real-world lessons in the field.

The Hidden Cost of AI

Training a single large language model (LLM) can emit as much carbon as five cars over their lifetimes, according to a 2019 study from the University of Massachusetts. When multiplied by the thousands of models deployed daily, the scale becomes staggering. Cloud providers such as AWS and Azure offer immense compute power, but without optimization, ML workloads can waste energy, inflate costs and undermine sustainability goals.

Traditional DevOps excels at automating software delivery, but ML introduces new challenges: Data pipelines, model training and inference — all resource-intensive. MLOps bridges this gap, streamlining ML lifecycle management. When paired with a sustainability lens, it becomes a powerful tool to minimize waste and maximize efficiency.

MLOps Meets Sustainability: A Practical Approach

Recently, while working on a SaaS platform focused on cloud sustainability, the challenge of making ML workflows greener was addressed. Here is how MLOps principles, applied with tools like Kubernetes, Terraform and GitHub Actions, can drive sustainable AI:
1. Optimize Resource Usage With Intelligent Orchestration
  Kubernetes is a DevOps staple for container orchestration, but it is also a sustainability hero. By deploying ML models on Kubernetes clusters with auto-scaling, we can dynamically adjust compute resources to match demand. For example, I have used Horizontal Pod Autoscaling (HPA) to scale ML inference pods, cutting idle resource usage by up to 40%. Add tools like Kubernetes Event-Driven Autoscaling (KEDA) to trigger scaling based on ML workload events, and you have a lean, green pipeline.
2. Automate Infrastructure for Efficiency
  Terraform’s infrastructure-as-code (IaC) approach is not just for speed — it is for sustainability. In one project, I crafted Terraform modules to provision cloud resources (e.g., AWS EC2 and Azure AKS) with sustainability in mind, embedding policies to shut down unused instances and prioritize low-carbon regions. This reduced costs by 30% and emissions by a measurable amount — proof that IaC can align technical and environmental goals.
3. Streamline ML Pipelines With GitHub Actions
  Continuous integration and continuous delivery (CI/CD) is not just for code — it is for models too. Leveraging GitHub Actions, I automated ML training and deployment workflows, integrating carbon-aware scheduling (e.g., running jobs in off-peak, low-emission hours). This cut build times by 30% in a recent project, while also shrinking our carbon footprint. MLOps automation ensures models are deployed efficiently, not excessively.
4. Monitor and Minimize Carbon Impact
  Visibility is key. Tools such as Prometheus and Grafana, paired with cloud-native monitoring (e.g., AWS CloudWatch), enable us to track resource utilization and estimate carbon emissions. In my SaaS work, I built dashboards to monitor ML workloads, identifying over-provisioned clusters that waste energy. One tweak — right-sizing a Kubernetes node pool — saved 20% in compute costs and reduced our environmental impact.
5. Leverage Sustainable ML Techniques
  MLOps isn’t just about operations — it is about smarter ML. Techniques like model pruning (reducing model size) and federated learning (training locally to cut data transfer) lower energy use. Pruned models deploy 50% faster on Kubernetes, with a corresponding drop in cloud resource demands. Sustainable AI starts with sustainable models.
Case Study: Cloud Sustainability SaaS

In a recent project, a SaaS platform was designed to make cloud usage greener.

Goal: Help clients optimize their cloud footprint, including ML workloads.

Approach: I designed MLOps pipelines that deployed sustainable ML models — think carbon-aware training schedules and auto-scaling inference clusters.

Result: Clients saw 40% cost savings and 100 tons of CO2 reduced across their AI operations. It wasn’t just about efficiency; it was about proving that MLOps can power a greener future.

Why Does This Matter Now?

The U.S. and global tech sectors are under pressure to meet climate goals — think net-zero pledges from AWS (2040) and Microsoft (2030). MLOps for Green AI isn’t a luxury; it is a necessity. It aligns with national priorities, cuts operational costs and positions companies as sustainability leaders. For DevOps engineers, it is a chance to lead the charge, blending technical expertise with environmental impact.

The Road Ahead

Building Green AI with MLOps comes with its challenges — balancing performance with sustainability requires trade-offs, and carbon tracking tools are still evolving. However, the necessary tools are here: Kubernetes for orchestration, Terraform for IaC, GitHub Actions for automation and a growing ecosystem of sustainable ML frameworks.

My next step? Open-sourcing a Terraform module for carbon-aware ML deployments — watch this space.

As we scale AI, let’s scale sustainability too. MLOps isn’t just about accelerating model deployment; it is about deploying them intelligently, for the planet and the bottom line. What is your take? How are you making AI greener in your work? Let’s start a conversation.
April 15, 2025
Survey Sees DevOps Teams Focusing More on IT Sustainability Issues

A survey of 504 DevOps practitioners finds 59% work for organizations that have a sustainability initiative specifically tied to application development and deployment.

Conducted by Techstrong Research, the survey also finds that 54% of respondents work for organizations that have sustainability initiatives tied to IT operations.

Other areas that have sustainability initiatives include improving application software efficiency (45%), site reliability engineering (39%), cloud migrations (37%), artificial intelligence (29%) and platform engineering (28%).

Mitch Ashley, principal analyst for Techstrong Research, said from cloud resource consumption to AI processing to software design decisions, sustainability is now part of the conversation. Sustainability considerations will have a strong influence on the software, infrastructure and operational decisions, he added.

The amount of carbon an IT organization is on one level just another type of cost metric. IT environments that generate excessive amounts of carbon are consuming IT infrastructure resources at rates that are probably inefficient. Optimizing the rate at which IT infrastructure is consumed will both reduce costs and reduce the overall amount of carbon emissions. The challenge is that as application workloads become more distributed it’s difficult to continuously assess how efficiently software is running.

The simple fact of the matter is that many IT teams, especially in a cloud era that makes it relatively trivial for developers to spin up virtual machines, have not been paying a lot of attention to the amount of infrastructure they consumed.

In addition, there are still plenty of opportunities to replace legacy applications with a more modern application that executes code much more efficiently. Much of the code running in production environments today was written at a time when how efficiently that code might run was less of a pressing concern than it is today.

Fortunately, there are more tools available that enable IT organizations to track power consumption as part of an effort to reduce carbon emissions.

The SustainableIT.org consortium has also developed a set of 240 industry agnostic metrics for tracking and measuring the impact of sustainability initiatives. The non-profit consortium also partnered with ServiceNow to infuse best sustainability practices into its IT management platforms.

The Green Software Foundation (GSF), meanwhile, is working with SustainableIT.org to define a set of best practices for writing energy-efficient code.

Additionally, Red Hat has donated an open source Kubernetes-based Efficient Power Level Exporter (Kepler) tool that captures power usage metrics from Kubernetes clusters to the to the Cloud Native Computing Foundation (CNCF).

Cloud service providers now also routinely make available reporting tools to enable organizations to more precisely track how much carbon they are generating.

Right now, tracking of carbon emissions is still a bigger issue in Europe than most anywhere else and it remains to be seen how much regulatory pressure might be applied to organizations in the U.S. as a new administration take office. One way or another, however, the pressure to consume IT infrastructure resources as efficiently as possible is only going to continue to rise in an era where organizations are not just concerned about environmental impact but also the total cost of IT.

For more information, download a copy of the DevOps Next report here.

November 8, 2024
Sustainable IT and Data Innovation: Driving the Future of DevOps

The worlds of sustainable IT and data innovation are converging, and two upcoming events are set to showcase the latest developments in these critical areas of sustainable IT and data innovation. The SustainableIT Impact Awards & Symposium and the DevOps + Data Impact 2024 promise to bring together industry leaders, experts, and practitioners to explore how technology can drive positive change in both business and society.

Inaugural Impact Summit & Awards, SustainableIT.org

Sustainability Takes Center Stage

Sustainable IT has become a key focus for organizations across the globe, as the environmental impact of technology comes under increasing scrutiny. The SustainableIT Impact Awards & Symposium will be held in Austin, Texas. It aims to celebrate the institutions and individuals who are leading the way in this field. The event, organized by the nonprofit SustainableIT.org, will feature a range of awards, including the Environmental Award, which recognizes success in reducing energy consumption and emissions across technology infrastructure, and the Governance Award, which celebrates institutions that have woven sustainability into their decision-making processes.

Meanwhile, the DevOps + Data Impact 2024, presented by Perforce and Delphix, also in Austin, will dive deep into the world of data management and its role in driving innovation. As enterprises grapple with the challenges of managing and leveraging vast amounts of data, the conference promises to offer valuable insights and solutions. Attendees can expect to learn about the latest trends in data analytics, security, and compliance. Plus they will gain insights on strategies for optimizing data management to drive business success.

Perforce’s Acquisition of Delphix: Driving Sustainable Data Management

Perforce’s recent acquisition of Delphix, a leading provider of DataOps solutions, reflects the growing importance of data management in the DevOps landscape. This strategic move aims to help customers automate and accelerate their DevOps pipelines while promoting sustainable IT practices. As Perforce CEO Jim Cassens noted, “Data is at the heart of how enterprises operate today and essential for successful software development, but accessing and managing that data is extremely challenging. Many teams do not have rapid access to solid, high-quality test data.”

By integrating Delphix’s capabilities, Perforce seeks to address these challenges and enable more efficient and sustainable data management practices. As Cassens stated in a blog post announcing the acquisition, “Another unique ability of the Delphix platform is how it virtualizes data and ultimately reduces storage footprints, which is good news for sustainability and operational costs.”

Bridging Sustainable IT and Data Innovation

The convergence of these two events points to a broader trend in the DevOps world: the need to balance sustainable IT and data innovation. Organizations are increasingly recognizing that sustainability efforts can drive efficiency and innovation. Accordingly, robust data management can support environmental goals.

As the DevOps landscape continues to evolve, look for more initiatives that emphasize balance with AI-driven energy management in data centers, sustainable IT and software development practices, and circular economy approaches to IT hardware. Certainly, the upcoming events in Austin serve as a barometer for these trends, offering a glimpse into how industry leaders are navigating the complex intersection of technology, data, and sustainability. By attending these events and engaging with the insights they provide, DevOps teams can position themselves at the forefront of sustainable innovation, driving both business success and positive environmental impact.

September 2, 2024
Five Ways AMD’s ZT Systems Acquisition May Enhance AI Sustainability

AMD’s recent acquisition of ZT Systems for $1.9 billion opens new doors for enhancing sustainability in the data center and AI markets. Here’s how this strategic move could advance AMD’s environmental goals, along with some potential challenges:

1. Improved Energy Efficiency

ZT Systems specializes in energy-efficient hyperscale server solutions. By combining their expertise with AMD’s high-performance processors, data center energy consumption could potentially decrease. Fundamentally, ZT Systems makes servers that use less energy in big data centers. AMD makes computer chips that are powerful but rely on less energy. When you put AMD’s chips in ZT Systems’ servers, data centers can do more work with fewer servers. This aligns with AMD’s goal of achieving a 30x increase in energy efficiency of AI and high-performance computing processors by 2025.

2. Greener Supply Chains

AMD has been actively working with suppliers to reduce environmental impact. Integrating ZT Systems’ practices may further strengthen these initiatives, potentially leading to more sustainable manufacturing processes across a broader network. The challenge lies in harmonizing supply chain practices across different regions, which may have varying environmental regulations and standards. For instance, one region may have strict carbon emission limits, while another is more lenient, requiring companies to manage different regulations. This complexity makes it harder to establish a consistent, eco-friendly supply chain strategy.

3. Smaller Data Center Footprints

ZT Systems’ rack-scale solutions could help AMD optimize AI infrastructure deployment. This may result in fewer physical servers needed, reducing both energy use and the measurable size of data centers. With less space required to house the servers, data centers can be smaller overall. Real estate is expensive so reducing costs associated with building and maintaining larger facilities is a positive. However, headwinds may arise with the transition to more space-saving solutions. It might require a larger initial investment and infrastructure changes.

4. Advanced Cooling Technologies

Effective cooling is crucial for data center efficiency. ZT Systems brings expertise in advanced cooling solutions, which could help AMD further reduce energy consumption in its data center offerings. The challenge here is ensuring these cooling solutions are adaptable to different data center environments and climates. However, ensuring these cooling solutions are adaptable to different data center environments and climates can be a challenge. For example, data centers located in hot and humid climates may require more advanced cooling systems to maintain optimal temperatures. Traditional air-cooled systems account for up to 40% of a data center’s energy consumption.

5. Accelerated Renewable Energy Adoption

AMD and ZT Systems’ combined expertise could accelerate renewable energy adoption in data centers, reducing fossil fuel dependence and carbon footprints. Their integrated solutions offer better power management for variable renewable sources. Data centers could optimize workloads to match energy availability, using solar power during the day and wind power at night. While renewable energy costs and availability vary by location, this approach enables more efficient use of intermittent energy supplies.

By leveraging ZT Systems’ capabilities, AMD is poised to make significant strides in delivering environmentally responsible technology solutions. This acquisition not only strengthens AMD’s market position but also reinforces its commitment to sustainability in the tech industry, despite the challenges that may arise.

August 20, 2024
Advancing Sustainable Practices in Data Centers: Insights from Platform.sh

The environmental impact of data centers is becoming increasingly critical. These facilities, essential to our digital infrastructure, consume vast amounts of energy and contribute significantly to carbon emissions. Clearly, recognizing the urgency of this issue, Platform.sh is pioneering sustainable solutions to mitigate their environmental footprint. By incorporating renewable energy, cutting-edge cooling systems, and robust carbon offset programs, they are leading the way in creating greener data centers.

Insights from Leah Goldfarb

In a recent interview with Techstrong TV, Leah Goldfarb, Environmental Impact Officer at Platform.sh, shared valuable insights into the company’s sustainability efforts. Goldfarb discussed the integration of renewable energy sources, advanced cooling systems, and carbon offset programs, highlighting how Platform.sh is leading the way in green data center practices.

Green Data Centers Explained

Green data centers focus on energy efficiency and the use of renewable energy to reduce their environmental footprint. These centers aim to lower carbon emissions and optimize energy use, promoting long-term sustainability. Notably, according to industry reports, the global green data center market is projected to grow significantly. Check out this impressive projection: USD 198.3 billion by 2032 from USD 61.4 billion in 2023. This surge is driven by increasing demand for sustainable solutions and stringent environmental regulations

Innovative Sustainability Strategies at Platform.sh

Adopting Renewable Energy:
Platform.sh integrates renewable energy sources such as wind and solar power into their data centers. This move reduces reliance on fossil fuels and decreases carbon emissions, contributing to a cleaner energy future. For example, some of their data centers are powered entirely by renewable energy, significantly cutting their environmental impact.

Designing Efficient Infrastructure:
The data centers at Platform.sh are designed with energy efficiency as a core principle. By investing in advanced technologies and optimizing their infrastructure, Platform.sh achieves lower power usage effectiveness (PUE). Markedly, this focus on innovation helps manage energy consumption more sustainably.

Utilizing Advanced Cooling Technologies:
Implementing state-of-the-art cooling systems and optimizing server utilization are critical for minimizing energy consumption. Indeed, intelligent workload management and automation ensure resources are used efficiently, reducing unnecessary energy use. Key point: these technologies are crucial for reducing the overall environmental impact of data centers.

Certifications and Compliance:
Notably, Platform.sh holds certifications like ISO 14001 for environmental management, demonstrating a commitment to maintaining high sustainability standards. Compliance with European Union regulations further ensures rigorous environmental practices. In essence, these certifications show that Platform.sh is dedicated to adhering to the best environmental standards in the industry.

Participating in Carbon Offset Programs:
To address residual emissions, Platform.sh uses carbon offset programs. These initiatives support global environmental efforts such as reforestation and renewable energy projects. By participating in these outlets, Platform.sh helps mitigate the environmental impact of their operations.

Benefits of Renewable Energy in Data Centers

Using renewable energy sources provides significant environmental and economic benefits. It helps reduce greenhouse gas emissions and conserves natural resources. Economically, it can lead to long-term cost savings by decreasing dependency on non-renewable energy sources and stabilizing energy costs. This dual benefit makes renewable energy a compelling choice for sustainable data center operations.

Emerging Trends in Sustainable Cloud Hosting for 2024

Sustainable cloud hosting’s latest trends span from automation to AI. However, green initiatives in digital transformation face challenges.

Automation and AI:
Automation and artificial intelligence enhance energy efficiency by enabling precise management of workloads and resources, thereby reducing energy waste. These technologies are becoming increasingly important in optimizing data center operations.

Decentralized Energy Grids:
Utilizing decentralized energy grids allows data centers to effectively use local renewable energy sources, minimizing transmission losses and improving efficiency. This trend is helping data centers become more self-sufficient and environmentally friendly.

Circular Economy Principles:

Embracing circular economy principles involves reusing and recycling hardware components, reducing electronic waste, thus extending the lifespan of data center equipment. Consequently, this approach not only conserves resources but also reduces the environmental footprint of data centers.

Challenges in Implementing Green Initiatives

Implementing green initiatives in data centers can be challenging. Common obstacles include high initial costs, technological limitations, and the need for specialized expertise. However, prioritizing sustainability and leveraging innovative solutions can help overcome these challenges. For instance, despite the initial investments, the long-term benefits of green technologies often justify the costs.

🎤Tell Your Sustainable Solution Story

We invite industry Thought Leaders in tech and sustainability to share their innovative solutions and strategies. Boost your company’s profile and amplify your brand by contributing to our discussions on sustainable practices.

Contact Bonnie Schneider at bonnie@techstronggroup.com to participate.

May 14, 2024
Tools for Sustainability in Cloud Computing

Cloud computing is a primary driver of modern technological advancements. However, as cloud services adoption continues to soar, so does a concern for their environmental impact. The further expansion of cloud infrastructure will demand more energy, which means the need for sustainable solutions has never been more acute.

Due to their scale, cloud vendors operating large data centers have a greater impact on minimizing the IT’s industry overall carbon footprint than do smaller on-premises deployments. In this article, we highlight the main benefits of migrating business operations to the large-scale cloud providers known for their established sustainable practices, and point to useful resources.

Follow the Leaders

Major cloud platforms, including Google Cloud Platform, Amazon Web Services and Microsoft Azure, have pledged to achieve net-zero carbon emissions in their data centers by 2030. These platforms are well-positioned to lead the adoption of sustainability principles in cloud engineering due to their reputation and global reach.

Microsoft’s study on the carbon benefits of cloud computing showed that an enterprise could reduce its carbon footprint by up to 98% by migrating to Microsoft Cloud, compared to running its IT operations in the on-premises data center.

As for Google Cloud, all of its cloud products are carbon neutral. Google achieves this by purchasing renewable energy to reduce the carbon footprint from operations performed on its infrastructure, as well as by using high-quality carbon credits to compensate for its remaining carbon emissions.

By moving your operations to the cloud and focusing on developing cloud-based applications, you can contribute to sustainable business practices and overall carbon footprint reduction.

Increase Energy Efficiency

Public cloud platforms rely on large-scale data centers with advanced infrastructure technologies. That allows them to reduce electricity consumption for overhead tasks such as lighting, cooling, and power conditioning. For example, according to Google’s environmental report, its data centers consume half the energy needed to power an average on-site data center.

Energy efficiency is also driven by higher server utilization rates and the ability to manage workloads flexibly. These are the main factors that allowed AWS to make its data centers 3.6 times more energy efficient than a median data center run by a U.S. company.

Additionally, the innovative approach applied to server architecture design in large data centers enables cloud service providers to develop custom hardware optimized for their customers’ workloads.

Big Names are Committed to Sustainable Software Development Practices

Another environmental benefit of running IT operations in the cloud is that public cloud platforms are committed to sustainable software engineering – and they want to make it feasible for anyone who uses their platforms.

Green (or sustainable) software starts at the design phase, considering carbon emissions as a project requirement. According to the Green Software Foundation, sustainable software applications are carbon efficient (can run on fewer machines and resources), energy efficient (require less energy to perform the same tasks), carbon aware (prioritize renewable energy sources as a power source) and hardware efficient (use less hardware).

IT operational efficiency is a fundamental of sustainable software engineering. Since cloud service providers already apply advanced technology to improve the energy and hardware efficiency of their infrastructure, cloud migration is a good start for supporting green coding.

Microsoft developed eight principles of sustainable software development for engineers to follow when they build, design, and deploy applications. You can apply these principles to all types of applications, whether deployed on-premise or in the cloud.

Amazon released a sustainable architecture guide with strategies for applying sustainability practices to optimize AWS infrastructure compute, storage and network layers. AWS also provides guidance on applying environmental best practices for designing and operating workloads in the cloud.

A Tools Roundup for Measuring and Reporting Carbon Emissions

In addition to implementing sustainability practices in their business operations, cloud platforms offer clients tools to monitor and measure carbon emissions linked to cloud service usage and report their overall carbon footprint following greenhouse gas protocol standards.

Google Cloud provides users with an emissions calculator. Using Carbon Footprint, you can track the emissions from running your applications. The tool lets you view and analyze data related to gross carbon emissions and use derived insights in emissions reports.

AWS also offers the Customer Carbon Footprint Tool for emissions tracking. You can calculate their carbon emissions associated with using AWS services, track emissions after migrating to the cloud, and use the data for sustainability reporting purposes.

Microsoft customers can access the Emissions Impact Dashboard to estimate their carbon footprint related to Microsoft Azure and Microsoft 365 cloud services. Microsoft also offers Cloud for Sustainability for tracking and managing carbon emissions. It gives you visibility into your organization’s environmental impact, aggregates data from various sources, and offers practical recommendations for reducing a carbon footprint.

Energizing Thoughts

Achieving “net zero” is not solely the responsibility of large-scale cloud providers. Businesses must also contribute. Itransition’s cloud consultants believe that if more companies move their IT operations from on-premises servers to the public cloud, overall energy consumption and related carbon footprint will decrease.

By leveraging renewable energy sources, optimizing infrastructure for energy efficiency and employing innovative technologies, these leading cloud platforms can attract a cohort of environmentally conscious clients. Those users may consider switching to the cloud not only for direct business benefits but also to make their operations more sustainable.

Photo credit: Aimable Mugabo on Unsplash

April 12, 2024
SEC Introduces Enhanced Climate-Related Disclosure Rules
The Securities and Exchange Commission (SEC) has adopted new rules to enhance and standardize climate-related disclosures by public companies and in public offerings. These rules aim to meet investor demands for consistent, comparable, and reliable information on the financial effects of climate-related risks on companies’ operations and their risk management strategies.

The Importance of Transparency

According to SEC Chair Gary Gensler, the initiative is based on the principle that investors should be able to make informed decisions based on companies’ complete and truthful disclosures. The regulations introduce comprehensive requirements for disclosure of significant climate-related risks and their efforts to address them.

Key Aspects of the SEC Rules:
- Mandatory disclosure of climate-related risks that significantly affect or are likely to affect a company’s strategy, operations, or financial condition.
- Details on efforts to mitigate or adapt to climate-related risks, including financial expenditures.
- Disclosure requirements for climate-related targets or goals and their impacts on the company.
- Requirements for certain companies to report Scope 1 and Scope 2 emissions, with an accompanying assurance report. (Scope 1 emissions are direct emissions from owned or controlled sources, while Scope 2 emissions are indirect emissions from the generation of purchased electricity, steam, heating, and cooling consumed by the reporting company.)
Phased Implementation and Review Process

Set to become effective 60 days after publication in the Federal Register, the rules will be implemented in phases based on the filer status of registrants.

Strategic Implications for DevOps and Cloud Computing

The new SEC rules could catalyze a strategic pivot in DevOps and cloud computing towards more sustainable practices. Specifically, they may further incentivize the development of tools and methodologies for accurately measuring and managing Scope 1 and Scope 2 emissions. New roles or teams may be dedicated to environmental impact analysis within IT departments. These shifts would ensure sustainability becomes a core consideration in tech deployment strategies.

Driving Innovation for Greener Technology

Software solutions that automate the tracking of energy consumption and emissions data across cloud and on-premises environments may see increasing demand. Particularly as organizations seek actionable insights for reducing carbon footprints.

This emphasis on detailed climate-related disclosures may also serve to drive innovation in cloud infrastructure design and management. For example, cloud service providers might introduce new features or services that enable customers to more easily assess and optimize the environmental impact of their cloud usage.

This could lead to the adoption of more energy-efficient computing resources, such as serverless architectures. Dynamically allocating resources could be prioritized to reduce unnecessary energy consumption

Advancing Eco-friendly Solutions with AI

These new rules amplify the global lean into AI for faster and more effective solutions. Utilizing predictive analytics and machine learning, AI can accurately forecast computing demand and make changes in real-time. This ensures that energy consumption closely matches actual usage, minimizing waste and reducing both operational costs and environmental impact.

Balancing AI’s Energy Demand

The flip side, though, is the high energy demands of AI. The new SEC rules will likely place an increasing focus on creating and implementing AI algorithms that are both potent and energy-efficient.

Look for greater focus on AI-driven cooling systems for data centers and the use “lightweight” AI models that require less data and computing power for training and

operation. Such models aim to maintain high efficiency while drastically cutting down the energy needed for their training and inference phases. All striving towards the broader goal of making AI more environmentally friendly.

Sustainability and Innovation: The Path Ahead

With the implementation of the SEC’s new rules, the landscape of corporate reporting and technology development is poised for significant change, emphasizing the importance of sustainability and environmental considerations in business strategies and operational practices.
March 6, 2024
A Manifesto for a More Sustainable Cloud Future

I’ve talked about cloud sustainability extensively in these pages previously, but until we start taking tangible steps to collectively reduce our cloud carbon footprint, things are not going to change. Achieving real change comes down to a smarter, more efficient use of resources and establishing a culture of sobriety—both financial and environmental—that flows through an organization. Below, I present the steps necessary for a more sustainable cloud future.

1. Shift the Mindset

First, and arguably most important, there needs to be a shift in mindset that sees IT managers, CIOs and other IT leaders become part of a much broader conversation about how to bake sustainability into the overall IT strategy from the start. In doing so, this offers a path to bring development, finance, ESR and business teams together to ensure greater financial and environmental accountability.

2. Right-Size to Optimize

Regular sustainability assessments can help to identify oversized resources and underused services, which paves the way for smarter allocation. Eliminating unnecessary services and right-sizing instances will help to identify recurring problem areas and serve to reduce carbon emissions. This will create a best practice of optimal resource usage and, consequently, contribute to ongoing savings made to an organization’s bottom line.

3. Embrace Dynamic Scaling

Most cloud providers now offer dynamic scaling capabilities that allow resources to be adjusted according to demand. By embracing these sorts of features, organizations can ensure that resources are used more efficiently while minimizing emissions. This sort of real-time optimization can strike an important balance between performance and environmental responsibility.

4. Fewer Servers Means Less Power Needed

Server virtualization—essentially the process of dividing a physical server into multiple unique virtual servers—offers a clear way of contributing to carbon reduction efforts. By optimizing resources and reducing the physical server count, the consumption of energy can be minimized. Fewer servers mean less power consumption, which directly translates into lower carbon emissions.

5. Only Energy-Efficient Architecture

Platforms such as serverless computing architectures have emerged as a greener alternative to traditional infrastructure, allocating resources only when needed. This creates greater energy efficiency and reduces total carbon emissions. Leveraging these sorts of energy-efficient instance types and hardware, such as ARM-based processors that consume less power than x86-based alternatives, provides a clear way to minimize power consumption and environmental impact.

6. Localization Matters

While the steps above are effective, their impact will be amplified by choosing cloud regions that are based on low-carbon energy sources. For example, regions that are reliant on coal or other high-carbon energy sources should be avoided or used sparingly to minimize environmental impact, but this is not as straightforward as it sounds due to seasonal variations, which might mean that renewable energy sources are subject to fluctuations; the costs associated with storing excess energy from intermittent sources which are both challenging and expensive and the cost of energy trading between regions which can impact carbon footprint.

The truth is that there is no one-size-fits-all solution when it comes to evaluating localization. Assess the energy mix, consider the availability and reliability of renewable energy sources and factor in any seasonal variations or dependencies on fossil fuel-based backup systems. This way, you can make a more informed decision that aligns with your sustainability goals and minimizes the environmental impact of your cloud infrastructure.

7. Align FinOps and GreenOps Strategies

Closely linked to more established terms like FinOps—another framework for managing operational expenditure across an organization, often in conjunction with the cloud and cloud computing—GreenOps is an emerging term that essentially means providing greater cost transparency while promoting a culture of environmental responsibility. The challenge is that making an entire IT ecosystem sustainable requires a step back and all areas of IT to consider their impact on the environment.

This requires collaboration right the way across an organization and with other lines of business which includes areas like HR, marketing, procurement and the accounting departments, all of whom must set business requirements, determine software development and DevOps processes, and adhere to operational principles. Aligning GreenOps and FinOps strategies will allow organizations to start monitoring cloud computing emissions by connecting emissions data directly to their cloud expenses so they can see both the costs to budget and the cost to the environment in the same place.

8. Remember to Turn Off the Tap

Finally, think of efficiency as being the cornerstone of any sustainable initiative, and cloud computing is no different. Optimize cloud instances and you will, in turn, reduce carbon emissions. By evaluating running service instances continuously and strategically and turning off those that are not needed during periods of low demand, organizations can significantly cut down on unnecessary energy consumption. This will not only reduce carbon footprint but also optimize operational costs.

By addressing these complexities head-on, I believe we, collectively, as an industry, have a once-in-a-lifetime opportunity to reduce our carbon footprint and pave the way for a greener, more efficient cloud computing landscape.

October 27, 2023