Green IT: 7 Lessons CTO Learned the Hard Way

Unpacking Green IT's Realities

Ludi Akue's presentation provides a much-needed, grounded perspective on Green IT, moving beyond theoretical ideals to share concrete, experience-driven lessons. The emphasis on starting with carbon footprint assessment, understanding electricity mix, and the paradoxes within software architecture (microservices vs. monoliths, FinOps vs. true green) is particularly valuable. Her acknowledgment that initial measurements might be imperfect but crucial for momentum is a key takeaway, encouraging practical first steps. The discussion on frontend optimization, highlighting its impact on user devices and equity, is also a crucial, often overlooked aspect.

However, a potential limitation is the depth of technical detail for each lesson. While the concepts are clearly explained, a more in-depth dive into specific tools or methodologies for, say, LCA, electricity mix analysis for cloud regions, or frontend optimization techniques, might have further enhanced its value for practitioners. The presentation touches upon the complexity of microservices and the potential for them to become distributed monoliths, a common pain point. Her suggestion to revert to modular monoliths for simplification in certain contexts is a pragmatic, though potentially controversial, architectural decision that warrants further exploration in terms of its long-term scalability and maintainability trade-offs. The integration of sustainability into OKRs and delivery backlogs, along with distributed leadership through champions, offers a strong organizational framework for adoption.

The audience that would most benefit are CTOs, engineering leads, architects, and DevOps professionals actively involved in infrastructure, software development, and cloud strategy. Developers, product managers, and even finance teams involved in FinOps would also find valuable insights. The presentation serves as a crucial guide for organizations looking to move beyond surface-level Green IT initiatives and implement meaningful, impactful changes, especially as the tech sector's emissions continue to rise and the impact of AI is projected to triple them. The lessons learned are directly applicable to building more sustainable and resilient digital systems in an increasingly volatile world.

Key Points

• The tech sector's emissions are increasing and are a significant contributor to global warming, surpassing the airline industry.
• AI is projected to triple tech sector emissions by 2030, making Green IT a critical imperative.
• Lesson 1: Start with Life Cycle Assessment (LCA) and focus on carbon footprint as a starting point, accepting that initial measurements may be imperfect.
• Lesson 2: Gain control and push for transparency in cloud infrastructure; understand electricity mix and carbon intensity to shape workloads and optimize region selection.
• Lesson 3: Simplify software architecture by focusing on core domains and delegating supporting domains; consider modular monoliths over complex microservices with technical debt.
• Lesson 4: Frontend optimization is crucial; webpage weight impacts user devices, equity, and emissions. Compress images, lazy load, trim scripts, and consider static rendering/edge caching.
• Lesson 5: FinOps does not automatically equal Green IT; cheaper cloud regions can be more carbon-intensive. Rethink decisions and be curious about underlying infrastructure.
• Lesson 6: Enable change through external accountability, training, education on planet boundaries and system thinking, and building feedback loops. Integrate sustainability into delivery backlogs and OKRs.
• Lesson 7 (briefly mentioned): Emerging AI patterns also need consideration for their environmental impact.

---

📖 Source: Presentation: What I Wish I Knew When I Started with Green IT