Taking Action to Reduce Carbon Emissions at Every Stage of Software

By Suraj Krishnan

In an earlier edition of The Observer, we covered our adoption of FinOps—how the organization’s embrace of financial transparency in the engineering process helped to significantly reduce cloud spending. FinOps is only one of the most recent examples—following DevOps and DevSecOps—of engineering teams adjusting their approaches to consider business goals like security and cost efficiency.

The pursuit of efficiency in software development doesn’t just have an impact on an organization’s bottom line—it can also help to reduce carbon emissions. One of the most impactful long-term climate solutions for software development is GreenOps, the practice of incorporating sustainability best practices into every step of the software development lifecycle (SDLC). Just as the developer community has embraced DevOps, DevSecOps, and FinOps to make their work more efficient, more secure, and more affordable, GreenOps will enable us to reduce emissions on an organization-wide level through our development practices.

Here’s what GreenOps has meant for New Relic and how other companies can learn from our experience to make progress on reducing their greenhouse gas (GHG) emissions.

GreenOps in practice

At New Relic, our cloud spend comprises a significant percentage of our total emissions: 45.41% in FY23. We recognize that we can make significant progress toward reducing our GHG emissions by streamlining our cloud use and production processes. This situation isn’t unique to New Relic—any organization that builds software can reduce their emissions by making their processes more efficient.

GreenOps is an operating model that incorporates sustainability best practices into every aspect of software development, including choosing what features to build and sunset. Different choices during the SDLC can result in redundancies and inefficiencies, and these redundancies lead to more energy use. Developers can choose which models and languages they use and how they structure their architecture to make operations more efficient, reducing energy use in the process.

First steps to reducing inefficiency

There is a common stumbling block when organizations first begin to embrace processes like DevSecOps, FinOps, or GreenOps: the impulse to take an all-or-nothing approach. The idea of completely changing a product or building a new solution to have a sustainability focus can be overwhelming and will more often than not lead to inaction. But no organization needs to take an all-or-nothing approach—all it takes is intentionality and a first step forward.

For New Relic, one of our most significant and intentional steps towards sustainability was our comprehensive data center exit. We completed this transition in October 2023, moving the entirety of our operations to the cloud. The results are already being felt: while an instance in a data center will run continuously, we can now turn our cloud instances on and off according to need. Our cloud use must be as precise as possible to improve our financial performance, but this precision also delivers parallel environmental benefits.?

What are the decisions you can make on a local, individual level that will have a positive result for energy efficiency? For many organizations, ARM-based instances can have an immediate and dramatic impact on energy efficiency. It’s a matter of organization: just as packing cubes can help you fit more items in your carry-on suitcase, ARM processors can help you achieve the same performance while reducing your GHG emissions. Hand-in-hand with our FinOps goals, New Relic is tracking towards a goal for 100% of our workloads to run on ARM-based instances. These improvements are also future-proof: we can continuously upgrade our instances to the most efficient and environmentally friendly on the market, ensuring that we’re constantly doing all we can to reduce costs and emissions.?

Organizations can also improve their sustainability by reviewing and streamlining their architectures. The rise of microservices architectures enabled many organizations to bring products to market faster, but it frequently led to inefficiencies and redundancies throughout the stack. At New Relic, we regularly review our architecture to streamline services and reduce waste.

Why it matters

We’re no longer in a situation in which engineering organizations have to trade cost efficiency for energy efficiency. You can make your services more reliable, more cost-effective, and more energy-effective simultaneously. It’s a matter of intentionality and will: if you take the first steps to make your cloud use more efficient, it will pay off on multiple fronts.

As more organizations follow this example and embrace efficiency, we expect the largest players in our industry to recognize the movement and take their own positive steps. The growing use of ARM-based instances is a massive step forward, and we hope to see similar developments from the largest cloud providers and data platforms.?

Any organization and any engineer can make a difference by adopting GreenOps principles. No step is too small, and it’s not too late to start. In the months to come, we will continue sharing examples of how we’re making our operations more efficient and sustainable, including metrics that track our progress. We hope other organizations will join us on the journey.