In most cases, cloud computing is greener than traditional data centers. Just how green depends on the details of each individual deployment. Credit: Thinkstock For businesses looking to score public relations and ESG (environmental, social, and governance) points for being green, where cloud computing ranks on the sustainable technology scale has been a burning question. How do we answer that? The trouble with cloud computing is that each deployment is different and includes a different mix of technologies. Total carbon footprint is based on a complex array of factors that determine how much power is needed. Is the cloud a sustainable technology? The answer is one that so many people hate: “It depends.” Today, all the hyperscalers boast that their clouds are green and are moving to zero emissions. That’s all true, and good for them; however, when it comes to the greenness of specific cloud deployments for specific enterprises we’re left saying, “yes, this part is green, but this part isn’t green.” It’s deployment-dependent. I’ve covered this before, so a few of these assertions are redundant if you follow me here. However, this is becoming such a hot topic, judging from press inquiries and questions I get from clients. I’ll stand by this: Public clouds are a greener option when compared with more traditional approaches to computing. But cloud may not be green for you depending on the way your company is specifically using cloud computing. Sorry for being the designated buzzkill. However, sustainability is very much a deployment-specific issue, even though it’s not painted that way in the press and by the cloud providers. Indeed, new ideas around this are emerging: The efficiency of the overall cloud architecture: As I’ve covered before, if we’re able to solve the same problems with 300 fewer technologies, then that architecture is driving the cloud deployment to be truly green. What’s interesting is that in many instances, a poorly designed cloud deployment running on a very green public cloud provider generates more carbon than a fully optimized architecture running within a traditional data center. This is not discussed as often as it should be. However, it’s a major determining factor if your cloud deployment is doing the planet any good or not. My point: We can do much better if the solution is optimized rather than deploying an underoptimized solution on what many consider greener technology. The points of presence where the cloud users are located. A cloud provider may have a data center in the United States tied to a wind or solar farm, which is great. However, your cloud services, applications, and data may not be using that data center. The data center your company uses may be in an area, state, or country that uses a coal-fired power plant. Is the provider green? Yes. Does using that provider make you green? Depends on how you’ve deployed your cloud solutions. Power-optimized application development. We’ve all heard of devops continuous testing services that determine if your application lives up to standards of security, performance, stability, etc. How about a standard of power consumption that uses the same automated testing approach? What about the ability to code and deploy an application that requires the minimum amount of power to carry out compute and storage operations ongoing? Most developers don’t focus on this. Considering that they have to prioritize performance, security, stability, etc., asking them to code and test for power efficiency in support of sustainability may be too much. However, it might require a very little amount of work for a large gain in power optimization. Applications and data storage systems can be “power-optimized” to reduce their consumption by half and significantly cut your cloud usage bill. Optimized workloads don’t require the allocation of as many cloud resources, run fewer power-consuming physical servers, and bill you for fewer resources. This is even more helpful for traditional, owned hardware deployed in traditional data centers. If power-optimized application development uses fewer resources by design, companies may delay the purchase of new hardware that will need to be plugged into the wall. None of this should surprise anyone. We just like simple messages that are correct no matter what. Cloud computing is a complex, distributed set of systems, and nothing is going to be simple around defining any of the perceived benefits, of which there are many. Sustainability is one of those things that in order to answer correctly, gets into a complex array of concepts to consider. I’m sorry about that, but it needs to be done. Related content opinion Stopping the rot in AI spending Emerging AI governance tools offer a step toward visibility and control of what’s happening in a company’s AI applications. By Matt Asay Oct 21, 2024 5 mins Generative AI Risk Management Artificial Intelligence opinion Making generative AI work for you Find the sweet spot where genAI boosts your productivity but doesn’t get you in over your head where you can’t tell good output from bad. By Matt Asay Oct 14, 2024 5 mins Generative AI Development Tools Emerging Technology opinion California’s vetoed AI bill: Bullet dodged, but not for long SB 1047 missed the mark. A far better solution to managing AI risks would be a unified federal regulatory approach that is adaptable, practical, and focused on real-world threats. By Kjell Carlsson Oct 08, 2024 8 mins Technology Industry Generative AI Artificial Intelligence opinion Crescendo makes AI boring—and profitable An AI startup is turning call centers into a successful model of using AI to support human employees. By Matt Asay Sep 30, 2024 6 mins Technology Industry Generative AI Artificial Intelligence Resources Videos