How Green is Your Cloud? Part 2: Measuring the Environmental Impact of IT Workloads

Tech Blog / How Green is Your Cloud? Part 2: Measuring the Environmental Impact of IT Workloads
Natural glowing Bulb in hand

In the first post of my 4 part-blog series about the cloud’s role in “green IT,” I talked about the importance of developing a sustainability plan for your IT workloads. As mentioned in that post, technology represents nearly 2% of the 38 billion tons of carbon emissions. To get a better understanding of how IT is contributing to this global problem, we can begin measuring the carbon footprint. And the best way to do that with the cloud is by using built-in, cloud-native tools from Google Cloud, Microsoft Azure and AWS.


Google Cloud

Google Cloud Platform (GCP) has a free service for monitoring CO2 emission called Carbon Footprint. It’s located in the console, and data can also be exported for use with other tools.

Microsoft Azure

The Microsoft Sustainability Calculator is a Power BI app. Here’s a great blog post Microsoft published back when it launched in 2020.

Amazon Web Services

The AWS Customer Carbon Footprint Tool was announced at AWS re:Invent 2021. Officially released in February 2022, it provides basic information about workloads’ carbon emissions.

Open-source tools like the Cloud Carbon Footprint have the advantage of being cloud neutral.

While these tools are easy to use, it’s valuable to understand how the findings are calculated. provides a detailed explanation. Following is its summary of the formula used to calculate emissions.

Here’s an easy-to-understand definition list for the Cloud Carbon Footprint equation:

  • Cloud Service Provider Usage – The amount of time that cloud resources are used (compute, storage, networking). Data can be obtained from billing reports or usage APIs (Application Programming Interfaces), such as AWS CloudWatch and AWS Cost Explorer.
  • Cloud Energy Conversion Factors – Represents energy consumed by cloud resources. Factors are different for each type of resource, such as EC2, Lambda, and storage. The unit for this variable in the equation is Watt.
  • Cloud Provider Power Usage Effectiveness – The cloud provider PUE indicates the efficiency of a cloud provider’s energy use. A PUE of 1 means all energy used is for computing with no wasted energy. The goal is to get as close to 1 as reasonably possible.
  • Grid Emissions Factors – Lists the type of local energy sources cloud regions use to power infrastructures. Sources include coal, nuclear, wind, gas, solar, and others. Each region produces a different amount of CO2 for each kWh generated. The unit for this variable is gCO2eq/kWh – gram of CO2 equivalent for each kWh.
  • Embodied Emissions – Represents the additional emissions generated by the manufacturing and disposal of cloud resources. An application using resources for extended periods will represent a small fraction of the total carbon emissions. In cloud environments where resources can be shared, emissions are lower than in on-premise environments where resources are dedicated; the CO2 is already emitted, whether or not resources are used.

Your ability to change these variables depends on where you are on your cloud journey. If you’ve already deployed applications using a given cloud provider and region, the easiest way to lower energy your carbon footprint is to modify the infrastructure and application architecture. If you haven’t selected a cloud provider and region, the PUE and Grid Emissions Factors are the variables to consider.

Variable Time of decision
Cloud Service Provider Usage When designing infrastructure and application architecture
Cloud Energy Conversion Factors When selecting compute, storage or networking cloud services e.g. (EC2 vs Lambda)
Cloud Provider Power Usage Effectiveness (PUE) When selecting a cloud provider
Grid Emissions Factors (GEF) When selecting a cloud provider region

Teads hosts a simple calculator on their website that is great for visualizing the complex math previously listed. As you can see in the following image, this EC2 workload example (m5.4xlarge in us-west-2) running for one year (8,760 hours) will produce 485 kg of CO2 emissions.

For comparative purposes, visit the calculator from to learn what would generate the equivalent of 485 kg CO2 emissions, the amount of CO2 from this AWS workload. As illustrated below, it may tell you that the same emissions would be generated by driving 2000 km (1200 miles) in a car running at 22 MPG (miles per gallon):

Measuring our IT workloads is a great first step. It gives us a starting point. Then we can monitor the evolution of the monthly emissions. Advanced IT organizations will be able to forecast a workload’s emissions before it is even deployed and use this information as part of the architecture decisions. In my next post, I’ll write about five steps you can take to reduce your environmental impact from IT workloads. Stay tuned!

Post Date: 06/20/2022

Matt Buchner, NTT DATA Services

About the Author:

Matt Buchner

Matt Buchner is the Lead Solution Architect at NTT DATA. He brings 14 years of international experience delivering technology solutions to solve complex business challenges in a rapidly evolving business and technology landscape.