Do the right thing: ideas on how to start driving carbon-aware cloud consumption in your organisation (Guest Blog from Airwalk Reply)

Do the right thing: ideas on how to start driving carbon-aware cloud consumption in your organisation (Guest Blog from Airwalk Reply)

Author: Luca Giannone, Senior Consultant, Airwalk Reply

The Problem

It is evident, that time is running out to mitigate the effect of climate change on our planet.

According to the International Energy Agency (IEA),approximately 1% of energy-related greenhouse gas (GHG) emissions, which are a key factor in global warming and climate change, originate from data centres and data transmission networks. In order to align with the Net Zero Emissions by 2050 Scenario, it's crucial that these emissions are reduced by 50% before the year 2030.

How do we deal with the emissions generated by the cloud services used by our organisation? First, we need to understand exactly what those emissions are.

Understanding the problem

The GHG emissions generated by the ICT activities are commonly called 'Digital emissions' and, like every emission caused by human activities, can be classified using three different scopes. The term 'scope' comes from the Greenhouse Gas protocol, which is the world's most widely used greenhouse gas accounting standard.

Source: The FinOps Foundation

Scope 1 emissions

Scope 1 covers emissions from sources that an organisation owns or controls directly - for example from your company's car fleet

Scope 2 emissions

Scope 2 are emissions that a company causes indirectly and come from where the energy it purchases and uses is produced. For example, the emissions caused when generating the electricity that we use in our buildings would fall into this category.

Scope 3 emissions

Scope 3 encompasses emissions that are not produced by the company itself and are not the result of activities from assets owned or controlled by them, but by those that it's indirectly responsible for up and down its value chain. An example of this is when we buy and use products and services (i.e. hardware production, cloud services, logistics, etc)

And now, before being able to reduce the emissions, we need to measure them, to create a baseline which can become one of the key ESG metrics and can be used to define corporate sustainability KPI.

Measuring the problem

The main cloud service providers (AWS, Azure, Google Cloud Platform) provide tools that allow organisations to estimate the carbon footprint of the cloud infrastructure which their digital products are based upon.

AWS Customer Carbon Footprint

The estimate provided by AWS includes only Scope 1 and Scope 2 emissions.

The Scope 2 portion of the estimate is calculated using the GHPG market-based method, which means that it reflects the emissions associated with the electricity they actually consume, rather than the average emissions of the grid on which they are located. This is because companies may have purchased electricity from sources with lower emissions, such as renewable energy sources, or they may have purchased carbon offsets to neutralise their emissions. The estimation is also factoring in the power usage effectiveness (PUE)

The provided data is summarised by time (month), AWS geography and AWS service.

Google Cloud Carbon Footprint

The estimate provided by Google includes Scope 1, 2 and 3 emissions.

The scope 2 emissions are estimated using a location-based method (based on the carbon intensity of the local grid area where the electricity usage occurs).

The methodology used to estimate the scope 3 emissions add several different components, such as the upstream lifecycle emissions of data centre equipment and the Google data centre employees' commute.

Azure Emissions Impact Dashboard

The estimate provided by Microsoft includes Scope 1, 2 and 3 emissions, using a methodology validated by the Stanford University in 2018.

The method to estimate the Scope 2 emissions is market-based, and the Scope 3 is estimated using a complex process that, amongst the other data centre related factors, includes the impact of the HW lifecycle to the overall footprint.

The provided data is summarised by time (month), region, Azure subscription and service.

In addition to the above proprietary tools, there are cloud agnostic open source tools that, using the cloud consumption data provided by the CSPs as inputs, can provide visibility and instruments to measure, baseline and reduce the cloud carbon footprint.

Cloud Carbon Footprint

It is an open source tool, currently supporting AWS, Google Cloud Platform and Microsoft Azure and providing carbon estimation that covers Scope 1, 2, and 3.

The emissions estimate is location-based and uses a methodology that adds the embodied carbon to the emissions generated by the cloud usage.

Unfortunately, the standard cloud services used by organisations are not the only ones generating digital emissions, we need also to consider those generated by the Digital Workplace platforms.

Both Google and Microsoft provide additional tools to quantify the emissions associated with the usage of Google Workspace and Microsoft 365.

Reduce the problem

There are several strategies that can be adopted to reduce the quantity of carbon emissions generated by the consumption of cloud services, some of them are overlapping with FinOps practices, as both of them are aimed at increasing the workload efficiency

Rightsizing

Rightsizing involves optimising cloud resources to match actual demand. This means avoiding overprovisioning, which leads to underutilised resources and wasted energy. By carefully evaluating workload requirements and using tools like auto-scaling, organisations can ensure they're using the right amount of resources at the right time.

Deleting idle resources

Unused or idle cloud resources continue to consume energy, even when they're not actively in use. Regularly identifying and deleting idle resources, such as unused instances, storage volumes, and virtual machines, can significantly reduce energy consumption and associated emissions.

Optimising data storage

Data storage can account for a significant portion of cloud energy consumption. Organisations can reduce their data storage footprint by implementing data compression, deduplication, and archiving strategies. Additionally, storing data in colder, less frequently accessed tiers can further reduce the energy consumption.

Spatial shifting

Spatial shifting involves relocating your computational activities to a different physical area with a lower current carbon intensity. This could entail moving to a region naturally endowed with energy sources that have lower carbon emissions. For instance, one might consider shifting operations between hemispheres based on the season to leverage more sunlight hours.

Temporal shifting

If spatial shifting of your computation to another region isn't feasible, consider temporal shifting, which involves adjusting your workload to times with lower carbon intensity. This could mean running tasks later in the day or during the night when solar or wind energy is more abundant. Technological advancements in weather forecasting enable reasonably accurate predictions of future carbon intensity, making temporal shifting a viable option.

Demand shaping

This is probably the most complicated one, is a strategy that aims to influence and optimise the timing and patterns of cloud resource consumption to minimise its environmental impact. It involves proactively shifting workloads to periods of lower carbon intensity or reducing demand during peak carbon emissions times. This approach can significantly reduce the greenhouse gas emissions associated with cloud computing, contributing to a more sustainable cloud usage model.

Wrapping up, we want to emphasise that this was just a high-level overview of the topic. There's so much more we could do to drive carbon awareness and make a real difference in the fight against climate change.

Public link

Please use the above public link if you want to share this noodl on another website.