Data: The sustainable fuel of the future

In an attempt to address these growing environmental concerns, I have set out some key considerations that can help reduce the footprint and environmental impact of this essential business activity.

Top 3 environmental concerns and recommendations for data usage

1. Data storage and treatment

• The environmental cost of data storage

The digital sector is responsible for around 1.4% of global carbon emissions, which is not far behind the 2.5% created by the aviation sector.

Now, this is largely in the form of electrical consumption. Greenpeace predicts that by 2025, the technology sector will consume 20% of the world’s total electricity — a significant increase from today’s 7%. This is easy to believe given the increase in cloud, AI, and IoT technology that leverage data for intelligent insights.

Data centers (used to store data) are energy-intensive hubs, with temperatures that have to be maintained 24/7. This itself typically accounts for around 43% of a data center’s total energy use. Beyond the energy cost of running a data center, you must also consider the embedded cost of carbon with building the data center and the hardware that resides within. Furthermore, beyond just the carbon footprint, data centers can consume vast amounts of water. In addition, the hardware is made up of precious metals.

Given the environmental challenges with an essential part of modern life, it is vital that we consider how we can make activities related to data treatment and storage as efficient as possible.

• Key considerations for data treatment and storage

• • Obsolescence, redundancy and duplication
    Organizations need to understand the importance of removing data that is no longer valuable. Having processes in place that remove outdated, redundant, and duplicated data will reduce the amount of storage required and the energy required to maintain this storage.
  • Tiering and compression
    Data should be appropriately tiered, i.e. organizations need to categorize data as primary, secondary, and archive material. After categorizing, compression techniques can then be applied to the data that doesn't need to be regularly accessed, thereby saving on the amount of storage required.
  • Working practices
    Organizations should consider best practices around file-sharing and email policies. Emailing file attachments instead of simply sharing links to documents stored centrally will reduce energy consumption. This also reduces the likelihood of multiple files being unnecessarily duplicated by being stored locally.

2. Data capture and processing

• The carbon cost of bandwidth

An internal Atos study conducted in 2021 was based on 2017 academic research. It estimated that 1GB on data transfer consumed 0.0375 kWh of energy, which can be converted to 0.0266 kg of CO2 emitted (CO2e). By applying these calculations, you can arrive at the carbon cost of worldwide data transfer.

In 2021, global carbon emissions were measured as 37.9 gigatons of CO2e, which means the total cost of bandwidth is around 0.026% of global emissions. Consequently, it is another area where considerations must be made in order to reduce the levels of data transfer and associated energy consumption.

• Key considerations for data capture and processing

• • End device design
    Physical devices that capture data should have multiple configurable power modes with flexible clock management. These can be used to leverage software that enables energy-saving features such as disabling unneeded features, sleeping as much as possible, and operating microcontroller units at a lower required frequency. End device design can even be extended to creating renewable energy at device levels required to power devices. Along these lines, Vodafone has developed self-powered mobile towers across the UK, enabling the deployment of new mobile sites in remote locations, eliminating major challenges such as the cost of connecting to the electricity grid.

• • Connectivity and capture protocols
    Always connected protocols should be avoided. Rather connections should only be established when necessary. This will help reduce network congestion that otherwise produces unnecessary retransmissions and energy use.
    Moreover, raw data storage should be reduced as much as possible. Instead, only store the information and insights that the raw data generates. Designs should be focused on reducing data creation/transmission as much as possible too.
  • Processing optimization
    As we have established, transferring data comes with a carbon cost. Therefore we should look to process the outputs as soon as this data has been captured to reduce data transmissions. We should look to focus on offloading tasks from the cloud. Data should be processed at a device level wherever possible, or it should be processed at the edge to avoid processing data in the cloud.

3. Data utilization and insights

Moving beyond the considerations for efficient data handling, we can now consider the benefits of data, using the insights to reduce carbon emissions related to our businesses.

It is important to follow established methodologies, such as the greenhouse gas (GHG) protocol when establishing emission sources.

This is a set of guidelines around the requirements and guidance for organizations that are preparing a corporate-level GHG emissions inventory.

As organizations look to craft their inventories and plans, they should also increase the level of maturity of their data methods.

The more accurate the data is, the clearer the picture of the emission footprint. Thus, this will increase the level of actionable intelligent insights that can be put into practice to reduce emissions across the entire value chain.

Additionally, it is possible to establish a recognized methodology for the carbon abatement associated with your own products and services if they support carbon emission reductions. For example, if a car insurance company launches an initiative to reduce the number of miles driven by its customers, carbon-savings could be measured by leveraging data and applying it to the carbon abatement methodology.

Currently, most carbon-reducing initiatives are measured against the company’s overall annual carbon emissions rather than the individual measurement of initiatives as depicted in the car insurance company’s example above. Not only does this limit the sharing of best practices as highly effective activities can't be easily demonstrated, but it also limits the transparency of how companies are effectively fighting climate change.

There are good examples of data being used to optimize processes such as Port Esbjerg and Atos collaborating to create a trimodal solution enabling companies to select the best mode and route with the lowest carbon footprint.

Embracing data for sustainability

Yes, we live in a data-driven world, and that is the way of the future. Rather than lament this fact, we should embrace it and be mindful of its environmental impact so that we look at this essential service through the same green lens that is often applied to more traditional services.

Embedding green methodology in our data strategy will not only efficiently manage the data lifecycle, but also drive us to leverage the insights from data to reduce emissions from our business’ core services and products.

Topics

Essentials
Environment
Digital transformation
Decarbonization
Data and Artificial Intelligence
IoT and Edge
Cloud