Ian Mitchell CITP FBCS, UKI STS — CTO Advisory Director, considers why sustainability is becoming one of IT’s hottest topics and how making architectural changes can help businesses become greener.

Artificial intelligence is proving to be a transformative force with a hidden cost. Dig behind an AI’s interface, and you’ll often find hugely powerful compute engines and large lakes of data. That data needs storing, moving, shifting, sorting and finally processing. All this adds up to a potentially huge carbon footprint. Against this backdrop, organisations and executives are beginning to focus on sustainable practices within their supply chain and operational ecosystem.

Given the growing importance of sustainability, executive boards are putting pressure on advisers and CIOs to consider the greenness of future IT landscapes — and rightly so.

This is where dedicated sustainability architects and a focus on sustainability within enterprise architecture are becoming hot topics within IT. This article explores some potentially impactful steps all IT organisations can take to reduce their carbon footprint — both in the short and the long term.

Embed sustainability in your architecture practices

A crucial yet simple first step is to review all IT processes and add controls, guidance and the like to drive a sustainable approach. Whilst this seems obvious, it is seldom done.

Start with a review of your enterprise architecture process, templates and principles, ensuring that sustainability is considered at every step of the enterprise design process. Undertaking this step encourages sustainability to be a critical element of the board approval of IT.

A solid next step is to update your design decision template to show the impact of various options from a carbon footprint perspective and cost.

As the custodian of an organisation's technical direction, the enterprise architecture team should also update guidance notes, standards, principles and other guide rail mechanisms. These will encourage product, solution and delivery teams to consider sustainability. 

It’s also worth finding and publishing data on IT solutions’ carbon and energy overheads. With the data in hand, consider embedding sustainability in dashboards and reporting. You could, for example, add a list of the organisation’s 10 least sustainable systems to the CIO dashboard. Data about risk and technical debt would help to bolster this insight.

Tag, monitor and alert cloud services

The most significant advantage of cloud services is the ready availability of resources at the click of a button. From a sustainability perspective, this is problematic because not only do underused resources cost money, but they also unnecessarily increase the company's carbon footprint. 

Bill tagging is a potential solution to this issue. In cloud IT, bill tagging involves assigning metadata, known as ‘tags’, to resources like virtual machines, databases and storage within a cloud environment. Each tag typically includes a key-value pair (for example ‘department: marketing’) that allows for easier resource identification, tracking and categorisation.

Deploying such an approach can help organisations monitor and allocate costs more precisely. You can, for example, break down spending by project, team, environment or any other custom defined category.

Consider testing

Addressing common challenges like leaving test environments running when they’re not in use is important for managing cloud resources and optimising costs effectively. Leveraging cloud provider billing APIs can help teams identify underutilised resources, enabling them to a) optimise usage and b) reduce costs and environmental impact.

One critical step is implementing a consistent tagging strategy across all resources, aligning tags with application services within the architecture. This approach empowers architects to manage costs and reduce carbon impact proactively. Tags can also identify cloud services that don’t need to be operational 24/7; for example, development environments can be set to shut down entirely during nights and weekends rather than just scaling down.

Another valuable measure is setting up billing alerts to notify stakeholders of significant usage increases. This allows quick responses to manage costs and sustainability impacts rather than waiting for monthly reports, which may be too late for timely action.

Reviewing ‘data in transit’ costs through billing tags can also reveal opportunities to adjust architecture patterns for cost savings. Similarly, right-sizing logging practices, such as turning off debug mode when unnecessary, can reduce storage costs by minimising unnecessary log data.

Finally, consider using specialised tools that connect directly to cloud environments, offering insights into your carbon footprint. Ensure these tools provide raw data, enabling more profound analysis and custom insights.

Focus not only on systems but also on data

Whilst on the data theme, it is worth looking at your data archives and associated strategy. You could be storing too much data and keeping it for too long. Similarly, it is tempting to leave everything in production databases if needed. However, if your procedure involves processing cases until they are closed — whether that be a client query, complaint, or just a simple product order to be processed — once complete, move the case data to a recoverable location that isn’t bloating your production case management databases.

Doing so will ensure data stored in the most expensive and resource intensive database is kept to a minimum whilst allowing the data to be still searchable from cheaper, less resource intensive stores.

Finally, from a data perspective, move processing near the data to avoid over-transportation.

A common mistake when processing data is taking vast amounts from one system and passing the information across multiple cloud zones, regions and even cloud suppliers. This often happens for purely historical reasons. Why not deploy lateral thinking and move the processing rather than the data? This may pay dividends in the long run, saving costs and environmental impact.            

Establishing the role of a sustainability architect in IT

Introduce a dedicated sustainability architect role to lead and embed sustainable practices. This role focuses on setting and promoting sustainability requirements, best practices, and patterns for IT architecture, design and coding. The sustainability architect will be responsible for developing comprehensive sustainability guidelines, policies and non-functional requirements (NFRs) to guide all IT teams in implementing sustainable procedures.

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

The sustainability architect will collaborate with testing teams to ensure that sustainability criteria are integrated into testing processes, making sustainability a fundamental aspect of system verification. As a consultant, I've successfully implemented this role for multiple clients, demonstrating its cost-neutral or even positive impact, making it easy to justify.

By establishing this role, your organisation reinforces its commitment to sustainability and cultivates a culture that encourages others to prioritise it. Regular reporting to the IT and executive level boards is recommended to maintain high-level support and awareness — an essential step in underscoring the organisation’s commitment to sustainable practices.

Summary and conclusion

We’ve covered a lot of ground. So, here is a summary and some extra ideas about how architects can help their organisations become more sustainable: 

  1. Design for sustainability from the start: ensure your enterprise and solutions have sustainability in mind
  2. Optimise resource usage: implement autoscaling for services and turn off unused environments or services
  3. Efficient API design: minimise API calls by combining or adjusting payloads to reduce the number of requests. Also, ensure data returned by APIs is relevant; avoid large payloads with unnecessary attributes
  4. Adopt event-based architectures: interfaces provide two routes for data: a read path and a write path. The read path gets or searches data from the back end and returns it to the front. The write path covers updates and edit and delete requests. With this in mind, consider using event based patterns on the write path. Doing so can flatten usage peaks and optimise standby resources during low-demand times
  5. Evaluate technology choices carefully: don’t rely solely on ‘serverless’ microservices, AI, or low-code solutions; consider sustainability impacts. Also, assess consumption by comparing services. For example, AWS EC2 may be more sustainable than other AWS-native services for specific needs
  6. Encourage efficient code: support developers in writing efficient code and considering sustainability during refactoring
  7. Collaborate with businesses and suppliers: work with the company to reduce paper and travel in processes. Ensure suppliers follow sustainable practices (examples are Climate Pledge or B Corp membership)
  8. Establish sustainable software patterns: involve the architecture team in creating sustainable software patterns for developers
  9. Create a ‘sustainability impact’ list: identify and track applications that need to be removed for sustainability, similar to managing tech debt

As the focus on sustainability becomes every more central, it’s vital that organisations and executives begin to focus on sustainable practices within their supply chain and operational ecosystem.