Seven Architecture Principles for Building a Greener Supply of Data

by Srini Koushik, Chief Technology Officer, Rackspace Technology

a succulent with a rxt circle

 

It’s easy to think about data as an intangible, cheap resource with a limitless lifespan, but the truth is data is an energy-hungry beast that costs much more than many think.

In my last blog, How IT Leaders Can Lead Us Towards a More Sustainable Planet, I encouraged CIOs, CTOs and IT leaders to apply three sustainability filters - make IT green, make IT viable and make IT equitable – to every decision they make across their technology infrastructure.

Environment, Economic, Equity

Now, I want to dive into how we can make IT green by digging into the high cost of data in the IT supply chain. For context, data storage on the cloud is growing exponentially, fueled by IoT solutions and rich media such as video, audio and high-definition images.

According to IDC, worldwide, data is expected to grow to 175 zettabytes by 2025, with 51% of the data in data centers and 49% on the public cloud. So, it’s important for technology leaders to look closely at how data consumption can impact both their business models and the world at large.

At Rackspace Technology®, our approach to data management centers on seven architecture principles. These principles align with guidance for how we manage and store data, how we utilize it, and how we apply technology to it — all with the intent to design and build a greener supply chain of data while meeting the ever-increasing need for new data and algorithms.

 

data containers with hands holding it

Be Better Custodians

First, it’s essential to rethink how we look at data. Just as you manage the efficiency of your home by turning off the lights when you’re not in a room, you can mimic these same principles across your organization.

  • Manage the Information Lifecycle - Implement information lifecycle management (ILM) to actively manage the classification and retention of information.
  • Optimize Cloud Storage - Use and optimize appropriate cloud storage solutions based on ILM policies and patterns of use.

 

Picture cards stacked

Be Deliberate with Your Data

It’s simple to think of wasting energy in terms we’re familiar with day-to-day. For example, you wouldn’t want to upload the same photo to your cloud storage 13 times just because you have the storage capacity — each move or copy expends energy.

  • Implement Database Technologies - Use database technologies that best suit the computing task to optimize compute time and performance.
  • Minimize Redundant Data Storage - Use data architecture patterns to minimize redundant data storage and promote just-in-time transformations.
  • Reduce the Movement of Data - Adopt data architectures that minimize data movement.

 

data container with leaf

Be More Progressive

Start using the right type of technologies by investing in out-of-the-box solutions to get the full benefit of your data while mitigating its impact. For example, utilizing low-energy storage solutions for the information you don’t use regularly could be the first step.

  • Leverage Purpose-Built Data Technologies - Leverage purpose-built data technologies that help avoid redundant copies of the same data – Snowflake®, Databricks®, etc.
  • Invest in Green Machine Learning - Leverage emerging disciplines of Green Machine Learning (using multi-information source optimization) to drive machine learning with optimized energy consumption.

Increased use of machine learning and deep learning techniques has driven up data storage and energy consumption. According to OpenAI, the energy required for deep learning has been doubling every few months, resulting in a 300,000x increase from 2012 to 2018. In addition, AI power consumption doubles every three to four months, and large AI training jobs have a life cycle carbon footprint of five cars.

In the end, there are significant business costs that can stem from a mismanaged data supply chain. But more importantly, a mismanaged data supply chain can come with high costs to our planet. The machines are churning ,even if you don’t see them. So, we must be more proactive in minimizing the waste of energy. And while the advancements in data capabilities signal progress, it also means data could become our biggest energy hog.

We are committed to implementing sustainable architectural principles at Rackspace Technology to make IT green. Read our complete Environmental, Social, and Governance Report to learn more about our holistic approach to building a more sustainable business infrastructure.

Also, check out my other article on Triple Pundit on Why IT Must Adopt a Broader Definition of Sustainability.

Let’s Start Solving Together™ to Make IT More Sustainable