- Digital Economy Dispatches
- Posts
- Digital Economy Dispatch #132 -- How Data Centers are Powering Up the Digital Economy
Digital Economy Dispatch #132 -- How Data Centers are Powering Up the Digital Economy
Digital Economy Dispatch #132 -- How Data Centers are Powering Up the Digital Economy
21st May 2023
To many people, the technologies that power up the digital economy may as well be magic. They just work (until they don’t). Phrases such as “internet-connected” and “cloud-based services” are meaningless jargon. They have no insight into the world beneath the surface of the mobile phone app. The disembodied voice of Alexa may as well be a phantom living under the sofa. Their car’s GPS could just as easily be based on magnets under the road surface rather than satellites in the sky.
Perhaps that’s not a bad thing. As these technologies melt into the fabric of our lives, they disappear (literally and figuratively) to become something else. A mysterious force that powers our universe. They are embedded into devices and tools we use every day, delivered as utilities that can be consumed at-will. We connect to them rather in the way most people in the UK use electricity and the National Grid, or the sewage treatment plants and miles of pipes that deliver water to our homes.
But this view also brings problems. In particular, the lack of understanding can have severe consequences when a knowledge of the details is needed to inform debate, or guide behaviours. Let’s take the Internet as an example, and the often-heard mantra of how important it is to “tell kids to be safe online”. Basic instructions about not sharing passwords is all very well. However, this well-meaning campaign struggles to have deeper impact if the audience for the message has an extremely limited view of how the Internet works, to what they are being connected, where their information flows and resides, and so on.
The Price of Fame
To dig deeper into the implications of this digital technology knowledge gap and its implications, there is an exercise that I use with students to start a discussion that you might want to try with your colleagues. Imagine that you go to a web search engine such as Google and look for something to buy such as “grey trousers” or “a red sofa”. Immediately you see a response with lots of information. Consider a few obvious questions about what just happened and see how confident you are about your answers:
Where did this information come from?
How do you know it is accurate and free from errors?
Who decided what information you see and how it was ordered?
Why do 2 people doing the same search at the same time see different results?
Broadly, we all have a sense of the answers. But when pressed, it is amazing how quickly the dialogue dries up. I have found that these kinds of questions and the discussions that take place about their answers gives a really useful starting point for considering the way the Internet works. It initiates a conversation about not only the technology that connects us, but also the ownership and management of information that flows across those links.
However, the question that causes most debate and controversy in this scenario is perhaps one of the most obvious:
What did it cost to perform that search?
The idea that a Google search has a cost comes as a shock to some people. Surely the cost is zero. However, a more considered view will suggest 2 distinct ways to view the economics of this activity. The first is to look at this transaction from the perspective of the “attention economy”. A term that was coined by Herb Simon to highlight the value that is placed on getting people’s attention in a world where there are so many calls on our time and so many distractions to keep us from taking action. This has become a fashionable way to view the cost-value equation, driving us toward a world of surveillance capitalism.
However, beyond Simon’s more esoteric view of cost, the second approach is to recognize that there is also a more traditional view of the costs of using digital technologies. In an age of Internet-based services, it is easy to forget what lies beneath. That is a big mistake. All such events require a computer to perform processing. This requires physical infrastructure to be constructed and power to be consumed. This is true whether that is on a mobile phone in your pocket or on a server a thousand miles away. In a world that is becoming increasingly aware of the environmental impact of our actions, establishing a more realistic view of the costs of the digital economy is essential.
What Lies Beneath
The powerhouses that sit beneath the Internet are data centers. Simply stated, data centers are the physical facilities used by companies and institutions to house the computers, network, storage, and other IT equipment that supports the digital economy. In the ever-evolving digital landscape, data centers serve as the backbone of our interconnected world. These massive facilities play a pivotal role in storing, processing, and delivering the staggering volumes of data that power our daily lives. They may be constructed to be specific to one company and co-located with their other facilities, or be shared and distributed across remote locations around the world.
The complexity and diversity of approaches to data centers can make them a difficult topic to understand. Something that many authors have tried to address. For example, in the excellent book "Tubes: Behind the scenes at the Internet" by Andrew Blum he takes a journey through the physical infrastructure of the internet. Blum literally follows the path of the “tubes” that connect us to reveal more about the hidden world of undersea cables, data centers, and network hubs. Following him on this path is very instructive.
What he discovers is much more than a jumble of intricate technologies. Rather, he demystifies the people and process that govern the basic operational foundation of the internet: Computers connected to other computers through cables under the seas, across continents, and now more frequently using satellites and other wireless networks. His most evocative descriptions are focused on his experiences visiting various data centers in the US and Europe. In the places “where the data sleeps”, Blum paints a picture of what it is like to step inside a large data center. Something that few have experienced. (You can get a good high level perspective by watching this short video).
What he finds are thousands of computing and data servers surrounded by a complex arrangement of generators and cooling devices in warehouses covering several acres of land. Through this voyage of discovery, he encounters a data center industry largely hidden in plain sight. The dimensions and impact of this industry are worth considering. According to a detailed data center landscape review in 2022, there are over 100 third-party Data Centre providers identified in the UK, with just under 300 data centre facilities occupying 1 million square metres of floor space. These data centers are responsible for hosting and processing the data that powers the UK's digital economy, which is estimated to be worth £180 billion per year.
Furthermore, the growth of the digital economy has led to an increasing demand for data centers. According to Data Centre Dynamics, the UK's data center market is expected to grow by 22% by 2025, with a total IT load of 3.3GW. This growth is driven by the increasing adoption of cloud computing, the internet of things, and AI, which require large amounts of computing power and storage. Indeed, demands for increased data center capacity is substantial with rapid adoption of generative AI tools such as ChatGPT and others.
As a result, as the demand for data centers grows, so does their energy consumption. Reports about the extent of their energy use vary, but are considered to range between 1-3% of all global energy use. This is raising concern and drawing attention to the high cost of Internet services.
Given this energy footprint, data centers are adopting more sustainable and energy-efficient practices. For example, many data centers are using renewable energy sources such as wind and solar power to reduce their carbon footprint. In addition, large data center providers such as Google are looking to advances in AI and machine learning to optimize their energy usage and reduce waste.
However, a better approach is to avoid energy use entirely. To address this challenge, investments are being made in a range of new technologies. This is essential when data centers generate a great deal of heat which must somehow be cooled or exchanged in a cost-effective way. Only about half of the energy consumed powers computing and data servers. The rest is required for cooling, lighting, and other non-computing activities. No surprise that adopting energy-efficient technologies such as virtualization and liquid cooling to reduce energy consumption has become critical.
Power to the People
Every digital action comes at a cost. To understand this, you need to ask some hard questions about how the internet works. When you do, you’ll find that at the heart of the digital economy are data centers. They provide the computing power and large storage banks essential to today’s intelligent, internet-connected solutions. But to do this, they consume a lot of energy. With increasing demand for digital services, data center growth will continue and energy consumption rise. By adopting more sustainable and energy-efficient practices, data centers are reducing their carbon footprint, lowering their operating costs, and ensuring the long-term viability of the digital economy.