As we progress through the third decade of the 21st century, the world of big data is well underway. Whether you’re buying a coffee at Starbucks, ordering a pair of headphones on Amazon or posting a video on TikTok, you’re contributing to the creation of data. Data creation, to put it mildly, is on steroids. By 2025 global data creation is expected to grow to 181 trillion gigabytes. And to put things in perspective, worldwide data created in 2010 was ‘only’ 2 trillion gigabytes.
The idea that we can generate such a vast amount of data perhaps becomes more understandable when we look at how we use some of the world’s most popular applications. There are estimated to be 99,000 search queries on Google every second, over 500 hours of video uploaded onto YouTube every minute and over 1 billion video views on TikTok each day. Add to this that a large chunk of humanity has yet to access the internet, and it’s apparent that Big Data is about to get much bigger.
So, do we have a problem? With all of this data being created, could we run out of space? Are we fast approaching a day when our favourite applications start setting restrictive limits to the amount of data we create? Maybe that day is closer than we think. Google has updated its inactive account policy, deleting accounts that haven’t been signed into for at least 2 years. On 1st June 2021, Google also reversed its unlimited free storage policy for Google Photos. Could we be on the cusp of a data drawback as we head towards a global data storage crisis?
Well, not really. First, not all data that’s created is saved. Second, storage capacity and the efficiency of storage systems have improved in line with the magnitude of data creation, with promising advances on the horizon such as data storage in single molecule magnets, nanostructured quartz glass discs and even DNA.
The immediate problem is the impact on the environment. In conjuring up an image of a data centre, you may think of rows of physical servers piled on top of each other in an enormous warehouse situated in a remote location such as the Arizona desert. Lured by the prospect of economic benefits and increased tax revenues, cities and towns authorize the development of these data centres despite the environmental cost, for example, of up to 1.25 million gallons of water a day.
Data centres, in fact, are located in various climes, but nonetheless have large environmental impacts. They require an enormous amount of energy to operate and huge amounts of water for liquid cooling and to prevent overheating of servers. In 2018 it was estimated that data centres in the United States had an annual water footprint of 135.47 billion gallons. Given that water is a scarce resource, alternative solutions to land-based data centres are being explored.
Enter the underwater data centre.
In 2020 Microsoft revealed positive results from Project Natick, an underwater data centre deployed 117 feet deep off the coast of the Orkney Islands in Scotland. Consisting of 864 servers, the 40-foot long data centre was kept cool underwater, reducing the need for tapping into scarce freshwater reserves and benefited from a failure rate that was 1/8th of land-based data centres.
Off the coast of Lingshui Li autonomous county, Hainan province in China, the Beijing Highlander Digital Technology Co has deployed its own underwater data centre, estimating a 30% saving on energy costs. By keeping a data centre sealed and water-tight, it remains oxygen and dust-free, preventing corrosion of hardware and reducing failure rates.
Given that 40% of the world’s population lives within 100km of the coast, there’s considerable potential for underwater data centres to contribute not only to a reduced carbon and water footprint, but also a better internet experience due to proximity of users and servers.
Data creation is very much intertwined with the environment given the enormous water and energy costs of storing that data. In another sense there’s a similarity in that we produce huge amounts of data and we produce huge amounts of carbon emissions. However the similarity ends when we look at the almost-opposite paths we’ve taken to the problems of both. With carbon emissions, the problem has been ignored or neglected to such an extent that our solutions are now a combination of mitigation and adaptation. With data ’emissions’, if we can use a comparable term, technology has kept up to minimize it and to store it efficiently; something that the long-overdue carbon capture and storage process has the potential to do for carbon emissions. Creative solutions are being explored such as the aforementioned underwater data centres, and innovations in design are allowing land-based data centres to operate at higher temperatures, reducing the reliance on air conditioning and cooling water. With other potential solutions on the horizon such as DNA storage, we can for the time being upload our YouTube videos, order our Starbucks coffees and watch our Netflix movies.