Data centers are the new plastic.
A paradigm shift from data centers to swarm-like infrastructures is not just a moral duty: it is an ecological necessity.
The creation of a thousand forests is in one acorn — that is, in layman terms, the smallest thing can spawn the greatest good. The reverse is also true: the destruction of something as meaningful as a thousand forests depends on the smallest things. This is not just a metaphor. Indeed, our most insignificant, day-to-day actions, however innocent they might seem, are causing a planetary ecological catastrophe right as we speak.
I’m not talking about plastic. I’m talking about data center-based clouds, the new plastic. We do about 1.2 trillion Google searches per year — and this is just one app. The Cloud is much more. Messages, social media, music, movies: the Cloud has become so pervasive across the services we use every day that often we are not even aware of it. According to Greenpeace International (see Mark P. Mills, The Cloud Begins With Coal, Aug’13; Greenpeace International, How Clean Is Your Cloud?, Apr’ 12.), the ICT (Information Communication Technologies) ecosystem is estimated to be responsible, as of today, for 10% of the total worldwide energy demand, about 1500 TWh/year, roughly equivalent to the combined energy production of Germany and Japan. Cloud storage, mainly centralized upon enormous, densely-packed data centers, constitutes a non-negligible part of it.
Why is this model so dominant? Because data centers are the first infrastructure that comes to mind when it comes to data storage. You need a multitude of servers working in parallel under direct supervision, then why not put them all together and filling up a building with them? Yet, as we discovered, the data center is not the most energy-efficient model. In fact, quite the contrary.
The energy consumption of a data center revolves around two factors: storage and transfer of data. As far as the storage is concerned, cooling and powering a standard hardware requires an average of 11.32 W per TB as shown in the table below.
On the other hand, the infrastructure to download a file from the cloud, routing it through a relay network across hundreds of kms, consumes about 24kJ per transferred GB, equivalent to 24 seconds of hair dryer use (see Jayant Baliga, Robert W. A. Ayre, Kerry Hinton, and Rodney S. Tucker, Green Cloud Computing: Balancing Energy in Processing,Storage, and Transport, Proceedings of the IEEE 99.1 (2011): 149–167).
Fortunately, an alternative to the data center paradigm already exists in distributed infrastructures like the Swarm, where data are stored not in proprietary server farms but over a glocal network of peer-to-peer interacting devices. Simply put, the Swarm is to servers what AirBnB is to homes. Cubbit, in fact, provides its user base with a server infrastructure without owning any servers. Thanks to the power of distribution, Cubbit beats the average storage energy consumption by 53 to 87%, requiring about 2.5 W for each TB stored in Cubbit.
Not only that: by employing a suite of machine learning algorithms as coordinator, Cubbit can optimize the geographical location of distributed files so that stored shards are close to the owner, therefore reaching an overall proximity to the user that is inconceivable for centralized solutions (check Cubbit’s green paper at https://www.cubbit.io/static/media/greenpaper.pdf for deeper insights).
Long story short, the distributed model is not just, as we previously argued, the only one able to guarantee true privacy. It is also the only one that, when it comes to eco-sustainability, actually walks the walk. Indeed, we should train ourselves to look at the data center as we look at plastic: a highly-polluting, world-threatening relic that we must consign to the remote past.
