Archiving of big data in DNA

Information has been the driving force behind the socio-economic growth of civilization since its very beginnings. Today, however, the storage, archiving and processing of data by dedicated centers no longer offer sufficient margins for optimization to cope with the deluge of data, and its problematic environmental impact. A recent report by the French Academy of Technologies explores a promising alternative to the conventional model: data archiving on a molecular scale, a 20-year project.

Indeed, the storage and archiving of digital megadata ("big data") using the current approach to data centers will not be sustainable beyond 2040. The global sphere of data (SGD) created by humanity was estimated in 2018 at 33 zettabytes (33 trillion billion bytes), on the same order as the estimated number of stars in the observable universe. SGD increases by a factor of two every two or three years, or about one hundred to one thousand every twenty years. This article does not deal with efforts that might be made to slow the growth of SGD, even though this aspect is part of the wider problem.

Most of this data is then stored in several million data centers (including corporate data centers and the cloud), which operate within transmission networks. Together, these already consume around 2-4% of electricity in advanced countries. Their overall construction and operating costs are in the region of a trillion euros. These centers cover one millionth of the earth's surface (i.e. around 150 km ² ); at the current rate, they will cover one ten-thousandth to one thousandth by 2040 (i.e. around 150,000 km ² , 1/3 of France). The storage technologies used by these centers are rapidly becoming obsolete in terms of format, read/write devices and media, which need to be copied every five to seven years to guarantee data integrity. They also pose growing problems of supply for scarce resources such as electronic-grade silicon.

Bearing in mind that by 2040, there will be a hundred to a thousand times more data to preserve, these figures show that the current conservation model will by then have become insufficient, as well as being environmentally unsustainable.

The promising alternative discussed in this article is offered by molecular carriers of information, such as DNA, used here as a chemical agent outside the living world, or other highly promising non-DNA heteropolymers. Potentially, DNA allows information densities ten million times greater than traditional memories: all current SGD would fit in a minivan. DNA is stable at ordinary temperatures for several millennia, with no energy consumption. It can be easily multiplied or destroyed at will....