The volume of data that is generated and stored around the world is enormous. Two years ago it was equal to 45 zettabytes (ZB), one trillion ordinary gigabytes. The worst thing is that their growth is exponential: the IDC consultancy estimates that they will reach 175 ZB in 2025. Technologies such as big data or artificial intelligence, as well as the development of biosciences, add an acceleration that far exceeds the capacity of current media, whether digital or optical.
The (partial) solution to the problem may not ultimately come from electronics, but from biology. A French company founded last year, Biomemory – jointly promoted by the Sorbonne and the CNRS (a counterpart of the Spanish CSCI) together with private investors – has undertaken the development of a device that is described as a “dumb container” for the permanent storage of robotic archives and libraries, considered cold data because they are very rarely consulted, but they take up space and reside on magnetic tapes whose life ranges between 15 and 25 years.
The foundation on which Biomemory innovation is based is molecular, it is based on the DNA that contains the genetic instructions of all living organisms. The founders are the biologists Stéphane Lemaire and Pierre Crozet, who have invented a method and an algorithm for encoding and retrieving data that works on a theoretical principle formulated in 1959, but never applied: the zeros and ones of computing (binary information ) are converted to the quaternary code of DNA, in which the letters A, C, G and T represent nucleotides that are the basis of the genetic material. In order to reduce their cost, they are organized as sequences that cannot be manipulated for any other purpose.
With this brief description, a layman can already fit Lemaire's argument: the copies of the information stored thanks to DNA could be kept intact "for several tens of thousands of years", unlike current media (magnetic tapes, hard drives , DVD or solid state memories), whose average duration is a few years. The example he describes is incredibly compact: once encoded and compressed, all 45 ZB of data would fit into a gram of DNA.
"Once extracted, the molecules that contain the data are purified, lyophilized and introduced into tiny stainless steel capsules with their inert atmosphere," adds Crozet. The capsule is laser welded, in such a way that "we managed to protect the DNA from the three elements that could damage it: water, oxygen and light."
The reading of the stored data can be done with a sequencer stimulated by the algorithm and its subsequent recovery – they assure – has 100% fidelity. At room temperature, this technology consumes no power, unlike digital systems that require cooling: by comparison, existing data centers consume 2% of the world's electricity. The problem to be solved is that the cost of its coding should be reduced 1,000 times to compare with commercial technologies, says entrepreneur Erfane Arwani, who represents the investors who support the project.
The primary interest of this initiative is technological and in this it coincides with the searches of the IT industry (Microsoft, Seagate, Western Digital, Dell). The moral of Biomemory is that the young company emanates from the scientific community and has an objective that contrasts with that of start-ups born around futile applications that receive media coverage that is not always deserved.
