“Working with the microscope”: a revolution in data storage

/ photo grover_net CC
In one of our previous materials, we wrote about a new development in the field of data storage. We, as an IaaS provider, are very interested in technologies that can change our understanding of data centers. Therefore, we decided to devote some more time to data warehouses. Very capacious data warehouses.
In 2014, Seagate announcedits new Kinetic HDD with Ethernet interface, the performance of which is several times higher than the performance of all the company's existing products at that time. The new invention simplifies the process of creating software and hardware storage architectures, reducing total cost of ownership (TCO) and allowing you to quickly respond to the growing needs of cloud infrastructure storage systems.
Each Kinetic HDD has two gigabit Ethernet ports, each with its own IP address, and the Kinetic Open Storage platformenables applications and hosts to communicate with each other directly through hard drives using the TCP / IP infrastructure of data centers using an open API. Such technology can seriously transform the infrastructure of data centers and the IT industry as a whole.
Another Seagate development, the essence of which is shingled magnetic recording (SMR) technology, formed the basis of more than four million drives that were put on the market two years after its launch. A feature of SMR technology is that when recording information on a disc, tracks overlap each other.
This allows you to increase the number of tracks on each plate and reduce the distance between them, which in turn leads to an increase in disk capacity by 25%. The problem of low recording speed in this case is solved with the help of special microprograms that optimize sequential recording operations.
All this leads us to the idea that hard drives will remain with us for a long time, because this, although not without flaws, is a time-tested technology that continues to improve to this day. However, sooner or later, a new way of storing data will appear in the world, which will surpass magnetic disk drives in all respects. Let's take a short digression into the future and take a look at promising developments in the field of data storage.
DNA-based storages
One of the most promising and promising technologies of the future is DNA-based storage . We wrote about it in one of our previous posts . The DNA molecule offers incredible recording density. A teaspoon-sized DNA disk can fit all the data that exists in the world - every song, every book, every video.
Another advantage of DNA repositories is their durability. According to Harvard scientist George Church, the DNA disc can be left anywhere, even in the desert, but the data will remain on it even 400,000 years later.
To encode how long data can be stored in DNA, scientists have coded83 kilobytes of data, namely two documents - the Federal Charter of 1291 and the Palimpsest of Archimedes. The choice of these documents, in their opinion, shows not only the potential applicability of the method, but also its historical importance. According to New Scientist, the cost of the experiment was $ 1,500. According to representatives of ETH Zurich, if DNA is frozen, this data will remain unchanged for a million years.
At this stage, the main obstacles to the mass introduction of a new information storage technology is timetaken to retrieve data. Even using modern decryption technologies, reading a DNA molecule takes many hours - several orders of magnitude more than reading a regular file on a computer. Therefore, this type of storage is not suitable for frequently used data. Moreover, scientists still enter information into artificial DNA and only after that put it in a bacterium.
Of course, at the current level of technological development, DNA synthesis is too expensive, but the prospects for its use for long-term storage of important information are obvious.
It is also worth noting one of the related developments, which allows you to store data in the so-called “soft substances”. The term “soft substance” may refer to liquids, polymers, and even biomaterials. According to a new study, microscopic particles in a liquid can be used to encode the same zeros and ones, just like in modern hard drives. In theory, clusters of such particles will one day be able to store up to 1 TB of data in a tablespoon of liquid.
Atomic data storage
“What if we could arrange the atoms in random order?” Asked the famous American physicist Richard Feynman during his lecture, “There is still a lot of space below,” in 1959. The scientist suggested that a "manipulator" of the appropriate size will allow the movement of individual atoms. This would mean that information, such as text, can be written using the atoms themselves, and then the entire Encyclopedia Britannica would fit on the tip of the pin. Thus, Feynman laid the future foundations of nanotechnology.
Three decades later, a group of scientists from IBM managed to do what Feynman described. Using a tunneling microscope, they laid out a company logo of 35 xenon atoms on a nickel surface, demonstrating for the first time the ability to move individual atoms.
To prevent the displacement of atoms and keep them in their places, the researchers had to lower the temperature to -269ºC, which is only 4ºC above absolute zero. The experiment was so expensive that writing more than three letters made no sense.
In July 2016, a team of scientists from the University of Delft, located in the Netherlands, made a new breakthrough in the field of data storage at the atomic level. Instead of three letters, they managed to write down a whole paragraph of text (approximately 1 kilobyte of data).
The main disadvantage of the new method is the high demands on external conditions. So that the atoms do not move arbitrarily, the storage ring must be cooled to the temperature of liquid nitrogen (minus 196 degrees Celsius). However, such a method is much cheaper than liquid helium cooling, used at the time by IBM.
Unlike IBM researchers, Dutch scientists stored information not in letters, but in binary code. The essence of the technology is to place chlorine atoms on a copper plate, where they naturally form a grid with square cells. By alternating atoms with empty spaces between them, as well as moving atoms across the plate to empty spaces, scientists get “patterns” that can be read - just like QR codes are read.
“Imagine Fifteen. Each bit consists of two positions on the surface of copper atoms and one chlorine atom, which we can move between these positions. If chlorine is located in the upper position, and the “hole” is under it, then this means unity. The return location is zero, ” said Project Manager Sander Otte. Chlorine atoms are located quite densely, due to which greater stability is achieved in data encoding than in the IBM experiment.
The lattice was so stable that scientists were able to build 1,016 atoms in a 96 by 126 nanometer sized area (for comparison, the human immunodeficiency virus occupies 120 nm). This density of information storage allows you to fit 78 trillion bits per square centimeter - this is hundreds of times higher than the capabilities of modern hard drives.
Such a high density will allow in the future to expand the memory capacity of phones, computers and data centers. But first, scientists need to adapt the technology to work at room temperature. In addition, the dubbing speed is still low - only 64 bits in 1-2 minutes. Dr. Otte believes that he could increase it to a million bits per second, but this is still a thousand times slower than modern HDDs.
Be that as it may, the idea of creating such a drive is very promising if you develop it in the right direction. It is noteworthy that Dr. Otte decided to write down the first 160 words from the famous Lecture of Freinman on his nuclear storage device. It turns out that he was right: we can actually arrange the atoms as we please.
Quantum data warehouse
Quantum encryption has long been seen as a prospect for the development of superfast computers, which are based on qubits (quantum bit), rather than traditional binary data. The fact is, by virtue of its nature, the information capacity of a qubit is higher than the usual bit of binary logic, however, the most interesting here is not even the density of data storage, but their security.
Researchers from the German Max Planck Institute for Quantum Optics presented their new development, which claims to be the future quantum storage facility. For the first time they succeededto preserve the state of a quantum bit in an artificial diamond crystal at room temperature for longer than one second. This discovery allows us to overcome the main obstacles to the creation of ultrafast quantum computers.
Note that the previous record for storing quantum information at room temperature was only a few milliseconds. When extremely low temperatures are reached, this indicator can be significantly increased, but such an approach is impractical if it is necessary to create computing devices for ordinary consumers.
The role of a qubit in the "diamond" memory is played by a carbon atom, more precisely, a carbon isotope atom C13. An isotope nucleus has a so-called nuclear spin that generates a magnetic moment, due to which it behaves like a magnet oriented in parallel (then the value of a qubit is equal to a conditional “unit”) or perpendicular (then its value is equal to a conditional “zero”) magnetic field applied from the outside.
Due to the effect of quantum superposition, the nucleus can be in two states at the same time - “parallel” and “perpendicular”, which allows us to write more information to the set of such qubits than to the classical binary register.
However, all information exchange operations with such qubits do not occur directly, but through the so-called nitrogen-vacant center, which, in fact, is the main invention of researchers, which allowed to extend the qubit's lifetime to one second.
According to researchers, in the future, quantum data storage systems can provide the most reliable means of authentication (passports, certificates) and secure payments (credit cards) that are protected from fakes and fraud. After all, if information is recorded in a quantum state, then it cannot be correctly calculated without knowing the initial recording parameters (for example, the direction of polarization of light).
Any attempt to find out the state of a quantum system changes it, and blind cloning is fundamentally impossible. For these reasons, popular brute force attacks against quantum systems are futile.
So far, do not rush to conclusions and declare that any of the described technologies will be able to enter the market. However, these developments will definitely contribute to the development of storage devices. In the meantime, it remains for us to monitor technological progress and store data on hard drives. According to the forecasts of the international consortium ASTC (Advanced Storage Technology Consortium), the capacity of the HDD will grow to 100 terabytes by 2025. It seems that this is enough for a while.
PS Interesting materials on the topic from our blog on Habré:
- “Mortgaged by nature": DNA-based data storage system
- Silence is Gold: 13 Things You Shouldn't Say to Developers and Testers
- Scientists again “puzzle”: Is brain upgrade possible and what can be expected from it
PPS In addition, we have prepared links to practical manuals if you have time to get acquainted with our IaaS provider 1cloud and explore its capabilities: