Publications about the developments of the Russian IT company Promobit, which produces its own data storage systems on Russian and foreign processors under the BITBLAZE brand, appear from time to time in the news feed of the Technosphere media resource. Russia". Maxim Koposov, the general director of the company, gave a long interview to the editor-in-chief Evgeny Belkin. In our conversation, we touched on a variety of topics – from the current situation around the Elbrus ecosystem in our country to the creation of our own microelectronics using modern technological processes, the development of an engineering school and the formation of public interest in the national scientific and technical history.
— Promobit has been operating in the Russian IT market for 15 years, and 11 years ago it first started working with domestic Elbrus processors. That is, practically your company is one of the first developers of systems on Russian hardware. How did you make such a decision in 2013?
— Since 2011, when the company was founded, we started developing our own data storage servers. Initially, they provided hosting services and created their own film exchange platforms. This required infrastructure, and we wanted it to be inexpensive. At first, they used used servers, then they saw that some companies produce them themselves, and decided to try it too.
The first was a high-density server under our BITBLAZE brand with 45 disks and 4 units. Then it was almost a record density. They began to make such platforms. First for yourself, then sell to other customers. But we understood that in this mode it is hopeless, because we need to invest in the creation of new models, in modernization, because the equipment becomes obsolete quite quickly, and we ourselves have already seen certain disadvantages. It was unclear where to get funds for investments: what we earned from selling was only enough to cover production costs, without development.
We looked at the direction in which the market is developing, and what trends are present. We saw that the Chinese were starting to produce such high-density servers and understood that Chinese equipment would always be cheaper due to high serial production. Even if it is not cheaper yet, it will be cheaper in a few years. On the other hand, there were Western manufacturers of the so-called "class A" (IBM, Dell and others), and their cost of production could be 10 times more, or even more. It would seem that the money is the same: if you sell your products 10 times more expensive, then you can already develop. But to sell for 10 times more expensive, you need to invest a lot of money in marketing, in training, in service, and we did not have such an opportunity. What should I do?
I remember studying the situation: in the market of cheap solutions – companies that cannot pay a lot of money – there are Chinese, in the market of large companies that want high–quality service, training programs, marketing - there were Americans. Perhaps there is some direction where there are neither Chinese nor Americans? The time of such reflections coincided with the moment when, quite by chance, I saw a video on the VKontakte social network with a dual-core Russian processor Elbrus-2C+ at 500 MHz. It showed a monoblock on which Linux was loaded. Konstantin Trushkin, Commercial director of MCST, the processor developer company Elbrus, presented the novelty. I went to the ICST website, saw the contacts there, wrote, and Konstantin Trushkin personally answered me. The correspondence and interaction began.
According to their results, in 2013, at the international exhibition Russian Arms Expo in Nizhny Tagil, we showed for the first time a BITBLAZE server with a large number of disks based on a Russian processor. There were positive reviews from the professional community, engineers, heads of departments of enterprises. But we did not immediately understand how to sell to specialists working in the field of the military-industrial complex.
Since the product was already there, we started looking for a niche to sell it in. It so happened that professional businessmen come from the market, from market analysis, from business ideas, and we came from the product. It turned out that the market was found because sanctions began in 2014. However, at that time, most IT specialists believed that everything would end quickly…
— Nevertheless, interest in the topic arose, and it grew. I remember perfectly well how the readers of our media resource reacted violently to the message dated May 1, 2014 about the beginning of the commercial sale of workstations on Elbrus-4S...
— Over time, interest began to appear, primarily from government agencies and subordinate institutions performing various tasks. For example, the Ministry of Communications of the Russian Federation (at that time it was not even called the Ministry of Digital Development, Communications and Mass Communications of the Russian Federation) and subordinate institutions built various state information systems. ru/archives/2170" target="_blank" rel="nofollow">Mir became the first state information system in which Russian data storage systems based on Elbrus processors began to be massively used, and we are proud that our company contributed to its creation
When a Russian citizen issues a foreign passport, his personal data and biometric parameters are collected. Our equipment is used to store such data and provides a high level of protection. 130 BITBLAZE servers for data storage were developed by Promobit, manufactured and supplied to equip this data center of the FSBI Research Institute Voskhod. The storage system used the most productive Elbrus-4S serial processors at that time…
 |
| 64-bit universal microprocessor Elbrus-4S. |
| Source: https://posho.in/ |
— It turns out that your company is to some extent a pioneer and conductor of domestic technologies and solutions in the IT field. There were such "entry points" from where import substitution begins: you supplied the equipment, and it showed that it is quite possible to use Russian equipment for those tasks where previously there was only foreign hardware and software. What can be noted from among such implemented projects?
— Of course, given that there are very few Elbruses themselves, especially if we are talking about the latest processor – the 16-core Elbrus-16C, it is important for the development of the industry to put it not just where they will pay more money, but where it will lead to some kind of turn in the development of the ecosystem. This is always important for us: in the future, such a supply will increase the demand for our products. In conditions of processor shortage, we tried to make just such deliveries.
So, in 2023, the Elbrus-16C processor was used to supply equipment to the Siberia company from Yakutsk, which used this processor to make its own system for medical institutions. Their software product works on our hardware solution, connected to a tomograph, which employs a radiologist who evaluates the health status of a person undergoing a tomographic examination. This is very important, and suggests that Russian products can be used for such complex tasks. But it is also a critical information infrastructure. To date, the confrontation between different countries is gradually reaching such a level that sanctions are imposed on the supply of medical equipment. The governments of unfriendly countries are setting up their manufacturers for this. And tomographic examinations have already become vital.
Another interesting story about Elbrus-16C is its delivery to our great friend Mikhail Shigorin, one of those who stood at the origins of Alt Linux, the oldest Russian distribution with an independent repository, which is now advancing on the market under the Basalt SPO brand.
Back in 2015-2016, Mikhail was among the first enthusiasts who contacted Elbrus. He has been engaged and is engaged in porting and adapting the distribution for use with Elbrus. This is a very important job, because the MCST company supplies its distribution kit along with the processor. It should be understood that the MCST is still focused on creating processors, and their distribution kit is just a way for them to show that the processor is working, a kind of demonstration element, and Alt Linux is still a distribution that has existed and been developing for many years. There is a wealth of experience, established procedures, updates, support and quality control. This is a distribution that is used in a lot of important critical applications, has regulatory certificates, for example, the FSTEC certificate on the absence of undeclared features or, as they say, "bookmarks".
And so Mikhail Shigorin used his personal funds to purchase equipment for Elbrus-16C. Thanks to this, in fact, I ported Alt Linux to the latest generation of Elbruses. It is worth noting that Alt Linux was previously ported to the previous generation, which includes Elbrus-8CV, and it could also work on Elbrus-16C. But the fact is that the processor "16C" is not only distinguished by a large number of cores and a smaller process, it represents a new, 6th generation of the Elbrus architecture, performs an even greater number of operations per clock cycle, it has additional extensions and, most importantly, it has virtualization. And without virtualization, it is almost impossible to imagine a modern server infrastructure.
First of all, when servers are purchased, installed in any data center, as a rule, a hypervisor is installed on them, which further allows them to be divided into a large number of virtual machines. Due to the fact that Elbrus-16C was in the hands of Mikhail Shigorin, firstly, he adapted Alt Linux in order to achieve maximum performance on this particular generation of architecture. Secondly, he was the first to check and configure virtualization on Elbrus-16C. By the way, the Elbrus-2C3 processor currently on the market is built on the same 6th version of the architecture and also supports virtualization. Now that we have some clients who say that virtualization does not work for them or something did not work out, we are sure that everything will work, and if necessary, you can seek professional advice from Mikhail Shigorin at Basalt SPO. Therefore, such a supply has become really important for the development of the entire ecosystem.
— In our conversation, we came to a topic that has been heard in the professional community in the last couple of months in connection with the decision of the MCST company to open source and make the Elbrus ecosystem accessible to all developers . Based on this decision, what are the prospects for further development of the topic with Russian Elbrus processors? And what about the factor that the number of these processors is limited now, but there is no production yet?
— In order to assess the importance of this decision, let's take a short digression into history. The MCST company develops a processor architecture, a command system, it develops a system on a chip, where there is a processor itself and additional peripheral interfaces, for example, PCI-Express, USB, memory controller. Next, the company creates a motherboard for this processor, then the operating system for this processor, of course, makes a compiler that builds this operating system in C and assembly language. As a result, a huge amount of tasks is obtained, which is solved by one fairly small team of MCST specialists. For example, we just talked about hypervisor and virtualization, let's also take a binary translation system for running x86 codes on Elbrus. Most of these tasks are closed by only a few highly qualified specialists of the company. In this mode, you can make a demonstrator, proof of concept, prototype. Some therefore accuse the company of being closed, that they do everything themselves, accuse of the lack of documentation in the public domain. Sometimes even in the absence of documentation as such. This is all explained by the lack of resources. That is, the resource is spent only on the most necessary things: to make it work, and all the forces are thrown at it.
Now let's take a situation where some Russian company that produced servers, for example, on Asus or Supermicro motherboards, and purchased them until 2022 in Taiwan, says: "We want to make our equipment on MCST motherboards." Such a request is not solved quickly. Then why is it solved quickly by the same companies Asus or Supermicro? Because these are huge companies working with the products of the processor manufacturer, the same Intel. They have built technical support for many years, established a system of control and testing, for example, compatibility with different memory modules, different devices. That is, they are purposefully working to create a market product, which includes not only the release of documentation, but also service, testing in all user scenarios, etc. MCST could not afford this due to its small size. Abroad, such issues were dealt with by external companies, that is, not Intel itself, which creates processors, but the same Supermicro and Asus. And we simply did not have such integrators, or rather OEM/ODM manufacturers. Why?
Firstly, because it probably didn't make any economic sense at first. That is, if a company wants to make money, it needs to supply the equipment that works exactly, has already been definitively tested, and at the same time is cheaper. Therefore, they supplied foreign equipment and became an integrator or marketing vendor who would deal with the task at the very end of the long chain from processor development to the client – to make a product based on ready-made imported platforms with support and documentation in Russian. And all the intermediate links – the development of motherboards, BIOS microcodes, appearance, body – everything was done, as a rule, by Taiwanese and Chinese companies.
When the sanctions began, and state regulatory measures began to work in Russia, they began to direct, at least, state customers to purchase equipment on Russian processors. Equipment manufacturing companies are faced with the fact that there are not enough of these intermediate links within the country, and they do not arise. Although there is a demand for them. Why don't they arise? Because it is difficult to interact with the MCST – the company does not have a designed product in the form of an OEM platform for computer assembly. Designed, this means with technical and marketing documentation, well-developed partner and service programs, etc. Because there are no resources for it, and because it's not their business. As I said above, there are just enough resources to make it work. In the classical sense, what the MCST usually does should be carried out by a cooperation of 5-7 companies, each of which is deeply immersed in its own specifics. To create your own product, you often need design and software documentation (source codes). Yes, it is physically there, but part of it was written when the work was carried out by order of the state, and there the terms of the contract were such that in matters of transferring source codes and documentation that are the intellectual property of the customer, the MCST is not authorized to make such decisions itself. Roughly speaking, it has no right to do this. Everything belongs to the customer in the person of one of the government departments and it is necessary to settle legal issues.
Now, after the ICST has declared its openness and has already published the source codes of the operating system, compiler, and system libraries, new opportunities for further development have opened up. By the way, before that, they had already started sharing reference designs of motherboards on request. That is, conditions are being created for teams to arise that form an ecosystem around the Elbrus processor.
Actually, our company is one of such teams, but we became such long before it became mainstream and justified from the point of view of the market. Now conditions are being created to ensure that there are more companies like ours.
It is worth noting separately the assessments already expressed by some market participants, for example, Associations of developers and electronics manufacturers. The head of the Association, Ivan Pokrovsky, noted that in the future, the path chosen by MCST for the development of the ecosystem could transform the entire company's activities, make MCST an IP vendor producing IP blocks of Elbrus computing cores and a compiler for them. Then other customers will be able to integrate such blocks in their products, including from other countries that do not have a problem of a sanctioned nature. It turns out that the architecture of "Elbrus" can exist despite the sanctions.
Then some other vendor will be able to produce the processors themselves, by analogy with the ARM architecture. ARM is based in the UK and licenses IP cores on which other companies produce products. For example, such as Mediatek or Qualcomm, whose processors are in every second or third smartphone in the world. Samsung and Huawei also produce their processors based on ARM cores. In our case, some companies located in sanctions-neutral countries can make their own microprocessors based on the Elbrus architecture and supply such processors to Russia…
 |
| "Elbrus-8C1". |
| Source: cnews.ru |
— That is, is there an option to close the existing problem of manufacturing Elbruses for some time due to the lack of equipment in Russia?
— Yes, to close the problem of their manufacture at the time of the creation of local production using modern technologies in Russia.
— The question immediately arises about the real import substitution in the production of microelectronics in our country. The year 2030 is often referred to as the point when we should go out to create our own equipment at least according to the topology of 28 nanometers. How realistic is this from the point of view of an expert who has been working in the IT industry for many years?
— It is difficult to judge at what point we are now, because building a microelectronic factory requires mastering a large and long chain of different processes, long cooperation and, one way or another, probably, the participation of international partners. Since we have set such a goal, the movement towards it in the face of sanctions from unfriendly countries is as non-public as possible. One can only assess from some open publications in the media what is happening. For example, in Nizhny Novgorod, a production technology has been created in laboratory conditions according to modern low-gauge standards. Russian X-ray lithographs appear somewhere and are being tested.
You can read about new experimental design works that are being ordered in various Russian research centers - in Zelenograd and other cities. We know that the production of lithographic machines has been preserved in Belarus since Soviet times, and they produce lithographs at 350 nm . Now they are working together with Russian enterprises to make this process more modern – first 180 nm, and then less.
That is, the pieces of the puzzle, they appear, and gradually there are more of them. There are probably some other projects that we don't know about. In general, the dynamics are visible, the movement towards the set goal. Most likely, the goal will be achieved, if not by 2030, then at most within plus or minus five years from this date.
In general, we have observed in various industries that it happens that the one who passes the path first moves along it for a long time. But after someone has passed first, the second one catches up with him quickly. According to the history of the development of the same Intel and AMD companies, this can be clearly traced.
— It turns out that in the current situation, the emergence of its own production is a matter of technological sovereignty, that is, Russia's national security. Promobit develops and manufactures secure data storage systems. Even if there are temporarily not enough "Elbruses", equipment is being created on processors from unfriendly countries, but with the highest possible degrees of protection, and software also plays a big role here ...
— In my opinion, the most correct way to ensure technological sovereignty is when the manufacturer uses mainly Russian solutions throughout the entire production cycle from the processor to the motherboard, and then to the system software, operating system and special software. The specialists who developed all these steps, all these levels, should be located in Russia. Only then do we have technological sovereignty from a factual point of view. And only then is it possible to check the product at different architectural levels for errors or specially left bookmarks.. For this purpose, there are relevant expert organizations and regulators that involve these expert organizations.
But the main thing is to have specialists who can do it. Then they can check everything already. After all, if a person undertakes to check what he cannot do, then the question arises, does he really understand everything that is happening? In our case, we develop software for our data storage systems on our own. Moreover, we are developing based not on some pre-existing SDS solution (software-defined storage) on the market, but on the basis of our ideas and our architecture. This is also due to the fact that if you rely only on what someone threw out in open source, you will hardly be able to overtake the market, because most often the most valuable and fresh ideas do not get into open source software immediately. First, they are patented, and the one who first thought of them earns money from them, or rather invested money in development. They are shared after some time, when the one who came up with them has already earned money. We see this in the example of mass-produced products. For example, the virtualization industry leader, the American company VMware, does not publish its main products in open source. Yes, there are open virtualization systems now, but they appeared much later, and although their popularity is growing, closed VMware still occupies the largest market share. With databases, we see how open source runs exactly the same way behind closed ones, and Oracle is still the leader in this market. Therefore, it is important to develop it yourself.
By the way, we had our own experience in open source development. There, at a certain point, we just ran into a performance limitation, which was dictated by the architecture in the original solution. That is, a "mistake in DNA," as they say. It is very sad when I have already invested a lot of resources in product development, years of work by programmers, to say: "No, we can't go on, we need to redo everything. And we do not know how to redo, because we initially went through refinement. And what the architects laid down inside, in general, we understand, in order to refine the level, and to redesign the level, we need another year to figure out this code." And that's why, with the current approach to development, we immediately came to the conclusion that we would proceed from our architecture and would not find out in a few years that we needed to do something in which we had not trained our skills at all. It is valuable for technological sovereignty to develop its own school of development. Not only the school of integration of ready-made solutions, but also the school of development.
It is necessary that the entire life cycle of products and their development cycle be controlled by Russian engineers working in Russian companies.
— In 2024, Promobit opened its retro-computer museum in Omsk . This project, as we recently said on the air of the regional television and radio company, is a link between what was done in the domestic IT sector before 1991 and the current leaders of import substitution. What does such a project give to society and the industry?
— The main idea is to really show visitors the continuity of development in our country, and above all, to interest the younger generation, future engineers and programmers. Do you remember the memes with Yuri Gagarin, where they lament that everything previously created by previous generations has been lost? We explain that a lot has been missed, but not irrevocably, the dynamics have decreased for a while, but there is a desire, love and interest in creating your own computing equipment and, in general, your own technical and design solutions.
We have our own school to rely on, there are examples and experience that can be used. At one time, there was a lot of replication, one unfortunate statement by one of the Ministers of education of the Russian Federation that creators were trained in the Soviet Union, but qualified consumers should be trained. The fallacy of this statement is now being shown to us by history itself. Our museum does not verbally pronounce this, but clearly demonstrates it. We would like to draw attention to famous Soviet engineers and programmers.
By the way, it was customary for us to talk only about foreign bright figures of people from high-tech industries, such as Steve Jobs, Bill Gates or Elon Musk, in cultural objects, in movies, in literature, but at the same time to avoid our outstanding engineers and developers of advanced technology – Sergei Lebedev, Isaac Brook, Bashir Rameev, Vsevolod Burtsev or Viktor Glushkov, the author of the first unfinished "Soviet Internet" — OGAS.
The origins of such ignorance of the industry's own technical history were laid even before the collapse of the USSR. The fact is that the country's leadership in the Soviet period was little surprised by the attention to PR, and, as a rule, in general, the principle of organizing production and the economy was such that it did not feel the need for PR. Rather, some could carry risks, because, as a rule, these specialists headed teams that worked on important and secret orders for the state, and only incidentally made some kind of products for the mass sector. As it was then called "consumer goods". And in the United States, the whole life is built on PR: if a developer loudly announced his achievements, he attracted the attention of investors, they always knew how to do it there.
The modern structure of Russia's economic structure has adopted a lot from them. Yes, we live under capitalism, in a situation where private initiatives to create teams are available and encouraged. And a private trader can go to the same private trader for money, and not to the state for investments. Of course, our private trader now needs to be able to declare himself. I think the previously mentioned erroneous trend of keeping silent about the achievements of our engineers has exhausted itself today. But once it was customary to ignore their achievements, despite the fact that they were really huge.
 |
| Retro-computer Museum in Omsk. |
| Source: tehnoomsk.ru |
— It turns out that our serious problem of the late Soviet and early post—Soviet period is the lack of information for society about what we can actually create and produce. As far as I can remember, because we are all from the same generation – we grew up in the 1980s and 90s - for a long time there was no information about our own, even in the 2000s. Because of this, it turned out that in Russia today they still write about Bill Gates or Elon Musk as innovators. We have to remember our own and educate society...
— We were taught that we can only be qualified consumers…
— We have been told for a very long time that we can only work in teams led by "cool" foreign companies. And at the same time, everyone saw that our programmers and engineers are among the best in the world. There was already such a general cognitive dissonance in the 2010s: if we are so cool, then why can't we develop everything ourselves?
— That's right. I recently thought about the fact that the "brain drain", a huge problem in the 1990s and 2000s, was a consequence of such a policy. Now, despite the fact that it is generally believed that the openness of borders is some kind of unconditional value, freedom to choose a country of residence, liberal values, human rights, and all that, ultimately, society has begun to understand that the quality of people's lives depends on how much they themselves invest their work in that environment, in where they are located: to the country and city where they live and work. The situation when it is difficult for our specialists to move abroad has become nothing more than an end to brain drain.
This is what gives hope that high-quality and promising local products will arise in Russia. Because those brains are there, and they just flowed out of the country for 20-30 years. Now they will accumulate within the country, and gradually the necessary number of qualified personnel will be accumulated.
Specialists returning to Russia from Europe and the United States these days create for themselves a quality of life here no worse than if they lived outside the country. They also improve the quality of life for others. This inspires optimistic thoughts, hopes and confidence that there will be Russian "Bill Gates" and "Ilona masks". It's just necessary that this critical mass of smart people stay here, they can build social ties among themselves, organize themselves into structures. With the assistance and assistance of the state. The state should regulate the industry in such a way as to help and support the creative initiative of the masses, and at the same time direct it to achieve its goals.
Now our state has formulated a goal — technological sovereignty. But then, probably, we will not be limited to this goal, because it is logical that the next goal will appear in the form of expansion into world markets. If we talk about computer technology, then this may be the expansion of this market itself into other areas. After achieving technological sovereignty, it is possible to set larger-scale tasks such as world leadership in a particular industry.
Evgeny Belkin was talking
