Russian DBMS market. TAdviser 2025 Review

Different segments of DBMS (general purpose, analytical, resident, graph, documentary, multimodel) show different dynamics. For example, in the world, graph DBMSs are the fastest growing segment with CAGR above 26%, according to a Gartner report^[3]

Ekaterina Kanunnikova, Product Director of Data Services at VK Tech, noted the trend of blurring the boundaries between different models of database organization. The demand for the transactional and analytical type is growing, companies want to simultaneously actively record data and immediately conduct analytics. Analytical processing functions in DBMS are important everywhere, even for [1]Highload systems with an explicit OLTP profile.

According to Konstantin Osipov, Managing Director of the Arenadata Group for Research and Development, the trend towards specialization, which emerged more than 15 years ago, remains. However, in total, all specialized DBMSs still cannot approach the two main classic segments - OLTP and OLAP.

According to the expert, the observed high growth of graph DBMS is due primarily to the low starting base. It is important that the functionality for working with graphs is gradually becoming part of universal DBMS - similar features have already been implemented in PostgreSQL through extensions, and are also present in solutions such as Cassandra, Picodata and Tarantool.

Specialized graph systems face challenges in building a sustainable business model due to their relatively narrow application niche. This is especially noticeable in the Russian market, where an independent segment of graph DBMS has not yet developed, and the main activity boils down to replacing departed Western products.

As Alexey Bekhtin, an analyst at DBMS RED Database of RED SOFT, said, there is a trend towards adding vector functions to general purpose DBMS, which allows you to use standard SQL queries to work with vector data, which greatly simplifies the integration of such functions into existing and familiar systems already certified by the FSTEC of Russia.

Definitely, in the selected solutions they will work more efficiently, but the introduction of vector functions in general-purpose DBMS opens up new opportunities for manipulating data, the expert said.

Analytical DBMS and general purpose DBMS are in demand in many industries because they solve the most common business problems. This creates a large market for boxed solutions, as enterprises are looking for ready-made products that can be quickly implemented without the long cost of development and configuration.

According to Ivan Panchenko, Deputy General Director of Postgres Professional, general-purpose DBMSs remain relevant. The expert referred to an article by Michael Stonebraker and Andy Pavlo: over the 20 years of DBMS development, there have been many attempts to create specialized systems, but later there was always a rollback to universal ones, whose capabilities are expanding. Russia has the same patterns, in addition, there is incomplete satisfaction with demand for domestic solutions, which is a more significant incentive for market development than global trends. The segment associated with the management of data for which AI is used is expected to grow. The demand for DBMS for solving analytical problems has existed for a long time and remains high. There are still few solutions of this class on the market that meet all the needs.

Leonid Savchenkov, CPO of the Yandex Cloud Data Platform, believes that general-purpose DBMS will continue to grow steadily, they remain the base for most business applications. The main development vectors are scalability, improved monitoring tools and improved security. The analytical DBMS will continue to strengthen the trend towards storage/compute separation, which allows you to more effectively manage resources when working with large amounts of data.

Yuri Osipov, Jatoba product manager for Gazinformservice, recalled that general purpose DBMSs and multi-model databases are different aspects of the world of data management that can be linked but are not interchangeable.

Relational databases (domestic forks of PostgreSQL) are suitable for cases related to transactions, while graph databases are needed for applications with a large number of relationships and data. The growth of the market for DBMS segments Yuri Osipov directly connects with the growth of organizations using DBMS in these segments.

The expert noted the trend in the growth of functionality to the level of Oracle Database. The extensibility of PostgreSQL allows you to support various types and models of data through extensions that can actually turn any domestic fork of PostgreSQL into a platform for building multi-model solutions.

SberTekha CEO Maxim Tyatyushev noted that the database management systems market will experience systemic evolution on the horizon of the next 10 years. Traditional relational solutions, primarily based on PostgreSQL, in his opinion, retain their significance as a universal standard. However, the growth vector is shifting towards distributed and fault-tolerant DBMSs in demand in global infrastructures and digital products with high SLA criteria.

{{quote "The appearance of vector and semantic DBMS reflects the transition to cognitive architectures, where AI solutions require searching not by key, but by meaning. In parallel, streaming and in-memory platforms are rapidly developing, providing real-time data processing and analytics.
The key transformation is the transition to ready-made business solutions. DBMS itself becomes a product from which Chief information officer and architects expect orchestration, security, monitoring, integration with AI, hybrid analytics (HTAP). Not specific technologies are in demand, but architectures that combine them, "Tyatyushev explained. }}

The trend on AI, in particular on generative, influenced all categories of DBMS, experts say. There are solutions that simplify the work of data engineers and analysts - automate the compilation of requests, building reports and working with data in natural language. AI is gradually being introduced into the processes of DBMS operation: it helps in problem diagnostics, predictive maintenance and resource optimization.

AI can significantly improve DBMS security by automating and optimizing access control, anomaly and intrusion detection, vulnerability and log file analysis, data leakage prevention (DLP).

AI builds SQL queries and starts processing their results, creates its own data schemes for solving specific problems, fills them, integrates them with sources. Agent systems gradually displace conservative assistants limited by predefined algorithms. This will lead to the replacement of all large parts of business processes with "black boxes" with AI agents.

Probably, the most conservative will remain the analytical segment and the classic OLTP, since they rely on business processes that require clear SLA, predictable load on hardware resources. The fact is that AI agents are good in ad-hoc tasks, but not in business processes with strict regulations: they are more creative and faster than people, but not very stable in terms of the amount of resources spent on solving the problem. Serverless DBMS will undergo stronger transformations, lakehouse - everything that was "sharpened" under Kubernetes.

AI is developing so fast that the conservative DBMS industry simply cannot keep up with it. Over time, the situation will change, and this request will cause a delayed growth of this segment, - said Ivan Panchenko, Deputy General Director of Postgres Professional.

According to the expert, the specialization of DBMS in this direction is only being formed, although there are already well-known and proven approaches. At the same time, many market participants are now working to include AI functionality. Postgres Professional contains AQO, a ready-made solution for optimizing query execution using ML technologies.

Alexey Bekhtin, analyst at DBMS RED Database of RED SOFT, noted that for general relational DBMSs, the AI trend rather puts forward new requirements for graphical tools for database administration and monitoring.

Administrators and database developers are increasingly talking about the need for AI-based intelligent assistants, which will offer help in writing new requests, explaining existing ones, finding bottlenecks in optimization, the expert explained.

Alexey Bekhtin mentioned the preparation of the release of the RED Monitor product, which will take place soon. This tool for monitoring the status of servers and databases should solve the most common problems in database administration, for example, relieving the load on administrators, leveling the lack of performance and minimizing database damage.

Maxim Tyatyushev said that Sber Tech is working on the use of AI as an assistant database administrator in the Platform V Kintsugi DBMS management console.

The AI assistant helps to configure the DBMS, detects anomalies in work and offers solutions to eliminate them, provides quick access to the necessary information in the knowledge base. Thus, we simplify the threshold for entering database management for faster hiring and adaptation of Junior specialists, - said the CEO of Sber Tech.

As Oleg Hyatsintov said, DIS Group focused on the development of the IDP intelligent data platform, which plans functionality with AI to quickly obtain missing information, ensure quality verification and data delivery, and build an visualization process. The data set should be provided as quickly as possible, taking into account the knowledge of the location of the data, the context of the request, the position of the user in the organizational structure, and the logic of the selection.

Barriers to AI

For a number of historical and economic factors, all agent frameworks currently have a number of critical shortcomings that make their industrial use difficult in complex landscapes, said Anton Balagaev, director of AI and expert solutions at Arenadata. As the expert explained, all significant frameworks offer agents to give each other instructions on their own, and not through the data bus. Hence the complexity of refactoring, duplication of logic, difficulty in impact analysis. Also, this format implies that agents themselves determine to whom to transfer the task, even if the models underlying them are poorly adapted for this. The security and control of the rights of agents to transfer the task to other agents can be compromised.

Before performing a difficult task, the AI agent is forced to plan its implementation. In most frameworks, such a plan is transmitted by text from one agent to another with the entire message history, regardless of whether an individual performer is able to convey it correctly and not distort it (this is especially bad for highly specialized models, for example, focused on coding). A very short time after the start of adapting agent frameworks to the Enterprise sector, such plans will be replaced by DAGs, originally presented as code. Anton Balagaev believes that this approach will allow target agents to perform individual tasks without atomizing to think about the entire chain of tasks, will make it possible to regulate the format of entry and exit, build strictly executed dependencies of individual goals and perform tasks in parallel.

Models lying under AI agents are obliged to work under conditions of sufficiently strict restrictions on the structure of their response to achieve the ultimate goal. Not all models can do this. The developers of popular agent frameworks mask the problem with an uncomplicated algorithm: if the model could not answer, the task is restarted. You can set the number of attempts, but this violates the predictability of the execution time and does not guarantee success. The technical community from the world of Open Source has already proposed a solution - constrained decoding. The developers of model execution environments (vLLM, SGLang) have implemented the ability to determine the model response scheme through JSON or Pydantic classes: constrained decoding technologies force the model to follow such a scheme on the first attempt. However, no popular agent framework still supports full-fledged constrained decoding without manually modifying the model call component.

In the near future, frameworks are expected to be available that can provide a level of reliability, transparency and security that is now unavailable when developing on popular frameworks such as Langgraph, Microsoft Autogen, Google Agent Development Kit, OpenAI Agents SDK, Crew AI, Atom, Pydantic AI.

In the future of two years, the communication of agents can be rewritten to pub/sub architecture, and their long-term memory will become an integral part of DWH.

According to Yuri Osipov, Jatoba product manager at Gazinformservice, customers are primarily concerned about the achievement of ORACLE Database functionality.

Ivan Panchenko, Deputy General Director of Postgres Professional, also noted that until now, customers did not really expect any breakthrough technologies from Russian DBMS. They needed the capabilities of foreign solutions that they had previously used to become available in Russian ones.

Alexey Bekhtin, analyst at DBMS RED Database of RED SOFT, shared his observation that customers are waiting for an increase in the performance of domestic solutions, a trend towards fault tolerance and high availability, as well as monitoring, automation of maintenance and deployment, is increasingly visible.

Ekaterina Kanunnikova, Product Director of Data Services at VK Tech, said that there are no radically new technological trends compared to the horizon of 1-2 years, old trends are still in effect. Thus, hybrid loads and operating modes are promising. We are talking about HTAP and the ability to perform real-time analytics with simultaneous high read and write speeds.

Of interest is MultiTiering, that is, the ability to store and move data between storage layers, for example, from RAM to disk and vice versa.

The expert noted the need to increase the security of the DBMS: traffic encryption, different methods of authentication, integration with internal client systems related to security. FSTEC certification for DBMS has become must-have for critical business systems.

Another trend is the movement towards the convergence of transactional and analytical loads on a single platform. The trigger is exponential growth in data volume, tightening requirements for real-time analytics and a sharp increase in the cost of storage.

At the same time, the Data Lakehouse concept becomes the optimal solution, implementing ACID transactions on top of economical object storage, providing vector processing of requests and a single interface for working with batch and stream data, - said Ekaterina Kanunnikova.

According to her, customers expect high performance, reliable storage and processing, a high level of fault tolerance, convenient tools for integrating with DBMS from their systems and convenient deployment and management of large clusters.

Leonid Savchenkov, CPO Yandex Cloud Data Platforms mentioned the presence of three trends. Support for several data storage models (relational, documentary, graph, key-value, etc.) within one system is relevant. A noticeable trend is the emergence of systems that can work "at the junction" of the cloud and On-Anticipate, as well as according to the principles of Multi-Region, Multi-Cloud. Another vector of DBMS development is automatic scaling and tuning of indices and parameters (autotuning), self-healing, automatic backup/restore, including under AI control.

Dmitry Lemesh, product manager at Nexign, noted a trend towards creating a conditional "ecosystem" around DBMS. Previously, various third-party tools such as pgAdmin, DBeaver, Percona were used to monitor and write requests, and now all of them must be in a single graphical interface that covers all the needs of the database administrator. An AI assistant is built into such an interface, which helps to correctly configure the database and gives advice on optimization. All this reduces the burden on the administrator and increases efficiency.

Oleg Hyatsintov, CTO of DIS Group mentioned the trend towards creating solutions of the Data Lakehouse class. Over the past year alone, at least five solutions of this class have appeared in Russia.

As Najim Mohammad, director of data strategy at Arenadata, said, the requirements for DBMS are increasingly shifting towards cloud architecture, to the possibilities of use in distributed environments. Modern DBMSs must support containerization, dynamic resource allocation, and have the ability to adapt to the load in real time.

Large companies are moving to federated data management models. Supporting an infrastructure with hundreds of installations, storefronts, instances requires scalability and stability, which stimulates the development of more flexible DBMS: as unpretentious as possible, less demanding of hardware.

According to Konstantin Osipov, Managing Director of the Arenadata Group for Research and Development, customers want, first of all, to increase the elasticity of data storage costs so that costs meet the real needs of the business. This request is satisfied through various approaches, including the transfer of data to S3 object stores, the separation of computing resources and data stores, and the introduction of multi-cloud architectures. Special attention is paid to optimizing the processing of unstructured data, which becomes an important competitive advantage of platform solutions that offer integrated tools for working with all types of data.

According to the expert, the most interesting in the future will be unformed niches: NoSQL, specialized databases, solutions for processing streaming and semi-structured data. Modern architectures rely on these technologies, and here the Russian market has yet to form.

According to the Strategy Partners report, the leading software deployment format in the market in 2023 was On-Anticipate, with a ~ 92% share. The strong demand for non-cloud solutions has been driven by business and government commitment to strict security standards, mandatory regulatory compliance, and the need to protect sensitive data. According to the same report, the B2B SMB segment accounted for only 7% of demand, which is associated with the wide practice of using free data management solutions.

According to Yuri Jatoba Osipov, the company's product manager, Gazinformservice Private Cloud is in demand and gaining growth, while Public Cloud is promising only for SMB.

While the On-Premise format remains leading in 2025, large companies are developing their own private clouds. This was announced by Ekaterina Kanunnikova, Product Director of Data Services at VK Tech.

Do not forget the public sector with the GEOP infrastructure (State Unified Cloud Platform). Therefore, the requirement for the DBMS to be Cloud Native is still relevant and promising, the expert emphasized.

According to her, Cloud Native - the ability to work in cloud environments - continues to be a relevant long-term trend, which can also include containerization, including support for Kubernetes.

Anton Blizgarev, Director of Cloud Business Development at Arenadata Group, confirmed that the share of On-Premise deployments remains dominant. A hybrid approach is often used: the productive part is hosted by On-Premise, and development and testing go to the cloud. At the same time, there are more and more solutions that are initially focused on working in flexible computing environments running Kubernetes with a separate S3 storage layer from computing. The expert believes that the trend towards a slow steady growth in the share of cloud-oriented DBMS is obvious. Clouds are becoming more and more productive, the previous restrictions are disappearing.

With the growing complexity of tasks, the increase in data volumes and the transition to more mature approaches to the IT infrastructure, the share of the cloud in the DBMS segment will grow, - Anton Blizgarev is sure.

According to Oleg Hyatsintov, CTO of DIS Group, Russian cloud services mainly work on the IaaS system. Western-type cloud services, which have all the solutions customers need at once, are not yet in Russia. Against the background of attacks on the Russian IT infrastructure, customers are concerned about the security issue. There is a temptation to make everything completely closed, but such products will not be able to develop, so we need to look for a hybrid approach. The development of Russian cloud capacities is promising for the prosperity of the country's economy as a whole.

The development of cloud services with multi-tenant solutions, first of all, is necessary to provide modern solutions to medium-sized businesses that need data, analytics and reporting no less than large organizations, but cannot afford to purchase expensive data platforms, "said Oleg Hyatsintov.

According to Leonid Savchenkov, CPO of the Yandex Cloud Data Platform, the segment of cloud-oriented and server-free DBMS in Russia remains quite promising. The benefits that the cloud model provides are especially needed in scenarios with high requirements for flexibility and speed of change.

We are recording a steady increase in interest in cloud DBMS: for example, the use of PostgreSQL in a managed format increased by 44% in 2024 compared to 2023, he said.

In 2025, Yandex B2B Tech opened business access to its own development platform for storing and processing big data YTsaurus. With its help, you can analyze data exabytes in companies and train complex ML models with billions of parameters. YTsaurus is available in two delivery formats: in the cloud and On-Anticipate. In the cloud, the solution is available as a managed service, that is, Yandex is fully engaged in platform support.

To grow the direction of cloud DBMS, it is first of all necessary to supplement application stores with application software, - said Alexey Bekhtin, analyst at DBMS RED Database of RED SOFT.

According to the expert, the use of DBMS in the cloud provides many advantages. The client can put the necessary application software in the cloud, buy a compatible DBMS to it, add reporting software if necessary, that is, collect all the necessary infrastructure. And for scaling, you need to buy power or servers.

External, not always directly technological, factors have an increasingly noticeable influence on the DBMS market in the Russian Federation.

Ecosystems are developing, marketers are fighting to increase the average check and are striving for maximum personalization of offers, and the state requires the protection of personal data. These and other trends are pushing up not only the DBMS market, but also the markets for encryption, information security, and data management.

Internal regulation by state institutions in the field of information processing, storage and protection is strengthened. We are talking primarily about data subject to the regime of critical information infrastructure (CII), as well as personal data and industry specifics (for example, in the financial or medical sector). Against the background of the growing number of cyber attacks, increased attention to information security forms a steady demand for FSTEC certification.

On September 1, 2025, federal law of 07.04.2025 No. 58-FZ "On Amendments to the Federal Law" On the Security of Critical Information Infrastructure of the Russian Federation "comes into force. State regulation in terms of the law on CII affects private organizations.

The high key rate affects all industries, the cost of money remains high. More attention is paid to ROI, the ability to reuse systems and data across processes. The business strives to manage the available resources as efficiently as possible.

Experts note extreme data growth and increased cost of server equipment, which leads to the need to optimize the infrastructure. Iron is growing in price steadily and regardless of fluctuations in other markets, storage, computing and support costs are increasing.

The issue of lack of qualified personnel is acute. For example, it is difficult to search for and hire engineers in such a complex and specific area as system programming. There is a shortage of database administrators, which slows down the introduction of new Russian products.

To implement large-scale tasks, it is not graduates of express courses in applied programming who are needed, but specialists with high-quality fundamental education and structural knowledge. This opinion was shared by Ivan Panchenko, Deputy General Director of Postgres Professional. This company develops courses for database administrators and developers, cooperates with universities, conducts internships for students and summer educational camps for schoolchildren, prepares for the publication of a book on PostgreSQL.

The import substitution driver with red thread passes through the topic of the formation of the Russian IT market from 2022. In May 2024, TAdviser interviewed the IT leaders of large Russian organizations about their experience in replacing Western DBMS with products from domestic developers. The results of the study were presented by TAdviser editor-in-chief Alexander Levashov on May 24 at the conference "Digital Industry of Industrial Russia" (TSIPR). The respondents called the key difficulties in switching to domestic DBMSs a large amount of refinements of applied systems, a shortage of competencies, personnel and resources for tasks not directly related to business needs, as well as an insufficient level of maturity of domestic DBMSs, which experts unanimously assess "on a C."

According to the conclusions of the study, the key difficulties/barriers to import substitution of DBMS, according to end customers, are factors: 1) a large amount of refinement of application systems; 2) lack of competencies, personnel, resources; 3) the shortage of maturity of domestic DBMS, development on Open Source.

As Ivan Panchenko said, Postgres Professional faced situations when the developers of some Russian DBMS appropriated the authorship of some of the functions created by the company, the code of which was in the public domain. The expert noted that similar phenomena are observed in the Russian segment of operating systems.

The approach taken now, when compound works are put on the same board with their own developments, is harmful to technological sovereignty, as it erases significant differences between them, says Ivan Panchenko.

According to him, Open Source is an excellent tool for community consolidation and cooperation at the level of industry consortia, without it the IT industry would not have such progress today. It's just that the attitude towards it should not be consumer, you need to invest in its development.

As for the first point from the conclusions of the TAdviser study, the expert noted that the large amount of necessary modifications of application systems is the basic difficulty in migrating even to a foreign DBMS.

For example, migrating from Oracle to Microsoft SQL Server is much more difficult to implement than from Oracle to Postgres Pro, he said.

Among the reasons for the difficulties are the initial unavailability of systems for processing, the incompleteness of internal design and technical documentation, the lack of full test coverage of application systems, without which migration verification is practically impossible.

The expert pointed to a significant, from his point of view, factor that affects the progress of import substitution, namely, the lack of an understandable phased state plan.

In our opinion, the market now has enough clear state strategy and a detailed import substitution plan. There are very strict requirements for the transition to Russian software, but there is not enough understandable phased plan leading to the fulfillment of these requirements - everyone must solve this problem himself, - Ivan Panchenko shared.

Leonid Savchenkov, CPO of the Yandex Cloud Data Platform agrees with the first two main conclusions of the TAdviser study, but believes that the assessment of the maturity of domestic and open source DBMS may be more balanced. The fact is that in terms of basic functionality they are not inferior to commercial Western products. The problem, as a rule, is not the "technical weakness" of the solutions themselves, but their less integration into the corporate IT landscape.

Previously, Western vendors offered holistic ecosystems in which DBMS was closely related to analytics, backup, security management tools, etc. When switching to new platforms, many companies have to build these bundles again, which requires a mature IT infrastructure and good expertise.

According to the expert, one of the key challenges that remains to be overcome is increasing the maturity and stability of the entire ecosystem of IT solutions, especially in terms of domestic Open Source.

The market lacks deep cooperation between participants in the development and development of open technologies. At the same time, the potential of such collaborations is enormous: they allow not only to accelerate technological progress, but also to build uniform standards, reduce costs and scale best practices.

Unfortunately, now we often observe a situation where, instead of joint development, some players are engaged only in repacking existing Open Source products under their own brand, practically without investing in the development of the technologies themselves. This slows down overall progress, interferes with the accumulation of expertise and makes it difficult to form a mature domestic market, the expert noted.

When choosing an Open Source DBMS, it is important to look not only at the brand, but also at the activity and scale of the community, project history, release dynamics and the presence of real implementations.

For example, in 2024 Yandex launched the distributed fault-tolerant Distributed SQL open source database YDB. It combines high availability and scalability with strict consistency and ACID transactions, supports simultaneous execution of transactional (OLTP), analytical (OLAP) and flow loads.

According to Leonid Savchenkov, one of the key problems that market participants face is working with large amounts of data in corporate storage, especially in conditions of limitations on the performance and cost of infrastructure.

When it comes to tens and hundreds of terabytes, standard approaches to data storage and processing cease to be effective - both in terms of the speed of query execution and in terms of economic feasibility, the expert emphasized.

Alexey Bekhtin, analyst at DBMS RED Database of RED SOFT, agrees with the main conclusions of the TAdviser study. He said that sometimes application software developers have an unreasonable fear of changes. Meanwhile, many tasks during the transition can be solved in automatic mode, especially if the software does not use the specific capabilities of individual Western DBMS, for example: stored procedures, specific data types, non-standard SQL syntax, capabilities around DBMS (failover cluster, administration tools, etc.). As for the upcoming challenges, the expert noted that first of all, functional improvements of Russian DBMS are planned.

There is a lot of work to be done to catch up with such Western market players as Oracle Database and MS SQL Server, but we are moving in this direction, studying the needs of customers and partners and relying primarily on them to close these needs as much as possible, the expert said.

According to Alexei Bekhtin, for the development of Russian solutions, it is necessary, first of all, to build an open dialogue between DBMS vendors, application software developers and customers.

When developers and customers start talking directly about their needs and problems, DBMS vendors will be able to make their improvements more focused and will surely know what improvements are needed in the market, "said Alexey Bekhtin.

Also, according to the expert, the use of foreign DBMS can become a problem for business due to state regulation, and the phenomenon of open source profanation can negatively affect the reputation of Open Source. However, key market players have many of their own improvements and a transparent development system, so it is not difficult to distinguish a clone from a high-quality development.

A large number of DBMSs are now based on Open Source development, but honest companies make a great contribution to the original DBMS. In addition, state regulation now plays a very large role, since certification requires high expertise in the development and elimination of vulnerabilities, "said Alexey Bekhtin.

A significant amount of improvements in applied systems remains one of the most complex technological barriers in import substitution, said Sergey Zolotarev, Director of Strategic Development of Arenadata Group. As the expert said, it is most often impossible to simply replace one product with another - a complete processing of architectural solutions is required. In particular, instead of one foreign product, several Russian analogues have to be introduced, which significantly complicates the process.

Russians have successes in the segment of analytical DBMS, while the situation with billing systems and ERP is more complicated: the performance of transactional PostgreSQL has not yet reached the Oracle level for high-load applications. In this case, the monolithic systems are decomposed into separate components with each transferred to a specialized DBMS.

There are cases where part of the load of high-load systems originally built on Exadata was successfully transferred to in-memory DBMSs such as Picodata. But this requires a complete redesign of the architecture of the application system.

Some companies in different industries still do not have a strategy for switching to Russian software, risking a situation where an accident or incompatibility of components will require emergency migration without proper testing and preparation.

Some companies postpone the transition, some do not plan it at all - the business simply does not see tangible benefits for itself in the short term. In the issue of migration to domestic DBMS, business is guided primarily by practical considerations: risks, budget, available resources and priorities. Leonid Savchenkov, CPO Data Platforms Yandex Cloud

The main trend based on the research carried out can be distinguished by the preservation of dependence on foreign decisions, despite the political and economic need... Methods of circumventing the sanctions regime can lead to the habit of using established foreign solutions, since they can be easier to support and integrate at the moment. Alexey Bekhtin, analyst of DBMS RED Database of RED SOFT

According to Alexei Bekhtin, based on the growing distrust of new local developments, some companies may have the opinion that the domestic quality alternative does not reach foreign counterparts, and a successful strategy may be to work to increase confidence in Russian products and popularize real success stories of their use.

As Vadim Yatsenko, General Director of Tantor Labs (part of Astra Group), said, one of the most promising areas is database machines (MBD) - specialized products created for the growing needs of the business. They automatically allocate resources and process tens of thousands of operations per second through deep hardware and software integration. Easy management and maintenance allows you to focus on goals rather than infrastructure issues.

Various approaches to the creation of MBD are presented on the Russian market: from a formal combination of software and servers to a deep engineering study that guarantees stability in critical scenarios.

Import substitution of DBMS is not just a replacement of one product with another, but a strategic transformation of the IT infrastructure, requiring a balanced approach, - Vadim Yatsenko is sure. - Database machines are becoming one of the key tools of this transition, combining readiness to work here and now with the prospects for further development.

Yuri Osipov, Jatoba product manager at Gazinformservice, agrees with the findings of the TAdviser study. In particular, with the fact that import substitution slows down the immaturity of domestic business decisions. The expert noted the difficulties that market participants face when simultaneously replacing many application systems, as well as the shortage of budgets, resources, personnel. According to him, there is currently no assistance from the state, it should be expressed in co-financing customers, while the attention of the Russian community of developers should be drawn to the need for cooperation.

Profanation of open source due to an increase in the number of DBMSs from "developers" of "clones," as some experts say, as a phenomenon, is no longer deepening, said Yuri Osipov: a pool of leaders has formed, and some projects will even be closed (for example, a deal to sell Proxima DB).

It is no longer profitable to create new forks, since market leaders cannot be in time, "summed up Yuri Osipov.

State support

In 2023, software manufacturers expected the state to consistently follow the announced course, support client companies introducing Russian software products, as well as further steps to stimulate demand for domestic software and restrict access to foreign software.

According to Leonid Savchenkov, among the effective measures of state support, tax benefits can be distinguished - they allowed companies to redistribute more funds for development and technological improvement (some of the benefits have already lost their force). IT mortgages have also proven themselves positively, which help to retain qualified specialists.

At the same time, direct state support for specific solutions or companies is not always unambiguous. It can create reputational or technological risks for the recipient, especially given that the IT market is inherently global. Many domestic developments depend on foreign technologies - whether they are hardware components, operating systems or widely used software libraries. Therefore, a one-sided focus on "local" solutions without regard to this interdependence can limit the sustainability and flexibility of the industry.

To implement the import substitution program and gain digital sovereignty by customers, it is not enough to force them to "take what they give," said Alexey Zhuravlev, Deputy General Director of the Arenadata Group.

We feel powerful support from the state, which is systemic. For example, recently, to analyze dependence on foreign software, assess the presence and maturity of Russian analogue solutions, Competence Centers were formed: industrial competence centers (ICCs), which include business customers of IT solutions, and Competence Centers for Software Development (CDC), the expert noted.

The ICC independently audits IT solutions to ensure that the declared functionality is available. This allows customers to see the appropriate options and the results of their tests, and developers to understand what missing functionality should be implemented to close the needs of customers.

Another initiative of support from the state is especially significant projects (OZP). Recently, the first wave projects were completed and the WBS 2.0 projects were selected. A number of projects that have received the status of OZP are implemented with the help of subsidies.

The efforts of the state aimed at supporting manufacturers when entering foreign markets are extremely important, because sufficient sales directions are needed to create competitive products.

The Ministry of Digital Development organizes periodic meetings with developers, at which the discussion of both the results of the work of support measures and new proposals are announced.

Ivan Panchenko, Deputy General Director of Postgres Professional, noted that, as such, there is no separate support for the DBMS market. As for the appearance of RFRIT and grants, the expert does not yet see (at least in the direction of DBMS) that this will lead to the emergence of some big new results. Another thing is that consumers are supported through financing projects for the transition to domestic software, which is much more efficient.

Today, the Russian software industry receives a large amount of state support, but the problem is its indiscriminate nature: now all developers have the same preferences without taking into account the importance of their product to solve the problems of import substitution and technological sovereignty. This, among other things, leads to the emergence of pseudo-products issued as full-fledged commercial software, - said the expert.

Ivan Panchenko emphasized the recent successful project, the first of its kind for the Federal Treasury of the Russian Federation, on the migration of one of the key state information systems - GIS GMP - to a distributed DBMS with horizontal scaling Postgres Pro Shardman. This GIS records and processes data on payments of all citizens to the budget, from fines for traffic violations to tax charges.

The 180 Terabyte base was seamlessly transferred to our DBMS. For comparison, the most common on the market are base migrations that do not exceed 5-10 TB, the expert noted.

Another, no less important, case for transferring a system critical for the functioning of any modern bank - ABS - from Oracle to the Postgres Pro Enterprise DBMS was implemented with Gazprombank. In terms of importance, criticality and responsibility of migrations in the domestic market, this was one of the most noticeable, noted Ivan Panchenko: banking ABSs work with huge amounts of data from individuals and do not even allow short-term downtime.

This successful experience has attracted great attention from the financial sector, and we are already noting an increase in interest in our decision from other banks, the expert shared.

Each company has its own technology stack and it is extremely important that the DBMS is compatible with the software in this stack. Imported systems, in particular, have always been well compatible with popular software and operating systems. This opinion was shared by Dmitry Lemesh, product manager at Nexign.

According to the expert, with the beginning of the period of active import substitution, Russian companies were faced with the fact that domestic software was not always compatible with each other. But confirmation of compatibility is one of the key aspects when choosing a product by customers.

One of the indicators of market maturity and solutions is considered compatibility Oleg Hyatsintov, CTO of DIS Group. According to the expert, compatibility for business is a signal that improvements in implementation or migration will require less, that functionality will not suffer, the speed of implementation will be higher, and the cost will be lower. At the same time, the importance of compatibility of IT solutions is spoken even in the government of the Russian Federation.

If the DBMS is not compatible with business applications, users will not be able to fully solve their problems, and this is critical for them. Therefore, on the one hand, customers are striving to move from the legacy stack and foreign systems to new solutions, on the other hand, these new solutions do not always provide the necessary functionality. Oleg Hyacintov, CTO of DIS Group

According to Sergey Zolotarev, Director for Strategic Development of the Arenadata Group, among the unresolved challenges is still the problem of integration between various software classes - the industry has not reached the required level of compatibility. Unlike Western solutions, where the interaction of products has been debugged for decades, Russian vendors paid insufficient attention to this aspect.

Individual companies, including Arenadata, are actively working in this direction - for example, conducting integration testing with components of BI systems, which is an important factor in the development of products, the expert noted.

Alexey Bekhtin, analyst at DBMS RED Database of RED SOFT, noted that consulting services are becoming more and more popular in the DBMS market every year, companies are increasingly considering migration from foreign products or old solutions. But often big solutions contain around them a large ecosystem of various services and products, and it becomes more and more difficult for the end client to solve problems that arise when creating or updating infrastructure. The expert noted that consulting and outsourcing can be useful before building up companies' own competencies, while investing in their employees should be preferred.

Clients pay special attention to the presence of proven experience in working with import-substituted technological stacks, said Dmitry Stolov, head of the directorate of complex services. Arenadata According to the expert, the landscape of consulting services has changed dramatically: the company's routine tasks are solved by inhouse teams, and specialists with unique expertise are looking for complex highly specialized cases. Clients began to present fundamentally new requirements: now they need not just consulting support, but deep specialization in solving atypical problems. Simple tasks are gradually "washed out" of demand, giving way to more complex and complex ones.

As a key trend, Dmitry Stolov noted the request for a deep immersion in the problems and work with modern technological platforms. Special attention is paid to maintenance during the operational stage of solutions, optimization of total cost of ownership. All this requires DBMS vendors to revise approaches to the formation of a service offer and the development of truly niche competencies.

Leonid Savchenkov, CPO of the Yandex Cloud Data Platform, noted that the demand for outsourcing and database administration services remains in the DBMS market in Russia, but is limited. Many companies prefer to keep key IT functions within the organization, especially with security and control requirements in mind. At the same time, cloud and managed solutions, as a form of partial outsourcing, are used in a number of scenarios where flexibility and speed of implementation are important.

Outsourcing and administration services are in demand for critical systems with proactive technical support, said Yuri Osipov, Jatoba product manager at Gazinformservice. As an example, the expert cited the company "T1 Integration."

• Database Management Systems
• Russian DBMS market: achievements, drivers and prospects. TAdviser 2024 review
• Experience and prospects for the use of Russian DBMS in large organizations: results of the TAdviser study
• Softline and TAdviser conducted a study on Russian DBMS 2023