Russia has successfully implemented a major database migration project, replacing Oracle Exadata with the domestically developed Postgres Pro Enterprise Certified. This project involved the migration of nine critical subsystems of the Federal Tax Service’s automated information system, Nalog-3. The initiative, implemented between 2022 and 2026, includes approximately 45% of the system’s transactional data, exceeding two petabytes. The transition is a major milestone in Moscow’s initiatives to enhance control over critical government IT infrastructure and decrease reliance on foreign enterprise software.
A Critical Government System Moves Away from Oracle
Nalog-3 is used by the Federal Tax Service (FTS) to automate a wide variety of tax administration functions, such as the processing and analysis of information, the maintenance of tax records, and the provision of data to authorized external users. Nalog-3 is a comprehensive information system that facilitates the agency’s operations throughout the nation.
The underlying database infrastructure is strategically significant for a system that manages a vast amount of information regarding individuals, businesses, and tax compliance. Historically, Nalog-3 has relied on Oracle Exadata, an integrated hardware and software platform specifically designed to handle highly demanding database workloads. Although the platform facilitated the system’s operations, the concentration of numerous subsystems around a centralized database also resulted in technological dependencies.
Consequently, the migration was not merely a software replacement exercise. This entailed the migration of critical operations to a different database platform while maintaining the performance requirements of a nationwide tax administration system, data integrity, and operational continuity.
In 2022, the Federal Tax Service collaborated with JSC GNIVC, the agency’s Main Scientific Research Computing Center, to identify priority subsystems for migration. In addition to the data that the tax authority’s control functions used, the selected components included information about legal entities and individuals.
Nine subsystems and over two petabytes of data
The project’s scope illustrates the technical requirements for replacing an established enterprise database in a large government environment. Approximately 45% of Nalog-3’s transactional data is comprised of the nine migrated subsystems.
The new environment’s individual databases have expanded to as much as 60 terabytes, while the overall system contains over two petabytes of transactional information. The migration of databases is far more complex and detailed than the transfer of ordinary business applications or relatively small departmental databases due to the sheer volume of data.
Postgres Professional, a Russian corporation, has developed the Postgres Pro Enterprise Certified replacement platform. The certified edition is intended to satisfy the relevant information-security certification requirements, rendering it pertinent to organizations that are subject to regulatory restrictions.
The decision is indicative of a broader Russian strategy to replace foreign enterprise software with domestic alternatives in government agencies and other organizations that are responsible for critical information infrastructure. Nevertheless, the project’s importance stems not only from the software’s nationality but also from its proven capacity to manage substantial production duties.
Three-Node Architecture Designed for Reliability
The architecture of the replacement database environment is one of the most critical modifications. The new system is deployed on hardware-software complexes in a fault-tolerant configuration with three nodes. Although independent reliability test results and detailed failover times have not been publicly disclosed, this design assists in ensuring service availability in the event of an individual component’s failure.
The initiative also decreased the concentration of workloads in a single, centralized database. Redistributing some processing across multiple separate, fault-tolerant clusters reduced the degree to which multiple critical functions relied on a single database environment.
This architectural modification is crucial because the mere act of replacing a database product does not automatically remove all operational risks. Resilience is also influenced by the organization of replication, recovery, storage, and applications. The project addressed the concentration of database workloads and software dependency by integrating the migration with a distributed cluster arrangement.
Migrating 40–50 TB in Approximately One Day
Another major challenge was the relocation of large databases without disrupting critical government operations. The project team conducted multiple tests of the transfer process in preparation environments prior to the production migration.
The data volumes of approximately 40–50 terabytes could be transmitted in approximately 24 hours, as documented by these trials. As a result of these tests, engineers were able to establish an appropriate maintenance window for the production transition.
The final migration required slightly more than a day. The procedure involved the transmission of data, the configuration of the corresponding application software, and the execution of backups. The functional and operational characteristics of the migrated subsystems were preserved, as indicated in the project announcement.
The relatively brief production window substantially impacts the information system that facilitates nationwide tax administration. A migration of this magnitude necessitates more than the mere transfer of database files; applications must continue to interpret the information accurately, and the replacement environment must accommodate the operational needs of existing services.
The results that were reported suggest that the transition was successfully completed without any reported loss of functionality in the migrated subsystems. Nevertheless, the migration’s internal procedures, validation methods, and recovery arrangements have not been publicly disclosed in their entirety.
Handling 3–4 TB of Additional Data Every Month
The continuous growth of data is also being used to test the new database environment. As per the project information, the migrated environment processes an additional three to four terabytes of data per month without any reported performance degradation.
The initial stages of implementation were significantly outpaced by the volumes of individual databases, which have now reached up to 60 TB. However, detailed benchmark figures and workload-by-workload comparisons have not been published. Additionally, certain operations have been reported to be faster than they were on Oracle Exadata.
Sergey Shturmin, the deputy general director of GNIVC, stated that the project team initially had reservations regarding the ability of Postgres Pro Enterprise Certified to accommodate databases of this magnitude. The team determined that the platform was capable of accommodating the workload after conducting experiments, conducting operational testing at various stages, and implementing it in production.
The capacity to execute a single migration is equally important as the sustained performance of a large database system. The replacement environment must continue to support the tax authority’s daily operations while accommodating the increasing records and transaction volumes, as indicated by the reported monthly growth rate.
What the Migration Means for Russian Technology Sovereignty
The implementation illustrates that a database platform developed in Russia is capable of supporting large databases and critical government data in a live operational environment. This is a critical practical consideration for authorities that are attempting to decrease their reliance on foreign software vendors.
The project does establish that Postgres Pro Enterprise Certified has been deployed at a significant scale within one of Russia’s major government information systems. Its long-term importance will be contingent upon the sustained reliability, security, maintainability, and performance of the system as data volumes increase.