OBIT has developed a data compression service for Rosseti Siberia

2025: Data Compression Service Development

IT The solution operator UPHOLSTERED"" has developed a compression service data based on artificial intelligence the company "." Rosseti Siberia The solution reduced storage load by 35%, avoided purchasing additional equipment, and reduced storage infrastructure costs by three times. This was announced UPHOLSTERED by "" on September 12, 2025.

Rosseti Siberia, a subsidiary of PJSC Rosseti, transmits and distributes electricity in the Siberian and Far Eastern federal districts. The service area is over 1.974 million km ². The growth of data in the information systems of Rosseti Siberia required the operational expansion of storage systems (DSS). Given the high cost of solutions DSS and restrictions on the purchase of imported equipment, it was necessary to find a technological and cost-effective way to store data volumes without losing their integrity and availability.

Specialists of the operator of IT solutions "OBIT" have developed a data compression service for these tasks, completely located in the customer's perimeter. It does not require an Internet connection, is compatible with domestic products based on the Linux kernel and seamlessly integrates with the customer's information systems through an open API. The key feature of the module is the use of a built-in neural network that provides processing of documents and images taking into account the quality, size and compliance with internal document management standards.

The architecture of the IT solution includes a system for initially receiving requests, which analyzes incoming ones files and sends them to the corresponding processing branch - "document" or "image." If the object is an image, it passes through a chain of modules for preparation, processing and final quality assessment. For - PDFfiles uses a sequence consisting of document decomposition modules, component compression and subsequent assembly with the selection of an optimized option.

File processing is performed in automatic multithreaded mode. Compression parameters are configured according to the customer's requirements. Particular attention was paid to maintaining the "readability" of the text, the integrity and compliance of compressed objects with JPEG and PDF standards. This is necessary for correct operation in third-party applications. Data compression takes place at a given time interval: documents entering the internal document management system are automatically processed by the system and replaced with smaller copies. To evaluate the results, a neural network is used, which processes the document before and after compression, recognizes the text and determines its compliance with the established compression parameters. This approach allows you to automate quality control at a high level and minimize human participation in the process.

The project was implemented in stages over six months: from analyzing the volumes of data and scenarios for their optimization to setting up a compression model for business processes. All work was carried out iteratively with testing and customization for specific tasks of the customer.

As a result of the introduction of the service in Rosseti Siberia, as of September 2025, 5 TB of archive data has already been optimized, thereby reducing the storage load by 35%. The decision made it possible to reduce data storage costs by more than three times, - commented Mikhail Telegin, Deputy General Director for Strategic Projects "OBIT."

The flexible architecture of the IT solution allows you to scale the service to other segments of the customer's business processes, including document management and analytics.

The case for Rosseti Siberia clearly demonstrates that specialized solutions allow you to effectively solve application problems in the field of IT infrastructure. The developed data compression module can be scaled and adapted to other large enterprises, especially in industries with a high volume of digital archives and resource optimization requirements. At the same time, the solution combines high manufacturability, applied efficiency - which is critically important in a rapidly changing IT environment, the expert noted.