Data center (DPC)

A data center (DPC) is a critically important IT infrastructure facility, which is a specialized building, room or complex of rooms equipped with everything necessary for placing, operating and maintaining server, network, telecommunications and auxiliary equipment. A data center provides centralized processing, storage and data transfer, providing computing resources for the operation of corporate information systems, web services, cloud platforms, artificial intelligence systems and other digital solutions. In essence, a data center is the "brain" of the modern digital economy, on whose stable operation the continuity of business processes of thousands of companies and the availability of state services depend. The reliability of a data center is measured by the availability level (Tier), defined by the Uptime Institute: from Tier I (basic level, availability 99.671%, allowable downtime ~28.8 hours per year) to Tier IV (fault-tolerant infrastructure, availability 99.995%, allowable downtime ~26 minutes per year). The choice of Tier level directly affects the cost of construction and operation of a data center.

Key subsystems and engineering infrastructure of a data center

A modern data center consists of many interconnected subsystems, each of which is critically important for its uninterrupted operation. The engineering infrastructure includes a guaranteed and uninterrupted power supply system: uninterruptible power supplies (UPS) with double conversion, diesel generator units (DGU) with automatic transfer switch (ATS), distribution of power supply to racks through PDU (Power Distribution Unit) with the ability to monitor energy consumption. The cooling system: precision air conditioners (CRAC/CRAH), chillers with free cooling, organization of hot and cold aisles (hot aisle/cold aisle containment) to improve cooling efficiency. The fire extinguishing system: gas fire extinguishing (usually using inert gases or refrigerants), early smoke detection system (VESDA). The monitoring and management system (DCIM — Data Center Infrastructure Management) for controlling temperature, humidity, energy consumption and the state of all equipment in real time. The computing infrastructure consists of server racks (42U, 47U), servers (rack and blade servers), storage systems (SAN), network equipment (switches, routers, firewalls) and a structured cabling system (SCS) with a category not lower than Cat6A to support 10/25/40GbE. The scaling technology allows designing a data center with the ability to incrementally increase capacity, which is especially important for fast-growing companies. Virtualization based on the virtualization technology makes it possible to efficiently utilize resources and flexibly manage loads.

Data center reliability levels (Tier Classification)

The Uptime Institute standard distinguishes four levels of data center reliability. Tier I (basic): a single path for power distribution and cooling, without component redundancy — a planned shutdown for maintenance and repair is inevitable, availability 99.671%. Tier II (with redundancy): a single path, but with component redundancy (N+1), which reduces the time of planned and unplanned downtime, availability 99.741%. Tier III (concurrently maintainable): multiple active paths of power and cooling distribution, each component can be maintained and repaired without stopping the data center, availability 99.982%. Tier IV (fault tolerant): fully redundant infrastructure with multiple active independent paths, capable of surviving any planned or emergency situation (including failure of any component or path) without stopping services, availability 99.995%. The choice of Tier level is determined by the criticality of business applications and the project budget: for most commercial companies, Tier III is optimal. Each subsequent level increases construction cost by approximately 1.5-2 times.

Data centers in Russia and import substitution

In Russia, the data center market is actively developing, especially in the segment of commercial data centers (colocation, dedicated servers, cloud services). The largest sites are located in Moscow, the Moscow region, St. Petersburg, Kazan, Yekaterinburg and other million-plus cities. Russian data centers are obliged to comply with the requirements of legislation regarding the processing of personal data (152-FZ), ensuring the security of critical information infrastructure (187-FZ) and requirements for storing data on the territory of the Russian Federation. For state information systems, additional requirements are imposed on data placement and the equipment used. When building or modernizing a data center, the Fintech company provides a full range of services — from site selection and design to installation and commissioning. The installation service includes the supply and installation of server racks, laying SCS, installation and configuration of UPS and cooling systems, commissioning and testing. An integrated approach to the selection and integration of a data center allows an enterprise to obtain a reliable, scalable and cost-effective IT infrastructure that meets all regulatory requirements.