The author substantiates the approaches toward building reliable, scalable and safe information system and its core of modern data management infrastructure. The study urgency is determined by the institutional digital transformation, substantially increasing data volume, diversity and arrival rate? as well as by growing requirements of information service continuity, information accuracy, applied services performance and information security. In this context, the databases cease to be a software complex component, though a significant one, and become critical element of operation management, decision making, organizational sustainability. The purpose of the study is to classify architectural, technological and organizational solutions for effecting design, operation and development of data processing information systems. The author examines the key models of data warehouses, application of relational, non-relational, and hybrid solutions, and criteria of choice between SQL-, NoSQL- and NewSQL- approaches. She analyzes shared-nothing and shared-everything architecture, vertical and horizontal scaling mechanisms, sharding, replication, partitioning, use of cluster configurations and Database-as-a-Service cloud services. The practical value of the study resides in the possibility to use the above provisions in design, modernization and support of highload corporate information systems. The findings may be useful for information system designers, database administrators, architects and DevOps- and DBRE-engineers, as well as the managers of digital and infrastructural projects for formulating requirements to data storage, processing and security.