Fedorus O. Application of PARCS systems for modeling operations of relational selection algebra.

Every day, more and more aspects of human existence and systems getting digital analogues, and those that already exist are overwhelmed by the amount of information they operate. Every minute of our existence is digitized and must be processed and saved for future inquiries. Such systems are forced to operate with incredibly large amounts of information while responding quickly to requests of varying complexity. Databases, responding to these challenges, began to seek ease in horizontal scaling and speeding up read/write operations. Because traditional databases did not keep up with user requirements, this led to the emergence of new systems called NoSQL. The term "NoSQL" has not been clearly defined and can be understood as "not only SQL" or "non-relational". These databases are very different from each other and from traditional relational databases. But there are several features that characterize the vast majority: • the ability to duplicate or distribute data across servers; • ability to scale horizontally; • ability to use RAM efficiently (data caching, storage of the entire database in RAM, etc.); • ability to store and process unstructured (or partially structured) data. Gradually, relational databases also began to provide similar functionality, but being designed with different requirements and without flexibility - the result has always been a trade-off between reliability, speed and functionality. Similarly, some non-relational databases begun to provide the functionality of traditional relational databases: transactions and links between tables. They tried to cover a wider range of user requirements. This course of actions indicates the impossibility of abandoning relational operations for searching in distributed systems.