Two classic experimental systems

Lecture 7. System R: general system organization, basics of the SQL language

The System R relational database management system was developed in the IBM research laboratory in 1975-1979. This work had a revolutionary influence on the development of the theory and practice of relational systems throughout the world. It was System R that practically proved the viability of the relational approach to database management.

After successfully completing the creation of this system and obtaining experimental results of its use, a number of commercially available relational systems were developed, including based on the direct development of System R (the capabilities of one of the commercially available relational systems, DB2, are described in Russian). K. Data's book "Guide to relational DBMS DB2). The experience gained in developing this system is extremely important. Practically in all later relational DBMS in one or another article The methods used in System R are used.

After completing the development of System R, IBM actively continued work on relational DBMS, and in several directions. The first direction we have already noted - the development of commercial relational databases. The second direction is building a distributed relational DBMS based on the ideas of System R. An experimental version of such a system, System R *, was successfully developed by IBM. This work also greatly enriched the experience of researchers and developers of distributed DBMS. Finally, the third area is research and development of relational systems designed for non-traditional applications.

An extensive bibliography is dedicated to System RMS. For information, we give it at the end of this lecture. Although the development of this system officially began in 1975, the first publications related to this system appeared as early as 1974. In particular, one of the first publications proposed the basis of the base language System R SQL (then this language was called SEQUEL, and until so many people call it that way; by the way, System R developers (and now Oracle) recommend pronouncing the SQL name exactly as SEQUEL). Since the publications appeared in the course of the practical implementation of the system, each of them reflects the state of affairs (ideological and practical) precisely at that stage of work when the corresponding article was written. Some ideas and ideas naturally changed as work progressed. Only the final publications provide a relatively complete picture of the system as a whole. On the other hand, many interesting moments are not reflected at all in these recent articles, and we will try to provide a more complete overview of the ideas and methods applied in System R. In this case, we will dwell on some possible alternative solutions that were found by the developers of System R, but practically were not used.