Lecture
Of course, despite their attractiveness, classical relational database management systems are limited. They are ideal for traditional applications such as ticket reservation systems or hotel rooms, and banking systems, but their use in design automation systems, intelligent training systems, and other knowledge-based systems is often difficult. This is primarily due to the primitiveness of the data structures underlying the relational data model. Flat normalized relationships are universal and theoretically sufficient to represent data of any subject area. However, in non-traditional applications in the database, there are hundreds, if not thousands of tables, on which constantly perform the expensive join operations necessary to recreate the complex data structures inherent in the subject area.
Another serious limitation of relational systems is their relatively weak capabilities in terms of representing the semantics of an application. The most that relational DBMSs provide is the ability to formulate and maintain data integrity constraints. As we noted in Lecture 6, after designing a relational database, many of the designer’s knowledge remain, at best, on paper due to the lack of appropriate means of expression in the system.
Aware of these limitations and shortcomings of relational systems, database researchers perform numerous projects based on ideas that go beyond the limits of the relational data model. In all likelihood, any of these works will become the basis of future database systems. It should be noted that the subject of modern research related to databases is extremely wide. In the final part of the course we present only a short overview of the most important areas.
Comments
To leave a comment
IBM System R — реляционная СУБД
Terms: IBM System R — реляционная СУБД