What Makes a Relational Database Management System Relational?

What Makes a Relational Database Management System Relational?

Having already discussed the meaning of both relational database and database management system, it may appear that the subject is settled. But the natural implications of the relational model are so numerous and profound that critics contend that, even today, a “truly relational” database management system doesn’t exist. For example, Dr. Edgar Codd wanted the database management system to treat views in the same manner as base tables whenever possible, but the problem of view updateability is unsolved to the present day. Codd listed more than 300 separate requirements that a database management system must meet in order to fulfill his vision properly, and I have time for just one of them: physical data independence. Here is the relevant quote from Codd’s book:

RP-1 Physical Data Independence: The DBMS permits a suitably authorized user to make changes in storage representation, in access method, or in both—for example, for performance reasons. Application programs and terminal activities remain logically unimpaired whenever any such changes are made.

What Codd meant was that you and I shouldn’t have to worry about implementation details such as the storage structures used to store data.

Summary

I hope that you now have an appreciation for the theoretical foundations of Oracle Database 12c. You can find more information about the subjects I touched on in the books mentioned in the bibliography at the end of the chapter. Here is a short summary of the concepts discussed in this chapter:

• A database is an information repository that must be competently administered using the principles laid out in the IT Infrastructure Library (ITIL), including security management, availability management, continuity management, change management, incident management, problem management, configuration management, release management, and capacity management.

• Relation is a precise mathematical term for what is loosely called a data table. Relational database technology swept aside earlier technologies because of the power and expressiveness of relational algebra and because it made performance the responsibility of the database management system instead of the application developer.

• A database management system provides efficient algorithms for the processing of table operations as well as indexing schemes for data storage. The query optimizer rearranges relational algebra expressions in the interests of efficiency but without changing the meaning or the results that are produced.

• A database management system is defined as a software layer that provides services such as transaction management, data security, and data integrity.

• A transaction is a logical unit of work characterized by atomicity, consistency, isolation, and durability.

• Relational database theory has many consequences, including logical data independence, which implies that changes to the way in which data is stored or indexed shouldn’t affect the logical behavior of application programs.

Further Reading

• Bachman, Charles. “The Programmer as Navigator.” ACM Turing Award Lecture, 1973. http://amturing.acm.org/award_winners/bachman_9385610.cfm. To properly understand Codd’s theory, you must understand what preceded it.

• Codd, E. F. “Relational Database: A Practical Foundation for Productivity.” ACM Turing Award Lecture, 1981. http://amturing.acm.org/award_winners/ codd_1000892.cfm. It’s a shame that almost no database professionals have read the papers and articles written by the inventor of relational database theory. In my opinion, the 1981 ACM Turing Award Lecture should be required reading for every database professional.

• Silberschatz, Abraham, Henry Korth, and S. Sudarshan. Database System Concepts. 6th ed. McGraw-Hill, 2010. If you’re going buy just one book on database theory, this is the one I recommend. This college textbook, now in its sixth edition, offers not only copious amounts of theory but also some coverage of commercial products such as Oracle Database.