III. Read the text and do the tasks following it

Database management systems (DBMS)— comprehensive software tools that allow users to create, maintain, and manipulate an integral base of business data—have become very popular in the business community as a means of supporting management information systems. In an integrated database, the records are logically keyed to one another so that all data on a topic can be retrieved by simple requests. Users will probably not get involved in designing databases (except for supplying job information needs to designers), but they will likely be involved in using a DBMS, either on a workstation or a terminal, or on a stand-alone microcomputer.

File management systems used to be the only way of managing data and files. In these systems, data was stored in a series of unrelated files on tape or disk. The major problems associated with file management systems are (1) data redundancy — the same data appears in more than one file; (2) tedious updating procedures — because the same data appeared in many places, updating files was time-consuming; (3) poor data integrity—if some redundant data elements were missed during file updating, they were no longer current and could cause inaccurate information to be produced; and (4) lack of data and program independence — programmers could not use the data file to develop new programs because the data and the programs were restricted by existing formats. To update either the applications program or the data file became a major task.

Database management systems were developed to (1) make data independent of the programs, so that it is easy to access and change; (2) eliminate data redundancy; (5) establish relationships among records in different files; (4) define data characteristics; (5) manage file directories; (6) maintain data integrity; and (7) provide a means of securing access to the database.

DBMS software often uses a query language as an interface between the user and the system. This interface allows users to easily ask questions of the DBMS and obtain information to answer the questions and produce reports.

DBMS software also includes capabilities to simplify report writing and maintain the database (utilities) as well as to allow different applications programs to use the database (DML). During the design of the database, a data dictionary is constructed that contains all the data descriptions used by the DBMS to locate and retrieve data. The DBMS also can include a transaction log of current activity. This log can be used to update necessary backup copies of the database in case of failure of or damage to the operating database system.

A DBMS is usually modeled after one of three structures; (1) hierarchical, (2) network, or (3) relational. These models differ in terms of the cost of implementation, speed, degree of data redundancy, ease with which they can satisfy information requirements, and ease with which they can be updated.

The hierarchical database model resembles a family tree; the records are organized in a one-to-many relationship, meaning that one parent record can have many child records. Records are retrieved from a hierarchical model by starting at the root record at the top and moving down through the structure. There is no connection between separate branches.

The network database model is similar to the hierarchical model, but it allows multiple one-to-many relationships (each child can have more than one parent) and relationships between records in different groups. Also, access to the database can be made from a number of points — not just from the top.

The relational database model is made up of many tables, called relations, in which related data elements are stored. The data elements are in rows, called tuples, and columns, called attributes. The main objective of the relational database model is to allow complex logical relationships between records to be expressed in a simple fashion.

In general, the hierarchical and network models are less expensive to implement and allow faster access to data. However, they are more difficult to update and aren't as effective at satisfying information requirements as the relational model can be. Because the hierarchical and network models are older than the relational model, they are used most often on large computers. The newer relational model is used extensively on microcomputers; however, because relational models are so flexible in satisfying information requests, and because they aren't difficult to update, relational models are being used more and more on large computers.

The process of database design is usually carried out exclusively by specialists; however, users may have occasion to set up small databases for microcomputers. First, the logical design is set up; then the physical design.

The logical design, which has to do with what the database is, from the business users' perspective, involves defining user information needs; analyzing data element requirements and logical groupings; finalizing the design; and creating the data dictionary. During the logical design the schema (the organization of the entire database, including the structure type) and the subschema (the way certain data elements are linked) are defined.

The physical design involves how the data will be physically arranged and stored on the storage devices used by the DBMS. During this design phase the database is actually set up, and the data is stored.

Databases are important business resources, and DBMS software is sophisticated; both of them must be carefully operated and controlled. For this purpose, the job of database administrator (DBA) has been created. The main responsibilities of a database administrator are (1) guiding database design; (2) overseeing database implementation and operation; (3) user coordination; (4) backup and recovery; (5) performance monitoring; and (6) system security.

In general, the main advantages of database managment systems are (1) elimination of data redundancy; (2) increased ease of file updating; (3) increased data independence and simplification of program and maintenance; (4) increased user productivity and efficiency; (5) increased security; and (6) standardization of data definitions.

The main disadvantages are (1) complexity; (2) high cost of implementation; and (3) vulnerability of consolidated business data in a central database.