M358 - Relational Databases - Information System Architectures
fellstrider.com - the logo!
Home| OU Study Rooms | M358 Index | Block 1 - Information Systems
File-based approach

Serial files are recorded one after another. They can only be accessed by starting at the beginning and working through until you find what you want. Direct access files are given a location. This position is determined when the file is written. Reading and writing to these files requires access via a computer program .

Such programs determine the location of the files. They also need to contain a definition of the record structure - data types for each field etc. Programming like this is convenient for single access application. However if more than one application is needed to access the same files, this will require the duplication of files. Updating one set of files will lead to subsequent inconsistencies between the two sets of files.

If there are additional requirements for files the access program must be altered to reflect the new file structure. Programs such as these are data dependent. Changing programs may lead to a change in the file. If there are a large number of files, accessed by a large number of programs, there could be confusion as to what should and shouldn't be changed.

There are management concerns with the file based system. Each program has to maintain its own file management. This will include access control, semantic constraints, data recovery, transactions and concurrency.

A file based system has problems from the point of view of data duplication, data dependence and data management. Is there an alternative?

Database approach

This is summarized as the management of data as a shared resource.

Image of a model of a database system

Main Features of a Database Management System (DBMS)

The diagram shows that all the necessary components within a database system go through a DBMS.

The database definition is known as a schema and is essential to any database system. This contains the specification of all data in a database and is used by the DBMS to control how the data is processed. The schema is independent of both the user processes and the DBMS. Data definitions are thus outside of any programs.

Database management systems

A DBMS may also be known as a:

Non-relational DBMS

Main functions of a DBMS

Data definition is through the creation and modification of a schema. This is initiated by a user via a database tool for defining data and its properties. There can be no database without the definition in an associated schema. Schema modification is a part of relational DBMS but not non-relational.

Constraint definition and enforcement is an important element of data properties within the schema. Constraints, once established are enforced automatically.

Access control is included in the schema. Defined in response to requests.

Data manipulation is controlled by the DBMS. A user can only update, retrieve or otherwise delete data that is allowed by the DBMS. Derived data is calculated by the DBMS.

Restructuring and reorganization are DBMS functions that allow the alteration of a database through the restructuring function.

Transaction support defines a transaction and ensures that its properties are maintained. DBMS automatically prevents partial transactions, however caused. Transactions must be complete entities.

Concurrency support allows multiple users to access a database concurrently. This will also prevent interference when more than one user transaction is updating the same data.

Recovery function wiil restore the database to a usable state in event of a failure/corruption. It is esential to the persistence of data that no completed transaction is lost. The process of recovery starts by the continuous logging of transactions as they occur.

Home| OU Study Rooms | M358 Index | Block 1 - Information Systems
Move on to Information System Architecture Page 2.

Valid CSS! Valid XHTML 1.0!

Comments, suggestions, ideas to
Stuart Banner