Database Management

Creating database systems have several steps, and these are design, implementation, and project management. In this article, the focus is on the first step, the database systems design. In this stage, you deal with issues such as how access is managed, how server systems are set up, and how big the database is.

The database system design usually starts off with a rough model, like most designs usually do. The model demonstrates how the data is stored and accessed. What kind of retrieval method is in place? How is the data indexed? What kind of search is applied? What queries can be run on the data? These are just some questions that are taken in consideration during the design phase. Even at this early part of the process, it's a good idea to implement database management already to make sure you're starting things right. A database manager with a solid foundation on database design is going to be your best partner in a database systems project.

As mentioned above, the database system design step involves choosing a model for your database. The common models for databases are the hierarchical, network, relational, object-relational, and object. The hierarchical model arranges data in a branching-out pattern with a 'single parent-multiple children' relationships, which leaves you with a diagram resembling a family tree. The network model, if represented in two-dimensional form, shows a lattice pattern because of its 'multiple parents-multiple children' relationship. The relational model arranges data in a table wherein pieces of information intersect at certain points, allowing relationships between data to be defined as well. The object-relational model is a relational model with the allowance for customized data types to be integrated in. Then there is the object model which works best with object-oriented languages such as C++ and java. The object model basically makes data entries look and work like objects or modules, which are referenced through queries and call functions.

The right model out of the five mentioned above for your database depends on how you're planning to use it. For example, a relational database is a good choice for an application whose main feature is an advanced search function that allows users to qualify their queries. Amazon.com, for one, benefits from a relational database model because of the vast number of customized searches done by users within a day.

Once you've chosen a model to work with, you need to build a prototype model on which you'll perform a 'pilot' test, or a test done on a small scale for strictly evaluation purposes. From there you can determine if your design is working well not just in theory, but in actuality as well. If it's the latter, you can move on to determining how to implement your model on a large-scale basis.