Server Types

Server Types

There are several kinds of databases which include:

• Flat file databases
• Relational databases
• Object databases
• Object relational databases

Flat files are simply files with a table of information which may be seperated by delimeters such as commas, colons, or semi-colons. Relational databases consist of several related tables of simple data. The tables are composed of rows and columns. Object databases store data in an object form rather than in tables. They store attributes and class information, but sometimes they also store and methods (behavior) in the database. Object relational databases are relational databases with data stored in tables, but they have a front end that converts objects to data and data to objects, making it seem to the application that objects are being stored.

Database servers include both a server program that serves remote clients and manages the database. They may use some means of standard communication between client and server to allow management of the data such as structured query language (SQL) for relational database servers.

The most popular databases today are Relational Database Management Systems (RDBMS). However, object database servers may someday overtake the relational database servers.

Relational Databases and Objects

If relational databases are used to store objects, the object must first be disassembled into parts, normalized, and placed in tables. This can take some time to do, and be a labor intensive process required for writing the code. To use the object, it must be reassembled.

Many relational databases are run on one single server and do not use a distributed architecture.

Object Database Servers

Object database servers may use an Object Query Language (OQL) as a standard language for communication. They may use an application programming interface (API) to allow the application to control the data or they may use both the API and OQL.

Current and Future Trends

Relational databases are still the most popular database in use today. There is good reason for this. They are easy to use and are normally efficient.

However as programming has changed, tools related to those changes must also change. Object oriented programming is becomming much more popular and as that occurs a more practical tool for long term storage of data is in demand. This tool must interface easily to the object oriented language in question. It must also be a standard tool so users are not tied to specific vendors amd should have a standard way of exchanging information between applications and the database. OQL was developed for this purpose, but it does not appear to be widely supported yet by object oriented database vendors.

Although object databases were first written many years ago, since they have not yet become popular, it appears that the market is not stable. There are several object oriented database vendors, and it is difficult to tell who will be in the market for the long haul. Therefore, I believe the purchase of an object oriented database is somewhat of a risk. This risk may be somewhat mitigated by the fact that programs can be written in object oriented language to isolate the programs from specific object database products.

If the benefits of the object oriented database are large enough for the particular application they are used for and the specific organization considering them, the risks are likely to be worth taking. However, if the benefits are marginal, it may be worth waiting another year or two for more market stability and uniformity