A database is a system of storing information where numerous users can easily share information in a standardized format. Over the last decade, with technology and software applications growing at such a fast pace, databases rapidly replaced the traditional paper and filing cabinet systems used in previous decades. It’s pretty standard now in most industries with a big ongoing push in the healthcare industry.
Over the years, many different types of databases have emerged and, over time, some models become outdated, while others continue to be a popular model for today's databases. A common characteristic is that every model has two primary components: structure and operation. Regardless of the model, each database will have a definitive structure and a method of operation to access data. Here are the most common types of databases:
The hierarchical model is one of the earlier models of databases. Its philosophy dictates data is structured according to chain of command and formatted in a "tree-like" arrangement; the hierarchical model is structured in a way where information can be searched from top to bottom. 1 It also is described as called a structure with a "parent-child" relationship. It has rapid performance, but is a little cumbersome because information has to be accessed from the "top" down. A parent can have multiple children, but a child can only have one parent.
For example: Vendor>Order Number>Service Rendered. This may slow down search performance because if you want to start your search for "Service Rendered" it will have to sort through the other two pieces first because the information sought is located at the bottom. The biggest downside to using this model is that the environment is restrictive and, as a result, this model rapidly declined in usage and will likely become obsolete at some point. Some legacy systems may still have this structure.
The "Network model is named from the network of connections between the data elements" (Gerard Post), and the objective of structuring the data in this format promotes the philosophical thinking that users could search data from different viewpoints because they are linked together; each structure is referred to as a network. It has many of the same qualities as the hierarchical model. The primary problem with this model is that the database developer must think of every possible query that might be asked of the database because the structure is in such a way where indexes are created before any questions are asked and a search performed.
This model doesn't support the efficiency that other newer models represent. This model is space-consuming and requires a lot of time management allotted to it for maintenance. The network model, like the Hierarchical, is also becoming dated.
A relational database is the most commonly used database model today. It is comprised of a group of logically connected tables (relations), and data is organized to serve applications with speed and efficiency. The concept of a relational database was originally developed in the 1970s by E.F. Codd. Since its birth, this model has significantly grown in popularity. It is a very flexible model for database operation and does a beautiful job of cross-referencing when relationships are accurately assessed and underlying tables are correctly designed. In the relational model it's crucial to eliminate "repeating" groups of data (such as a customer's name appearing in more than one table) and to correctly identify the relationships between the entities (this is referred to as "normalization").
For instance, a customer is an example of an entity, the address is an attribute and these properties are stored in a "customer table"; an order they made is also an object and the order ID number and product purchased are properties stored under an "orders table." The database designer would link these two tables as having a "relationship" because customers make orders. "Employees table" might be another table, but since employees aren't typically associated with "orders" there would not be a relationship established there. Not every table needs to be linked together in the relational model. It's an ideal design for a database and it's not surprising that it remains the most popular model today because of its flexibility to encompass many objects.
This is an interesting database model since it carries a similar philosophy as the relational model because of the common link of being constructed with a theoretical foundation. The concept of the object-oriented model is relatively new and acts with the philosophy that properties can be "grouped" as objects with classes and subclasses. The objects include properties and methods; it supports links between varying objects and includes all their "child" properties. In order to understand the concepts of this model, it's helpful to have general knowledge of how object-oriented concepts work (some computer programming languages operate under this philosophy also). The concepts of object-oriented design are using objects, classes, methods, messages and inheritance. Objects may hold other objects, making inheritance an important concept in this model.
The above designs illustrate some of the different approaches that database designers might take in developing one of these databases that society uses all the time. Databases are very useful and necessary way to organize information in a digital world. People use databases in some shape or form every day. For instance, ATM machines get money for consumers from their accounts which are a part of the bank's database system. When people shop online, they access store databases; even submitting this article adds it into a database.
Fast forward to 2015 and most organizations have electronic records management or use some form of digital database management. However, what the future holds should be interesting. While relational databases are still firmly rooted in use, will it always be that way? What is often referred to is "post-relational" database models is becoming a more common term to hear. Computerworld has a good three-part series on the topic, also referring to upcoming models as "NoSQL alternatives" (SQL standing for standardized query language - a simple, but powerful and effective computer language).
There seems to be disagreement on whether or not relational database models are heading towards extinction. For the time being, however, it appears this model still is very useful for most organizations.
Additional reference: Post, Gerald, Database Mgt. Systems, 2005