What is Database Management Software?
Database Management software or systems (DBMS)is the foundation stone of any business intelligence system.
Firstly, what does a DBMS do? Why does it exist?
To store the mountains of data or data records (a data record is set of specific data about an entity e.g. a person, their name, address, telephone number, data of birth etc) and to be able to retrieve a single record very quickly i.e. in microseconds from millions of records.
Software was developed to store the data in a particular format and to create indexes or tabs to quickly retrieve data rather than sequentially reading every record which would take a very long time! The reason why a DBMS uses indexes to retrieve data, is the disk storage mechanisms used are slow to read sequentially through the data.
Note, if the data was all loaded into memory, which is extremely fast, then possibly we would not such sophisticated database management systems that need careful design to maximise performance. One business intelligence (BI) system, Qlikview, operates this way.
In the old days (20 years) ago there used to several types of database systems. The main ones were hierarchical, network and relational.
Today relational database is the predominant type because of its simplicity.
Each object occurrence or row has a unique identifier used to retrieve the data correctly. This unique identifier is often called the primary key and is used to build indexes for fast retrieval of the data.
See the example table below that represents a typical database table for a customer. The Customer ID column could be the Unique Primary key.
The design of a relational database needs to be done carefully. A process called data normalisation is commonly used to ensure unique identification of the data and that there is no redundancy of data in the system i.e. the same information stored more than once. This is optimal because you would to update the same data several times in the system to ensure accurate records, and this slows down the update performance time.
The data normalisation process is a three step process and is often call putting the data into Third Normal Form (TNF).
Here are some common relational database management software systems today are:
Data warehouses are usually based on medium to large database management software systems.