What is Database Virtualization?

Database virtualization centers on using database software to mask the physical location and configuration of a database from querying programs. Although the actual database used by the program might exist on several servers, the programs using it see it as if it is a normal database. Using virtualization technology simplifies database management, makes databases faster and lowers the cost of maintaining large database systems.

A standard database exists on a single computer. The database uses a single server or server cluster to store an information set. This set expands and contracts as information is stored or removed. These types of databases typically are easy to maintain, but they are less useful for the user because the number of simultaneous users and database queries are limited by the power of the database’s hardware. In addition, any problem with the computer or its network will render the database inaccessible.

Database virtualization alleviates some of these problems by decentralizing the database. A virtualized database can exist on several computers, in many locations and on multiple types of database software. In essence, the database is made of little pieces of databases connected to the system.

This process is made possible through something called a database virtualization layer. When a query is made to a database, it enters the software that makes up this layer. The layer then accesses a listing of database pieces and locations, finds the place holding the information and sends the request to that location. The initial query thinks that the database virtualization layer is the actual database that it spoke to and brings the information back as though nothing was different.

Database virtualization is better for people who use and maintain the database. From the user's perspective, he or she gains several advantages without any significant change to the applications. The user can continue to use the same programs as before, because the database virtualization layer controls the queries rather than the system. The information is more quickly accessible than before, because fewer people are querying the single location. Lastly, database downtime is reduced because the database is spread out over many locations.

From the administrator’s perspective, database virtualization works out very well. The database is already fragmented, so it is easy to have each part of the database existing in multiple locations at the same time. By having all of the database sections replicated, individual sections are easy to take offline for maintenance and repair. In case of a computer or network failure, the data is preserved in alternate locations. This backup and replication process is largely automatic, performed by the database virtualization software, freeing up time for the administrator to work on other things.