In order to support universal data access, it is important maintain a data dictionary or information about the stored data itself. This is known as Meta data. The provision of meta data allows the end-users to access the store data without any prior information about the where the data is located and how it is stored. For example, if the user wishes to access data related to sales, than the system will find out what sales refers to and where the data is stored by access the information stored in the meta data layer of the data warehouse, access the stored data and display it to the user (Humphries, Hawkins and Dy 1999).
Process Management Layer
The use of Process Management lays is in scheduling of various tasks that are required to create and maintain data warehouse as well as meta data for its storage.
Application Messaging Layer
The application messaging layer acts as a transportation system in the data warehouse. This layer contains much more than networking protocols. For example, it can be used to collection business transaction as well as messages and deliver them to specified locations at specified time.
Data Warehouse Layer
The Data Warehouse layer is the place where the actual data that is use for informational purposes is stored. This layer can also be defined to provide a logical or virtual view of the data that it stores. This lays stores the copies of external as well as operational databases in a form that is easy to access and highly flexible.
Data Staging layer
The final layer in the architecture of a data warehouse is the data staging layer. This layer comprises of the processes which are used to select, edit, summarize and load data warehouse as well as information access data from operational or external databases. Previously, this layer was the jumbled up with complex programming done to perform functions of this layer, however, it has been simplified through the availability of commercially available customized software. This layer may also hold applications which are used to identify patterns in the data is the stored or processes regularly (Ponniah 2001).