1. In today's world, every business enterprise either big or small maintains a large set of databases.
2. In such circumstances, it often turns out to be critical to acquire useful information from such databases, especially when each of these databases is not correlated.
3. Hence, to manage such circumstances, several researchers had put forth their efforts and came out with data mining techniques, like knowledge discovery in databases.
4. However, these researchers turned their attention towards other techniques due to the runtime failure of KDD.
5.Because of the existence of a large number of databases, their details of storage, their structures, etc.
6. Nowadays, every modern business enterprises immensely depend on a new type of aspect referred to as " Dataware House".
7. Here, the data warehouse itself refers to a huge database possessing connections to all other databases and serves the applications of all business functions.
8. Whenever we deal with data designing at the component level, it focuses representation of data structures accessible by various components forming a given software.
9. In order to do that, the following are certain important principles favoring it.
Data-centered architecture:
1. In the above architecture, essential data reside at the center of the architecture.
2. All the client software reauthorized to access this data.
3. This client software can easily manipulate the centered data they can delete, update, add, etc.
4. As the data manipulation can be done independently, hence, the centered data can be transformed into what called blackboard due to which a message is supplied to the client if he/she intends to access the data which is already been modified by other clients.
5. This architecture supports integrability which is due to the fact that the architecture is promoting independent accessing and manipulation of data store.
6. All the client software are authorized to access several of their processes independently, also as the data store can also act as a blackboard, this client can send and receive messages among themselves.
7. Whenever we require that a given data be transformed into a certain output by means of few transforming components in those conditions data flow architecture is prescribed.
8. Each filter in the above architecture transforms the data received by it into certain output independently and delivers it to other filters through pipes.
9. Also, each of these filters is independent of the functioning of all other filters.
10. At times, it happens that a condition is met referred to as batch sequential, in which the data lowing gets degenerated into a single line of transforms.
11. In those conditions, the architecture accesses a batch of data and transforms it into a certain output using one or more of these filters.
12.Reuse of transformations is possible.
13. New transformations are easy to add, making evolution easy.
14. It is simple to implement and transformations may execute sequentially or in parallel.
15. It is easy to understand the system in terms of input and output processing.
16. It requires a common format for data transfer that can be recognized by all transformations.
17. Interactive systems are difficult to write using the data-flow model because of the need for a stream of data to be processed.