Home

Research

Chapter 10. Conclusion and Future work

10.1. Conclusion

SagaMap plays a vital role in providing support for automated data warehouses [8]. It is simple to use, highly interactive and provides an easy means to creating a new data warehouse. It also acts as a reliable tool to quickly explore schema of the source database in order to generate schema for the data warehouse.

SagaMap hides the underlying complex mechanisms from its users, except where it is absolutely appropriate and necessary to expose them. In effect, even non-technical users can create, populate and update data warehouses with minimal time and effort.

Attributes from source tables can be mapped into new attributes in the warehouse database tables using aggregate functions. Then, relevant data is automatically transported from the source database to the newly created warehouse. The tool thus integrates warehouse creation, schema mapping and data population into a single general-purpose tool.

This tool has been designed as a component of a framework, whose users are Database Administrators. They will also be able to synchronize updates of multiple copies of the data warehouse. Warehouse images that need to be updated are taken offline and applications that need to access the data warehouse can now access any of the other image warehouses. The Image Switcher built into this framework switches between databases in a way that is totally transparent to applications so that they do not realize existence of multiple copies of the data warehouse. It also ensures that ongoing transactions on a particular database are not interrupted when the database is scheduled to be taken offline.

SagaMap is system-independent. This gives it advantages such as portability and wide application. This tool can access any database with minimal effort since there is no hard coding of information in the application.

10.2. Future Work

The current system has been designed to support only a single Oracle database as the source database. In the future, the system can be enhanced to support multiple source databases of heterogeneous nature. Since no other database other than Oracle 9i supports the Metadata API, there is no direct way to extract the DDL information and it has be accomplished by indirect ways. Hence the cost of such an upgrade to support multiple source databases will be expensive in terms of implementation effort, but nevertheless, it will give users a great deal of flexibility.

In the system, DDL information is stored in XML files. When users need to manipulate it, information stored in the XML files is read and displayed to users. Users may choose to edit the information and save it to the XML file. In the future, performance of this operation can be improved by storing XML documents in memory and writing them as files only when essential. 

In the current system, the users are limited to creating new attributes based on existing attributes present in the corresponding table in the source database. In future the users should be able to create new attributes based on existing attributes from across multiple tables in the source database.